PDA

View Full Version : cad/cam, what are you guys using?



jacampb2
04-26-2008, 07:03 PM
I have been using Acad autodesk '06 for a few years now, and have gotten fairly good with it, both for 2d and 3d modeling. From reading my mach3 manual again, a cam program or post processor is highly recommended. So, what software are you guys using? Preferably affordably priced. I haven't looked at the included software with mach3, haven't had time yet. I know I saw a lot of discussion in a thread a while back, but I am not turning it up. I could have sworn it was one of Evan's threads, but I am not seeing it. If you have the link, that is fine by me too.

Anyhow, I would like to be able to do 3d, or is it 2.5d, I am not really sure what the heck the difference is.

Tell me what you all recommend.

Thanks,
Jason

BobWarfield
04-27-2008, 01:23 AM
Rhino3D for CAD, OneCNC for CAM. That latter is nice, but not cheap. I talked them into a student discount, and it still wasn't cheap. But it is a nice package.

Cheers,

BW

mochinist
04-27-2008, 01:53 AM
acad and solidworks for cad and featurecam for cad/cam, nothing cheap though. You might try searching thru a few of Evans threads from this year, I think he is using cambam now, but not positive.

these two links will give you a better idea of what the differance between 2.5d and 3d is

http://www.featurecam.com/general/software/featuremill2d.asp
http://www.featurecam.com/general/software/featuremill3d.asp

jacampb2
04-27-2008, 07:38 AM
Thanks guys, I will keep researching today, I was out of time last night, and thought you all might have some good suggestions. I will look through the posted software first.

Thanks,
Jason

brian Rupnow
04-27-2008, 07:38 AM
I am a design engineer, and work in 3D Solidworks. It is an amazing program, and I can not say enough good things about it. if I want to export to a shop for CNC work I can export in .dwg, .dxf, or .iges formats. www.rupnowdesign.com

jacampb2
04-27-2008, 08:02 AM
Thanks! I tried to learn solidworks back in the late '90s(?) I never had a lot of luck with it. I probably did not put enough effort into it, as I picked up Acad fairly easily when I had a need for it. Out of curiosity, are you saying solidworks is easier to model in over all, or just for 3d assemblies.

Thanks,
Jason

Evan
04-27-2008, 09:16 AM
I am using CamBam now for most of my CAM work. However the version I am using isn't publicly available as it is in beta test. For CAD I use a variety of programs from a very old copy to Turbo Cad to several of the free trial packages to plain old finger cam.

jacampb2
04-27-2008, 09:18 AM
Okay, I have been reading for about 2hrs about 2.5 vs. 3d and I am still confused. It appears that 2.5d basically means a 3d shape can be machined from several layers of different depth cuts, but cannot machine any overhangs.

I think this will be sufficient for my machine, but I do have a 4th axis rotab that I intend to use eventually. Will 2.5d cam handle things like gears and spline machined w/ a 4th axis?

Thanks for the answers so far. I did search back through Evans post, I thought for sure he listed all of the software he is using in one of the threads, maybe it was in a reply to someone else, and I only searched his threads. Anyhow, I can't find it, but cambam looks pretty nice. I will probably try it out.

Thanks,
Jason

IOWOLF
04-27-2008, 09:22 AM
I am using CamBam now for most of my CAM work. However the version I am using isn't publicly available as it is in beta test. For CAD I use a variety of programs from a very old copy to Turbo Cad to several of the free trial packages to plain old finger cam.

Big surprise Evan, You always have the best and newest stuff.:)

Stepside
04-27-2008, 09:44 AM
2.5D means the mill can do 3D objects but it move only 2 axis at one time a 3D machine moves any or all of the 3 axis at the same time.

I draw with RhinoV4 and exploring Visual Mill and MadCam for machining. Rhino costs less than Solidworks and AutoCad but it is superior for developing 3D objects for milling. If you have AutoCad experience then Rhino is real easy to pick up.

John Stevenson
04-27-2008, 11:35 AM
2.5D means the mill can do 3D objects but it move only 2 axis at one time a 3D machine moves any or all of the 3 axis at the same time.



Slight correction.
"2.5D means the mill can do 3D objects but it move only interpolate 2 axis at one time"

Many controllers more than machines, can do 3D provided that the Z move is in short line moves.

.

brian Rupnow
04-27-2008, 02:16 PM
Thanks! I tried to learn solidworks back in the late '90s(?) I never had a lot of luck with it. I probably did not put enough effort into it, as I picked up Acad fairly easily when I had a need for it. Out of curiosity, are you saying solidworks is easier to model in over all, or just for 3d assemblies.

Thanks,
Jason

Solidworks is easier for everything!!! First you design the parts---This is fairly simple, as you draw and dimension one face of the part, same as in 2D---then you extrude it to the proper depth to create the 3D part model. After you make all the "parts" which are saved as "part files", you drag and drop all the individual "parts" into an assembly. Solidworks provides you with tools to assemble these parts, same as you would use in real life with real parts. For instance, if 2 parts have bolt holes that are bolted together, you highlight the two circular faces and put a "concentric" mate on them---this will align the two holes. Then you decide which faces on the 2 parts would be next to one and other, and highlight the 2 faces and put a "coincident" mate on them. This moves the two faces of the parts together. then the only freedom of movement the 2 parts have is the fact that they can rotate about the centerline of the bolt hole, so you select 2 faces and either put a coincident mate on them or an angle mate to fix that third range of movement. You can assemble over 1000 peices this way on a complex machine!!! Then this file is saved as an "assembly" file. To make 2D drawings, you just drag and drop the parts or assemblies into a predefined "drawing" sheet, and it creates all the views automatically. you never have to 'draw". You create parts and assembly files, and all the "drawing" is a fully automated function. Remember---You have already defined the math data of the parts and assemblies when you created them. Its probably not quite as simple as I make it sound, ---- I use it 8 hours a day 5 days a week, and I teach courses in it as well, but it certainly is a wonderfull program.---Brian

gundog
04-27-2008, 11:43 PM
I just bought one of these router tables and read this it helped me understand the difference about 2D / 2.5D / 3D.

I am in the process of putting it together now I should be giving it a test run in the next day or two.

http://www.shopbottools.com/3-d_work_v2.htm

Mike

DR
04-28-2008, 04:07 AM
2.5D means the mill can do 3D objects but it move only 2 axis at one time a 3D machine moves any or all of the 3 axis at the same time.

.................




Actually, if you do a little research you'll find those terms have no specific meaning beyond what the writer wants them to mean.

I originally learned 2.5D meant it would move all three axes simultaneously, but only as a linear move. 3D was a machine that would do out-of-plain arcs using ail three axes simultaneously.

Both of my mills will do my definition of 2.5D. One will also do limited 3D using axis rotation for an out-of-plain arc, but not fully 3D since the arc must be in a plain rotated around the X,Y or Z axis.

I have since learned my ideas of these terms do not match those of others.

John Stevenson
04-28-2008, 04:34 AM
DR is correct in what he's written.
Full 3D is able to move in three simultaneous arcs and is a feature of the controller more than software and the machine.

True 3D controllers are rare and expensive.

Most controllers can't and fudge it by stepping one axis to get the 3D move.

It is also possible to interpolate an arc using other than XY.
There are 3 standard G codes for this in most controllers G17, G18 and G19

G17 is the usual XY arc like a circle cut in the flat.

G18 is an arc in XZ, imagine a pendulum swinging from Z along the X side to side axis

G19 is an arc in YZ, again a pendulum singing from Z in the Y plane

Evan
04-28-2008, 04:43 AM
I submit that the term "3D" has a very well defined meaning. It takes six variables to define the transformation of an object in 3D space. X, Y, Z location and A, B and C rotation around the X, Y and Z axes.

Transformation means the possible change in position of an object. In order to be able to machine an object fully in 3D space the cutter must be able to make transformations that are the reciprocal of the transformation of an object being moved in 3D.

A simple example is the ability to make a spiral cut with variable radius such as a pipe thread at any angle and any orientation. That doesn't mean the cutter must be able to reach up from underneath for example but the same effective result must be possible. That would normally be achieved by rotating the object in a "4th" axis.

Mach 3 supports operation in up to all six axes of translation and rotation but cannot do full interpolation in any arbitrary three axes at once.

So, to be considered capable of full 3d then the machine must be able to interpolate in any 3 of 6 axes simultaneously, which is what John said in quite a few less words.

Ah, I see John added a few words while I was typing. :D

J Tiers
04-28-2008, 08:26 AM
I am simply using the "cad" who cam in the door w' me.

Follows me around and I see him in the mirror.

DR
04-28-2008, 09:11 AM
I submit that the term "3D" has a very well defined meaning.

..................................




Yes, but not when describing a CNC machine.

The confusion comes when a machine's capability is described using the terms 2D, 2.5D and 3D. In that context they have no universally understood and agreed upon meaning. It's best to question the user as to exactly what is meant.

My first machine was described in the factory sales literature as 2.5D. I questioned why it was not considered 3D since it could machine a 3D surace. The explanation was the 3D surface would have to be approximated by a series of connected (short) linear moves. Obviously, this was the factory's interpretation. Not everyone agrees with it.

Swarf&Sparks
04-28-2008, 09:25 AM
Is this not a question of aliasing?
With digital equipment, some degree of aliasing is inevitable.
Interpolation, or interpretation?

Evan
04-28-2008, 10:19 AM
Is this not a question of aliasing?
With digital equipment, some degree of aliasing is inevitable.
Interpolation, or interpretation?

Not necessarily. As long as the sample rate exceeds the required resolution by at least twice aliasing does not occur. In other words, if we require a position accuracy of .001 then we must calculate the position to at least .0005 or better to avoid aliasing errors.

I think I know what you are getting at though. If a series of independent straight line moves is calculated and made it is indistinguishable from true interpolation. It's really only a question of where such calculations take place. The real limiting factor is the mechanical ability of the machine to present the tool to the work and vice versa.

Swarf&Sparks
04-28-2008, 10:50 AM
I'm not sure that the Nyquist limit applies to precision engineering Evan, but I'm always prepared to listen :D

karlgabel2
04-28-2008, 11:16 AM
I use Solidworks and love it! Have used TurboCad and AutoCad Lite. TurboCad was okay for 2D but poor for 3D but that was a long time ago (about 8 years ago). Could never get the hang of AutoCad Lite - command interface was a pain. Solidworks turns out to be very intuitive for me. I've used it professionally for electronic packaging and use it for mechanical drawings for extracurricular pursuits. Very nice!

lazlo
04-28-2008, 11:27 AM
I'm not sure that the Nyquist limit applies to precision engineering Evan, but I'm always prepared to listen :D

It doesn't apply here. The Nyquist theorem states that you have to digitally sample an analog signal at twice the highest frequency component to correctly re-construct it.

What you're doing with a 2.5 D CNC controller is interpolating a smooth curve, and you're absolutely right, if you're not simultaneously driving all three axis', as you interpolate the curve, you'll still have stair-steps in the cut.

As you increase the interpolated resolution finer and finer, the stair-steps will be less visible, until you exceed the DOC of the cutter, at which point it should be indistinguishable from a true, simultaneous 3 axis controller.

Swarf&Sparks
04-28-2008, 11:30 AM
My point exactly Lazlo
Perhaps 'aliasing' was a poor choice of words.

"jaggies" was the highly technical term we used back in the CGA vs Herc days :D

lazlo
04-28-2008, 11:42 AM
I'm not sure what the right term is either -- dithering is probably closer than aliasing, but dithering has a specific meaning in digital signal processing (which is effectively what you're doing: digital/CNC replication of an analog signal: a smooth curve) that doesn't quite fit either.

Swarf&Sparks
04-28-2008, 11:58 AM
Hmm, "jitter" perhaps?

Or maybe we should let this one through to the keeper/goalie/catcher.

Evan
04-28-2008, 12:09 PM
It doesn't apply here. The Nyquist theorem states that you have to digitally sample an analog signal at twice the highest frequency component to correctly re-construct it.
Sure it applies. When you calculate a position you are in effect creating a sample of the curve in question. That is exactly what the Nyquist sampling theorem deals with.

[edit]

You can run a Fourier transform of the end point coordinates of all the line segments used to approximate a curve and it will show the frequencies that are contained in the sample set. This is only the inverse of sampling a frequency in order to represent it without aliasing.

Swarf&Sparks
04-28-2008, 12:33 PM
But it will still only approximate the curve (or waveform)

Evan
04-28-2008, 12:42 PM
In practice, yes. In theory no, but the theory makes certain assumptions that aren't attainable in practice.

The Nyquist/Shannon/Whittaker sampling theorem is applicable to virtually any analog to digital to analog processing both in time and space domains. In particular, the reconstruction part of the theorem is called the Shannon/Whittaker interpolation formula and has direct bearing on the problem at hand here.



There are two limiting conditions that the function x(t) must satisfy in order for the interpolation formula to be guaranteed to reconstruct it exactly:

x(t) must be bandlimited. In other words, the function must have a Fourier transform http://upload.wikimedia.org/math/f/7/6/f763eabe48d1654050df68d42a72a238.png for |f| > B for some maximum frequency, B > 0.
The sampling rate, fs, must exceed twice the bandwidth, 2B. Equivalently: http://upload.wikimedia.org/math/d/3/3/d3300d92c4a840bd81daa988e9ebfe3a.png The interpolation formula reconstructs the original signal x(t), as long as these two conditions are met. Otherwise, aliasing occurs; that is, frequencies at or above fs/2 are erroneously reconstructed.


http://en.wikipedia.org/wiki/Whittaker%E2%80%93Shannon_interpolation_formula

lazlo
04-28-2008, 12:52 PM
Sure it applies. When you calculate a position you are in effect creating a sample of the curve in question. That is exactly what the Nyquist sampling theorem deals with.

No. Nyquist says how often you need to digitally sample an analog signal to correctly reconstruct it.
So if you have a 2 Hz sine wave, and you sample it at 2 Hz, you can get a straight line if you sample at exactly the wrong points.

With your 2.5D CNC, you're interpolating a curve, and Nyquist doesn't apply:

Plug a unit sine wave into a CAD/CAM model. It has one frequency component, with a frequency of 1. You'll still need an infinite number of interpolated steps to cut a smooth version of it.


But it will still only approximate the curve (or waveform)

Exactly.

Swarf&Sparks
04-28-2008, 01:02 PM
My apologies for the hijack, Jason :)

Evan
04-28-2008, 02:12 PM
With your 2.5D CNC, you're interpolating a curve, and Nyquist doesn't apply:
You cannot win this argument since your opinion is incorrect. Sampling theorem is completely extensible to time AND space. This is a problem for the reconstruction part of sampling theorem. It's 100% applicable. That is why it is called the Whitakker-Shannon Interpolation Formula. Nyquist was only a minor contributor to the theorem.


The sampling theorem is usually formulated for functions of a single variable. Consequently, the theorem is directly applicable to time-dependent signals and is normally formulated in that context. However, the sampling theorem can be extended in a straightforward way to functions of arbitrarily many variables.
...
Similar to one-dimensional discrete-time signals, images can also suffer from aliasing if the sampling resolution, or pixel density, is inadequate.

http://en.wikipedia.org/wiki/Nyquist%E2%80%93Shannon_sampling_theorem

There is no difference between an image of a curve and a mathematical description of a curve. That's how vector graphics works. It can be extended to any number of dimensions you choose. The machining of a curve from a vector graphic representation in a CAD drawing is merely a physical realization of the curve and the process of describing to the machine controller is entirely within the realm of sampling theorem.

That is all I have to say about it.

lazlo
04-28-2008, 02:22 PM
You cannot win this argument since your opinion is incorrect.

That is all I have to say about it.

You mean, that's all Wikipedia has to say about it. :rolleyes:

So here's a challenge for you Evan: draw a unit (1 "Hz", but translate into spatial dimensions = 1 meter, 1 centimeter, whatever) sine wave. Since the since wave has 1 frequency component, so you don't have to do any Fourier analysis. Nyquist says that you can correctly sample that waveform with 2 points (three vectors). Now cut that sine wave in CAM with three vectors :)

Evan
04-28-2008, 03:20 PM
You do the math. If you find any errors you can edit the wikipedia page yourself.

You will note that Nyquist et. al. says it can be done theoretically but there are practical limitations in the extreme case. Read the page, you may learn something.

George Bulliss
04-28-2008, 03:35 PM
What was the title of the thread again?

lazlo
04-28-2008, 03:41 PM
You do the math. If you find any errors you can edit the wikipedia page yourself.

I haven't read the Wikipedia page for Nyquist's Sampling Theorem, but I'm reasonably sure they've got it correct.

The problem is, that sampling theory has virtually nothing to do with interpolating curves with discrete controllers. Nyquist's sampling theory, as Swarf 'N Sparks correctly noted, tells you how often you need to digitally sample an analog signal.


Read the page, you may learn something.

Don't need to Evan. I have two degrees in electrical engineering. I suggest that you're not going to learn undergraduate digital filter theory by reading a Wikipedia entry.

lazlo
04-28-2008, 04:04 PM
What was the title of the thread again?

What Evan is trying to say, in a round-about way, is that 2.5D is "just as good" as 3D CAM, because the 2.5D controller can interpolate enough discrete steps to give a good approximation of a smooth curve.

That probably true, but it has nothing to do with Nyquist sampling theory.

On a related note, why doesn't MACH (or EMC2) support true simultaneous 3 (or 4) axis control? There's plenty of compute power on the average PC workstation...


True 3D controllers are rare and expensive.

Aren't most commerical controllers true 3D (or 4D or 5D) controllers? From looking at the cool Haas and Mazak videos, those are definitely true simultaneous CAM....

JSGAuto
04-28-2008, 04:17 PM
I use Pro/E. Suprised others don't use it, I supose its due to the cost.

The complete package is nice, modeling, transfer to a drawing, then go right into Pro/NC for CNC programing. Exports in all the usual languages.

Not new CAD user friendly.

Go with Solidworks or Rino (wich I think is a sweet deal).

Jim

John Stevenson
04-28-2008, 04:20 PM
Aren't most commerical controllers true 3D (or 4D or 5D) controllers? From looking at the cool Haas and Mazak videos, those are definitely true simultaneous CAM....

Can't comment on that as I have never run anything that new.
One thing I have seen recently is a company just up the road from me make Aerospace parts.

They have just bought a brand new Haaas with a 2 metre X travel to make wing spars for one of the Airbus's

When I was in there the other week I was watching this work and boy is it fast.
I asked what the rapids were and was told 100metres per minute [ mearly 4,000in / min ]
I then asked what speed they were cutting and was told 100 metres per minute :eek:

Looking at the code as it scrolled thru it was all G00 moves but I did notice that even this machine was stepping in the Z, only interpolating in the X and Y.

I was told the Haas was built to a price that was just good enough for the job, any better and it wouldn't have been affordable.

.

John Stevenson
04-28-2008, 04:22 PM
I use Pro/E. Suprised others don't use it, I supose its due to the cost.


Jim

I'm not suprised at all as this is a home shop board.
Not a lot of point dragging a $500 lump of dead iron home to throw $15,000 at it to draw pretty pictures ;) :D

.

lazlo
04-28-2008, 04:35 PM
Not a lot of point dragging a $500 lump of dead iron home to throw $15,000 at it to draw pretty pictures ;) :D

Well, folks keep raving about Rhino, but that's $995. There's an educational discount down to $199, but I looked into the educational version of SolidWorks when I was taking welding classes at the local CC, and SolidWorks requires you to re-buy the educational license every 2 years -- does Rhino do that?

rdesign
04-28-2008, 05:46 PM
I spent a few years doing machine design in AutoCAD, 2 years in ProE (2000, 2000i and i^2) (never ran wildfire) before moving to solidworks. I run solidworks every day at work. I own Rhino V4 and Rhino CAM basic at home. I have used mastercam X but not enough to really offer an opinion. I have been able to do everything I need with rhino CAM, I only do 4th axis positioning manually on my CNC. RhinoCAM basic does not do 4 axis unless you spend more on the pro version or a 4th axis option specifically. You can get creative and do a lot of 4 axis work without a motorized 4th though.

Rhino/Rhino cam can pay for themselves after a handful of projects and I am happy with them. It is definitely not Solidworks or ProE for part modeling though.

Mcruff
04-28-2008, 05:55 PM
Well I went and put my flame proof suit on, so let the fireballs loose!!

I suggest BobCad!

I have used it for years, since version 16. It has quirks and bugs but for a hobbiest it is still hard to beat. The new version 22 is a totallly different ball game than all the other version. It has a real post processor. You can wheel and deal with them on the phone and get it down to about $200-$300. There sales department is horrible and worse than most used car lots when it comes to calling back or not taking no for an answer.
I have used it for work off and on for the last 10 years. That being said if you buy it make sure you join a support forum, you will need it as there documentation is horrible. There is a guy named Sorin that teaches classes on it that will help immensely.
Here is a link ot his website:
http://www.cadcamtrainer.com/
Some things it does very easy, some things it takes some fiddling to get the post just right but all in all for the money it is well worth it for a hobbiest or for a fledgling company to start out with.

Oh and by the way I have used lots of other sofwares to compare it with so I do know what high end software is capable of.

lazlo
04-28-2008, 06:10 PM
I own Rhino V4 and Rhino CAM basic at home.

If you don't mind me asking -- how much did you pay for them? Is there a Rhino CAD/CAM bundle?

Stepside
04-28-2008, 07:58 PM
Rhino V4 is the latest version. The $995.00 is full boat/List , you can do better from some of the resellers. The student edition is also the real meal deal. It is the full commercial product. It does not time out and you can upgrade when a new version becomes available. You can avail youself to the same support that those who bought the commmercial get.

RhinoCam is really MecSofts Visual Mill set-up to run inside RhinoV4. It is not a Rhino product. That said it works very well with the RhinoV4.

Compare the cost of SolidWorks, AutoCad, MasterCam with Rhino and RhinoCam and then look at what you want to accomplish. There are some lesser known software packages that are less money than those I have mentioned that could be more than adequate for ones projects. Big is always better when growing pumpkins and a few other possibilities. The correct/best tool for the project is best the rest of the time.

toastydeath
04-28-2008, 08:24 PM
Aren't most commerical controllers true 3D (or 4D or 5D) controllers? From looking at the cool Haas and Mazak videos, those are definitely true simultaneous CAM....

Yes. All of our Fanuc controls (even from the 80's) have 4 axis controls. Whether they have a physical 4th axis is a different matter, the control still supports it and all the plugs are there. My school's late 90's Haas is the same way. '97 (I believe?) VF-0, three axis. There's another early 90's Mitsubishi control we have, just as another point of reference, and it is full 4D. When I do a G1 X1. Y1. Z1. F10., I'd throw a s***-fit if all three axes did not move simultaneously and within the following limits of the machine.

2.5D is only common with the hobby/desktop machine market. When you get a VMC with a real control on it, it does what you tell it to.

RobbieKnobbie
04-28-2008, 08:43 PM
I spent a few years doing machine design in AutoCAD, 2 years in ProE (2000, 2000i and i^2) (never ran wildfire) before moving to solidworks. I run solidworks every day at work.

I pretty much grew up on AutoCAD, spent a little time with Pro/E Wildfire 2 - but never made friends with it.

I'm using Solidworks now and couldn't be much happier. Excellent package and amazingly powerful. None of the learning curve issues or fussiness of Pro/E.

From speaking with a Pro/E rep recently, they're evidently taking quite a beating from SW - even after they lowered the price to SolidWorks levels.

Funny market. Always changing.

Evan
04-28-2008, 09:10 PM
With your two degrees Robert you really should know more about sampling theorem. Nyquist-Shannon sampling theorem is applicable to any discipline where data points are used to represent continuous functions. An arc or curve are merely a small part of a periodic function. While the theory says that sampling at twice the frequency is sufficient to describe the curve that is not true in practice because of the limitations imposed by mathematics. It would require computing an infinite series, not possible in reality.

In practice sampling at only twice the fundamental frequency of the function only assures that aliasing won't occur. It cannot in practice be used to accurately represent the function sampled. Below is an illustration of this. When sampling a sine wave the function sampled describes a circle. This is precisely what is happening when a circle or arc is described in a series of short straight line segments in CNC machining. If we sample at twice the function fundamental frequency the best we can achieve is a square wave that represents the maximum amplitude of the function as shown in the top sine wave. At worst we end up sampling at the zero crossings and end up with nothing as shown in the next down sine wave. The circles at the right show the result of the sampling and also represent what we can infer from the data points when we try to reconstruct the function from the data alone. In the second case all we get is a horizontal line with the two samples per period on either end.

If we increase the sample rate to 4 times the Nyquist frequency then we can at best create the the polygon shown on the right from the 4 data points per period. In the bottom case we have increased the sample rate to eight times per period and now begin to approximate the circle.

http://vts.bc.ca/pics3/sample.jpg

When we produce a circle by circular interpolation we have the advantage of already knowing what shape we want so we are able to create as many data points as necessary to represent the curve to whatever degree of accuracy our system permits. The process of creating the samples is completely subject to the limits imposed by sampling theorem and the degree of accuracy is determined by the relation of the number of samples to the Nyquist frequency as determined from the Fourier frequency plot of the curve.

oldtiffie
04-28-2008, 09:41 PM
Fine two-sided (lazlo/Evan) discussion/debate so far, which I am very interested in as I can follow most of it and can relate to it. Keep the discussion going please.

But on a more practical and "down to earth" note, can we have a series of parallel discussions of/on CAD and post-processors that are cheap and able to be bought and used by the average small HSM with a limited budget and with no real "hi-end" requirements. Just enough to draw simple projects, export via DXF, post-process them and use them on a small lathe/mill using Mach3 - or what-ever other small cheap effective CAD/CAM/post-processing software can be recommended/acquired.

Any advice will be appreciated.

J Tiers
04-28-2008, 10:04 PM
Evan, There is one thing that is left out of your explanation...... And it is the reason that laslo is correct, as you do in fact seem to be admitting now.

The INFORMATION is sufficient when the sampling is over 2x the data........

But there is the matter of the reconstruction filter...

The sampled forms you drew out, and the polygons etc, ALL INCLUDE FREQUENCIES SUBSTANTIALLY GREATER THAN THE ORIGINAL DATA.

The "other half" of the system is to filter out all frequencies OVER the original data, so as to "reconstruct" the original wave.

Your sampled square wave, for example, when filtered, as above, has only its fundamental frequency left. However, since its rate is EXACTLY 2x the data rate, it cannot give a correct amplitude, and it is still "distorted".

The problem with the CNC and sampling is that there is nothing in the system analogous to the reconstruction filter, and so the high frequency components are still present.

The system traces from data point to data point, and there is all kinds of distortion. This will still be true with 3x the 'frequency" or 4. In fact, it cannot be avoided.

The only "filter" you have available is the cutter diameter. With sufficient cutter diameter, you may reduce teh distortion considerably. But I don't know, (without performing some analysis that I don't want to do, and would have to think about) that it is quite the same thing as the signal filter as per the above.

lazlo
04-28-2008, 10:14 PM
Now you're cutting and pasting random quotes and pictures from Nyquist sampling theory, but I'm waiting for an Ebay snipe to trigger, so I'll play:


With your two degrees Robert you really should know more about sampling theorem.

I actually have three degrees ;) A Bachelor's and Master's in Electrical Engineering, and a second Master's in Computer Science. But let's get back to CNC interpolation...


Nyquist-Shannon sampling theorem is applicable to any discipline where data points are used to represent continuous functions.

This quote is correct, but taken somewhat out of context. Nyquist sampling theory applies to digitally sampling any analog signal. It's a rule applied to digital signal processing, including audio and graphics.

Unfortunately, cutting a curve with a CNC toolpath is not sampling an analog signal, and therefore has nothing to do with Nyquist sampling theory.


While the theory says that sampling at twice the frequency is sufficient to describe the curve that is not true in practice because of the limitations imposed by mathematics. It would require computing an infinite series, not possible in reality.

You don't understand how Fourier Analysis works, or applies here. A single component signal, like the pure sine wave I proposed earlier in the thread, and which you have copied in your post from some random Nyquist tutorial, is a single term function, and not an infinite series.

An infinite series occurs when you're sampling a complex signal with an infinite number of frequency components. Fourier analysis breaks each frequency component into a separate sine wave, such that the sum of all the sine waves == the original signal.

If you have a mathematical discontinuity, such as a square wave signal, it takes an infinite series of frequency components to represent. But the reason I picked the pure sine wave for your challenge, and the reason it's in the Nyquist diagram you copied, is because a pure sine wave has a single frequency component, and therefore no Fourier analysis is necessary:


Plug a unit sine wave into a CAD/CAM model. It has one frequency component, with a frequency of 1. You'll still need an infinite number of interpolated steps to cut a smooth version of it.


In practice sampling at only twice the fundamental frequency of the function only assures that aliasing won't occur. It cannot in practice be used to accurately represent the function sampled. Below is an illustration of this.

That's not what that Nyquist diagram is showing Evan. That's the classic undergraduate Nyquist sampling diagram that I described to you in this post:

http://bbs.homeshopmachinist.net/showpost.php?p=353319&postcount=30


No. Nyquist says how often you need to digitally sample an analog signal to correctly reconstruct it.

So if you have a 2 Hz sine wave, and you sample it at 2 Hz, you can get a straight line if you sample at exactly the wrong points.

Which is exactly the example in the section of the Nyquist diagram you snatched. That's not a coincidence by the way, this is the same exact example you use to teach Nyquist to undergraduates:

2 Hz Sine Wave with 2 Hz Sampling:
http://i164.photobucket.com/albums/u15/rtgeorge_album/nyquist1.gif

And the third picture is with ideal Nyquist sampling: a 2 Hz pure sine wave with 4 Hz (Nyquist) sampling frequency, and the since wave is captured perfectly:

2 Hz Sine Wave with 4 Hz Sampling:
http://i164.photobucket.com/albums/u15/rtgeorge_album/nyquist2.gif

This is also the exact sine wave in the challenge I proposed to you to show you why Nyquist Sampling theory has nothing to do with CNC toolpaths: Take that 2 Hz Sine wave, and make a 2.5D or 3D model of it. Nyquist says, and your diagram shows, 2 sample points per cycle is necessary and sufficient to describe the sine wave. Now cut that sine wave on a CNC toolpath.

Despite the fact that's it's a pure sine wave, with a single frequency term, it will still take an infinite number of interpolated cutting steps to cut a smooth version of it.


So here's a challenge for you Evan: draw a unit (1 "Hz", but translate into spatial dimensions = 1 meter, 1 centimeter, whatever) sine wave. Since the since wave has 1 frequency component, so you don't have to do any Fourier analysis. Nyquist says that you can correctly sample that waveform with 2 points (three vectors). Now cut that sine wave in CAM with three vectors :)

So if you really understand Nyquist Sampling Theory, it will take you 10 minutes to generate that sine wave in CAD and create a GCode file of it. It will be immediately apparently to you that it will take an infinite number of vectors to cut a smooth version of it, as Swarf&Sparks aptly pointed out.

lazlo
04-28-2008, 10:17 PM
I'm waiting for an Ebay snipe to trigger, so I'll play:

Tiffie and Swarf&Sparks -- I just won my snipe, and it appears that my NIB Henrob set is, in fact, now available :)

http://cgi.ebay.com/ws/eBayISAPI.dll?ViewItem&item=120250920468

J Tiers
04-28-2008, 10:50 PM
Laslo:

First please note my reply above which was posted while you were typing industriously....... I agree with most of what you said previously.

Then think for a second, and I think you will see a reason Nyquist MIGHT apply.

You DO have a set of data points, representing, if you will, samples of some arbitrary waveform......

if you satisfy the sampling requirement, you DO have (as I noted, and Evan claims) the data which will reconstruct that wave.

What is missing, as I noted above, is the reconstruction filter.

I somewhat jokingly suggested cutter diameter as the "filter"............... It could just work......... I admit to not having analysed the matter other that a quick qualitative estimate....

toastydeath
04-28-2008, 10:52 PM
lazlo is spot on with interpolated cnc paths, which is 99% of machining and the default mode for machine controllers.

The caveat is that high speed machining controllers do not treat the path like other controls do, and will fit a spline to the points. They also use acceleration and jerk based algorithms to produce a stepless toolpath.

lazlo
04-28-2008, 11:02 PM
First please note my reply above which was posted while you were typing industriously....... I agree with most of what you said previously.

Then think for a second, and I think you will see a reason Nyquist MIGHT apply.

You DO have a set of data points, representing, if you will, samples of some arbitrary waveform......

if you satisfy the sampling requirement, you DO have (as I noted, and Evan claims) the data which will reconstruct that wave.

Right, agree completely JT -- in the sine wave example, you have a 2 Hz sine and 4 datapoints, and if you had the right reconstruction filter, the CAM controller could reconstruct the path.


What is missing, as I noted above, is the reconstruction filter.

I see two problems: like you say, the reconstruction filter, which in the case of the sine wave, would have to generate a b-spline -type function. The second problem is that CNC controllers, especially the 2.5D hobbyist controllers, are programmed with G-Code interpolated per axis, and not digital reconstruction filters :)

So the process I was describing in the CAD/CAM challenge to Evan was that in order to cut a pure sine wave, the CAM program has generate enough vectors to smoothly interpolate the curve, even though mathematically it can be described by a single continuous function: sin(X), and discretely as 4 Nyquist points.

Now, if you were very clever, perhaps you could write a MACH plug-in that would take a Mathlab description of the reconstruction filter, the Nyquist points, and generate curved toolpaths. That would "fix" the interpolated curve problem that 2.5D controllers have....

oldtiffie
04-29-2008, 12:23 AM
OK.

I'd like to step in here as a novice and neophyte as regards the theory and address the practical application in my shop with a fairly common mill (HF-45 column mill).

I would like to know and ask how this affects me in my shop with my average CAD, post-processor and CNC set-up for the work I want to within my small budget and limited knowledge and capabilities.

Good discussion so far. It is staying "on track" and neither (too?) vindictive or personal - thanks for that.

I can follow most of it. My calculus is long forgotten as is most of my (analogue) servo theory and practice/application.

I can appreciate the number of points to be addressed in any function.

I am aware too that most/all CAD systems resolve/regard/address most splines and curves as a very large number of very short lines - at least in the graphical outcome. The better CAD systems have higher resolutions and number of decimal places addressed or resolved to as well.

OK? Or not?

The CNC file for use in a CNC machine (say Mach3) is a post-processed version of a DXF file which is exported from the CAD system as a series of vectors (straight lines at a pre-determined orientation).

Still OK? Or not?

The post-processed file output is a series of pulses that cause the CNC "motors" to "step"/"pulse"/"jerk" around their axis in very small pre-determined rotational increments.

Still OK? Or not?

These small discrete individual steps are transmitted to a (say) ball-screw/lead-screw that has mass and inertia as well as an elastic torsional distortional component and so will tend to "smooth" the "steps". This "smoothed" output from the ball/lead-screw is transmitted to the (say) mill table "X" or "Y" motion which is further "smoothed" by the mass of the machine table/s.

Still OK? Or not?

I can appreciate that some "dither" must be transmitted as a function of "stepping" and that the "dither" improves the "sensitivity" of the machine movement response in terms of minimising the effects of "stiction" but that too much dither can and will cause undue wear on machine components as well as at worst showing up in the work-piece finish.

I guess that at the end of it all, that there must be allowances and compromises made to have the best result of the theory and computing at the work-face.

I would guess that the average HSM-er would be aiming to get a "satisfactory" (for him) CAD and CNC system.

So let's address that as well.

Not-with-standing the theory so well expressed and explained thus far, at what point is the computed resolution adequate for the movement of the job/work-piece in relation to the (end-milling?) cutter to provide a sufficiently accurate result as regards size, finish and time generally and with regard to the type/class of machine and CAD/CNC system in the average HSM shop?

toastydeath
04-29-2008, 01:16 AM
In the spirit of your questions tiffie, I am going to keep my responses strictly in the realm of HSM. Please bear in mind i am not mentioning exceptions or things that happen on expensive control platforms.


I can appreciate the number of points to be addressed in any function.

I am aware too that most/all CAD systems resolve/regard/address most splines and curves as a very large number of very short lines - at least in the graphical outcome. The better CAD systems have higher resolutions and number of decimal places addressed or resolved to as well.

OK? Or not?


Sort of OK. You are spot on in the small lines thing. The wrong part is that most machines are limited to four decimal places, and CAM systems pretty much spit out four unless you need more for some reason.



The CNC file for use in a CNC machine (say Mach3) is a post-processed version of a DXF file which is exported from the CAD system as a series of vectors (straight lines at a pre-determined orientation).

Still OK? Or not?


Yep. A series of G01 moves, usually made incrementally at the machine's maximum resolution.



The post-processed file output is a series of pulses that cause the CNC "motors" to "step"/"pulse"/"jerk" around their axis in very small pre-determined rotational increments.

Still OK? Or not?


Sort of. There's a wide variety even in HSM controls on how they handle the increments. Some do small jerks, some try and make it smooth. Usually yes, you wind up with a bunch of small segements. You'll see reference made to "velocity based control" when you go to servos, in that a servomotor can drive at a specific velocity rather than stepping. Obviously this has advantages when you try to speed things up.



These small discrete individual steps are transmitted to a (say) ball-screw/lead-screw that has mass and inertia as well as an elastic torsional distortional component and so will tend to "smooth" the "steps". This "smoothed" output from the ball/lead-screw is transmitted to the (say) mill table "X" or "Y" motion which is further "smoothed" by the mass of the machine table/s.

Still OK? Or not?


No. The control must actively be aware of the machine dynamics in order to average things out. Get a hobby mill cutting a large circle at 100 ipm or so. The combination of control deficiency (pure interpolation) and machine dynamics will produce a very nasty scalloping. Most hobby mills have very limited look ahead (which starts to get into splines and curve fitting in a very basic way) if any, and the interpolation approach begins to break down at moderate feed rates. Many hobby mills won't cut a decently sized circle at over 40 ipm.



Not-with-standing the theory so well expressed and explained thus far, at what point is the computed resolution adequate for the movement of the job/work-piece in relation to the (end-milling?) cutter to provide a sufficiently accurate result as regards size, finish and time generally and with regard to the type/class of machine and CAD/CNC system in the average HSM shop?

If I was a betting man, i'd say around .0002-.0003 for most hobby mills. The control scheme (pure interpolation) breaks down with a bunch of short, quick moves. So you are limited in speed anyway. In terms of accuracy, you don't need more than your equipment can handle. If you had a pair of decent servos driving ballscrews, I'd say .0001" resolution is plenty, and is what most older commercial VMCs go with for interpolation.

Again, with interpolation, the slower you go, the more resolution you can make use of. So if you wanted a real dang smooth circle, you'd have to feed at fractional inches per minute and use encoders that supported better resolution. You won't gain accuracy (at all), but you do gain more points on the curve.

oldtiffie
04-29-2008, 01:33 AM
Thanks TD for taking the time and putting in the effort and skills that you have in crafting that excellent reply.

I have checked to see that you are still "on-line".

I will respond to your post and will include a couple of other "as it is" matters and comments as well later.

We have to go out for a while so I guess you will see it later today or tomorrow.

Evan
04-29-2008, 02:06 AM
Now you're cutting and pasting random quotes and pictures from Nyquist sampling theory, but I'm waiting for an Ebay snipe to trigger, so I'll play
You don't get it do you? I drew that diagram myself.


Unfortunately, cutting a curve with a CNC toolpath is not sampling an analog signal, and therefore has nothing to do with Nyquist sampling theory
No, it isn't. It is synthesizing samples to reconstruct an analog function. You don't seem to have noticed that is what I am talking about, yet. I am talking about the construction of continuous functions with discrete data points, which is the second half of Nyquist-Shannon sampling theory.



You don't understand how Fourier Analysis works, or applies here. A single component signal, like the pure sine wave I proposed earlier in the thread, and which you have copied in your post from some random Nyquist tutorial, is a single term function, and not an infinite series.
The tutorial is mine.

You don't understand sampling theory. To reconstruct a sine wave sampled at the Nyquist frequency theoretically requires the summing of an infinite series of terms. A square wave is the product of an infinite series of sine wave harmonics. To reconstruct the sine wave fundamental from the square wave produced by sampling at the Nyquist frequency a deconvolution must be performed which is the inverse function of constructing a square wave from a series of sine waves.



If you have a mathematical discontinuity, such as a square wave signal, it takes an infinite series of frequency components to represent. But the reason I picked the pure sine wave for your challenge, and the reason it's in the Nyquist diagram you copied, is because a pure sine wave has a single frequency component, and therefore no Fourier analysis is necessary:
You have certainly stuck your foot in your mouth this time Robert.

I must know what I am talking about. You have said so three times now.


I see two problems: like you say, the reconstruction filter, which in the case of the sine wave, would have to generate a b-spline -type function. The second problem is that CNC controllers, especially the 2.5D hobbyist controllers, are programmed with G-Code interpolated per axis, and not digital reconstruction filters
What do you call approximating a circle then? Since pi is irrational you don't have any choice but to approximate it with a series of data points. It doesn't matter if you use splines or anything else, they can all be deconstructed to a series of straight line moves since the machine must be instructed to move in discrete steps. [edit] There is a name for that. It's called sampling.

Jerry,

There is a filter in Mach 3 that takes care of the high frequency components cause by oversampling. When Mach 3 is running in constant velocity mode it cuts corners, literally. At the points that correspond to a discontinuity, a "jaggie", Mach 3 departs from the absolute toolpath and and in effect provides a low pass filtering function. The faster the cutting speed the greater the departure from the calculated toolpath.

oldtiffie
04-29-2008, 06:28 AM
I have been using Acad autodesk '06 for a few years now, and have gotten fairly good with it, both for 2d and 3d modeling. From reading my mach3 manual again, a cam program or post processor is highly recommended. So, what software are you guys using? Preferably affordably priced. I haven't looked at the included software with mach3, haven't had time yet. I know I saw a lot of discussion in a thread a while back, but I am not turning it up. I could have sworn it was one of Evan's threads, but I am not seeing it. If you have the link, that is fine by me too.

Anyhow, I would like to be able to do 3d, or is it 2.5d, I am not really sure what the heck the difference is.

Tell me what you all recommend.

Thanks,
Jason

I thought that given the number of posts that it might be timely to re-visit Jason's Original Post (OP) - see above quote.

Toasty Death (TD) in his post at his post #57 at:
http://bbs.homeshopmachinist.net/showpost.php?p=353455&postcount=57
did an excellent job of addressing issues that I had raised at my post #65 at:
http://bbs.homeshopmachinist.net/showpost.php?p=353450&postcount=56

Some further relevant issues at the following concurrent thread at posts my #5 and #7 and John Stevenson's post #8 at:
http://bbs.homeshopmachinist.net/showthread.php?t=28861

So to keep it in the context of the (I hope) typical HSM-er who may have or be interested in CNC at non-Professional/Academic level, I thought I'd press on.

Those who are or continue to want to read or explore the Professional/Academic level should/can read/refer to the excellent and informative posts between lazlo and Evan on this thread.

For myself, TD's advice that a step/move value of 0.0001" (1 tenth) should suffice for a/the HSM environment seemed very reasonable.

That suggests that a 0.2000" lead ball-screw would need 0.2000/0.0001 = 2,000 pulses per revolution of the ball-screw.

He also suggested that cutting a circle on a HSM machine (my presumption) using say Mach3 should not exceed 40" per minute. That suggests - for circles - the ball-screw should not exceed 40/0.2 = 200 RPM which means a pulse rate - as I understand it - of 200 x 2,000 = 400, 000 (4 x 10^5) pulses per minute = 6,666.6667 = 6.667 x 10^3 pulses per second.

OK now? If not please advise.

As the CNC process relies on dead-recking (ie where it - the table - should be) as opposed to knowing where it is (by feed-back as would be the case in a servo system) the only apparent way of knowing where the table actually is relative the the "X" and "Y" ways is to use absolute measurements via a very good Digital Read-out (DRO) system. That might assist in finding errors due to "back-lash", "end-float" and "gib adjustment" etc.

So the quality and rigidity of the machine and its set-up are very relevant as well.

So, as those may be regarded as "given" the only variables that we (or at least I) need to know are those as asked by the OP are or seem to be:

1.
what is the best "bang for buck" CAD system/software?:

2.
what is the best "bang for buck" post-processor?:

3.
what is the best "bang for buck" CNC system/software?:

4.
and if the answer to (3) is not Mach3, why isn't it?

John Stevenson
04-29-2008, 06:46 AM
2.
what is the best "bang for buck" post-processor?:

There isn't one.
A post processor is part of the CAM program and not a stand alone article.
A post processor from one CAM program will not run with another.

Posts as they are called can be designed either by the user or the CAM vendors to do just what is needed.

If you want to always retract to 30mm above the work to miss clamps this can be written in.

A good post is only good for that user.

.

oldtiffie
04-29-2008, 07:38 AM
Thanks for the correction John - appreciated.

Swarf&Sparks
04-29-2008, 07:45 AM
This has been (is) a fascinating thread. Thanks all for constructive posts.

Mick (tiff) my only experience with CNC is home-built kit (a router) and Vcarve software with Mach3 running the machine.

I have no hesitation in recommending mach3. It is let down by poor documentation, but there is so much info out there on various sites, that this is no big problem. I could spend all day re-programming macros (buttons) but don't need to (so far).

One last comment. Where the hell was sieg/syil when I started this project 3-4 years ago? :D

oldtiffie
04-29-2008, 08:09 AM
Thanks Lin.

Yes it is fascinating - good questions, good reponses, staying on topic - you name a good point and its here.

And with luck it will continue.



.......................................
One last comment. Where the hell was sieg/syil when I started this project 3-4 years ago?

To paraphrase you a bit Lin - where the hell is the new Seig now??

For John Stevenson:
any further advice re. time of "show"/"release" and comments re features John?

I realise that you are constrained/embargo-ed and have confidentiality agreements etc. as is normal commercial practice, but any "hints" or advice?

I'm sure that I am not the only one interested.

Perhaps a new/separate thread or a resuscitated older relevant thread just for continuity?

John Stevenson
04-29-2008, 08:10 AM
One last comment. Where the hell was Sieg when I started this project 3-4 years ago? :D

At the start of a learning curve both in manufacturing and marketing.
At the same time a lot of other points were coming together.
There were probably two major players in the breakout board area.
Two for stepper drives, both with problems at that time.

Now that has evolved into a new industry in it's own right.

Mach3 had just evolved from Mach2.
Go into help about and read the list of beta testers who contributed hours of work and time to get Mach to where it is today.

You can't see the list on the Vectric forum as it only shows for beta testers but they have a similar list and may people appear on both lists.

It is to ALL these unsung hero's that the success of home shop CNC is all about.

They are the ones who have made this all possible and allowed others to stand on their shoulders.

And a big thanks to Virgin Airways for their frequent flyer miles between London Heathrow and Shanghai :p

.

Swarf&Sparks
04-29-2008, 08:15 AM
No axe to grind, nor any connection, but Syil Australia are showing X4 ready to ship.

http://www.syil.com.au/product_X4Plus.php

John Stevenson
04-29-2008, 08:15 AM
For John Stevenson:
any further advice re. time of "show"/"release" and comments re features John?

I realise that you are constrained/embargo-ed and have confidentiality agreements etc. as is normal commercial practice, but any "hints" or advice?

I'm sure that I am not the only one interested.

Perhaps a new/separate thread or a resuscitated older relevant thread just for continuity?

There will be a new thread but a quicky here is that they will be announced with prices this week in the UK and both machines will be at Harrogate under power in two weekends time.
I belive the first shipment is due to land in June in the UK.

I have both machines here at the moment building enclosures for safe running at the shows. They were air freighted in two weeks ago.

I have got to nip back over to China just after the Harrogate show to do some work on the new lathes.

.

John Stevenson
04-29-2008, 08:19 AM
No axe to grind, nor any connection, but Syil Australia are showing X4 ready to ship.

http://www.syil.com.au/product_X4Plus.php

They are having that built for them in the north of China as Sieg will not supply them CNC machines.
It very political.

Interesting in that they are still using the old SX3 head.

.

J Tiers
04-29-2008, 08:30 AM
Jerry,

There is a filter in Mach 3 that takes care of the high frequency components cause by oversampling. When Mach 3 is running in constant velocity mode it cuts corners, literally. At the points that correspond to a discontinuity, a "jaggie", Mach 3 departs from the absolute toolpath and and in effect provides a low pass filtering function. The faster the cutting speed the greater the departure from the calculated toolpath.

I don't know if the filter you describe will help.

it's easy to find out, try laslo's test.

The filter not only must "exist", it must also be at a "frequency" under half the sample rate..... otherwise it does not eliminate the effects.

Your square wave at exactly 2x the sine frequency is exactly why the filter is needed. Go back and look...... At that frequency you lose the amplitude data. At a slightly higher sample frequency, you can reconstruct it.

But you have to filter at below the critical frequency.

IF the filter you describe can be set to be so, which I do not know, AND if it is a sufficient "brick wall" to remove harmonics effectively, then it COULD work.

Since it is a compromise intended to reduce jaggies, and otherwise stay out of the way, I suspect it isn't what would be needed. But it certainly IS the SORT of thing required.

There is also the question as to whether the "sample rate" was sufficiently consistent to be filterable. If not, it will introduce other forms of "distortion".

Since the whole thing (from original data points through filters, cutter diamter, etc) is presumably not a "system" designed according to the requirements of sampling theory, I strongly doubt it really represents a classis 'sampled system", and so probably will not operate in the way that might be expected.

Swarf&Sparks
04-29-2008, 08:34 AM
Thanks for the constructive comments John.
I had more than a sneaking suspicion that you had something to do with the X4 development.
I also note that they are using epoxy aggregate stiffening.
Comment, or confidential?

Evan
04-29-2008, 08:43 AM
Whether it is designed as a "sampling system" or not is irrelevant. Sampling theory still applies. The machine and the controller are not a sampling system but a sample reconstructor. When it cuts an arc it is taking a series of samples produced by the CAM software and approximating the original shape of the curve. The sampling frequencies with any reasonable number of samples are far above the ability of the machine to reproduce and are only an academic consideration in most cases. They can however result in resonances that can destructively interfere in the proper operation of the steppers.

oldtiffie
04-29-2008, 09:20 AM
No axe to grind, nor any connection, but Syil Australia are showing X4 ready to ship.

http://www.syil.com.au/product_X4Plus.php


They are having that built for them in the north of China as Sieg will not supply them CNC machines.
It very political.

Interesting in that they are still using the old SX3 head.

.

Thanks for posting that link Lin - very interesting indeed.

That was a very interesting post John - very interesting indeed.

I had problems getting info re pricing, availability, support, model availability and features etc. from the OZ distributor some time ago - last 12 months.

I let it "drop" as I just had a "bad feeling" about it - nothing I could put my finger on.

Putting the "politics" issue aside, its seems that we in OZ may not be getting the full latest version.

What is the difference between the X3 and X4 heads? If the difference is significant and if we can't get the latest model - at Seig quality standards - as is in the rest of the world - then I'll just "lose interest".

I have read the link provided by Swarf&Sparks (Lin)) at:
http://www.syil.com.au/product_X4Plus.php

I wonder how much of the computing/CNC stuff is included and if it is an "all in and set to work on your site" arrangement.

If it is "all in" then it is within our budget at AU$6,450 (~ US$6,000) and seems a good buy. But reading the specs there are not included. If it is a reasonable "extra" we might just "make it" - but I need to be convinced as it could go over AU$7,000.

I will email them and see what "gives".

That said, I'd still like to know the features of the X4 and its cost and availability before I even begin to consider it seriously.

Any help/advice in due course will be appreciated.

I did notice that the Seig X4 has
Solid concrete polymer on a steel base, precision guides

Lazlo, Cameron and some others should be interested and heartened by that.

Swarf&Sparks
04-29-2008, 09:27 AM
As you say, Mick, they are starting to approach Tormach pricing.
At least at your end of the continent.
http://www.tormach.com/Product_PCNC_main.html

Evan
04-29-2008, 09:27 AM
I'll have a chance to inspect one soon. A good friend of mine has one on order and should be picking it up right about now in Calgary. He asked me what he should buy and I recommended the X4.

toastydeath
04-29-2008, 09:41 AM
Again, some caveats: All my answers are not ethical and I realize this. If someone has philosophical or moral issues with my decisions, the only solace I can offer is to thumb my nose and go *pthhhhbt*.


1.
what is the best "bang for buck" CAD system/software?:

I feel Solidworks is the best of class for quick designs and drawings. While not as powerful as some other packages, if you notice, I will be shocked. Perfect for a home shop machinist.

I pirated my copy. Oh Noes!



2.
what is the best "bang for buck" post-processor?:

By post-processor, I guess you mean "CAM system." I usually program things at the tool, but in the event of something complicated, I pirated Mastercam X2. Oh noes again!



3.
what is the best "bang for buck" CNC system/software?:


The best bang for the buck is the old 80's VMC you got a steal on, which shouldn't be too hard to do at auctions. That is, of course, reliant on going out to find that bargain rather than buying something easily online.

This recommendation relies on many environmental and economic factors of course, and it may not apply to people who only have the space for a desktop mill. I still stand by my recommendation.



4.
and if the answer to (3) is not Mach3, why isn't it?

Even a VMC from the 80's had the hardware and control to cut at 200 IPM. As someone once mentioned when I brought this up last time, you do take a risk in hardware might blow and availability can be a concern, so obviously that needs to weigh in. But if you get the model of control and find a supplier of parts, in the UNLIKELY situation there's a problem, you will be covered. Fanuc and Yasnac parts are widely available to early revisions since they are so common, but it's good to check availability beforehand.

An old VMC from a good make will have horsepower, amazing rigidity, better accuracy (yes, even worn out), a bloody freakin' toolchanger, and what I consider more important than anything else, a good user interface. I've dorked around with a couple different pc-based systems and found all of them to be mediocre at best compared to the old monochromatic green Fanuc controls I have become familiar with.

Maybe this is just my inexperienced opinion with the retrofit controls, but they seem more focused on eye candy rather than getting actual work done. The differences in how the control handles the little things, for me, makes the difference between being frustrated and being able to get work done.

oldtiffie
04-29-2008, 09:43 AM
As you say, Mick, they are starting to approach Tormach pricing.
At least at your end of the continent.
http://www.tormach.com/Product_PCNC_main.html

Thanks for the excellent advice/comment and link Lin.

I had a look at the Tormach site and was surprised that they want a computer with a parallel port on the main board and that a lap-top will not do the job as in:
http://www.tormach.com/document_library/SB0012_Control_Computers.pdf

That is OK with me as I intend to use a desk-top computer with the parallel port on the main board.

The specs at their site were interesting as they are similar to the Seig.

At US$6,800 (~ AU$7,300) it is still dearer than the X4 and has only 3 ports as opposed to the Seig 4.

I don't even know if Tormach is available or supported in OZ.

To date the Seig is the preferred option.

Swarf&Sparks
04-29-2008, 09:45 AM
It may have escaped your attention TD, but Mick is in Oz, that is Australia! (as am I, even more isolated in the west)

"Old iron" is not an option.

Swarf&Sparks
04-29-2008, 09:52 AM
Fair 'nuff Mick.
I was simply pointing out that the chinese CNC mills are starting to get into the US-made price range. (if one believes that the tormach is entirely US made :p )

I defer to sir John's knowledge of small/home shop CNC and invite comparison between the syil/tormach.

oldtiffie
04-29-2008, 10:06 AM
Thanks TD for your post at #75.

Excellent again.

As Swarf&Sparks (Lin) says, I am in OZ - south-east corner. Lin is in Western Australia.

I am closer to New Zealand than I am to where Lin lives - what a worry!!

Also as Lin says, "old iron" is a "no-no" for lots of reasons - availability and cost of machine and delivery even from parts of OZ is horrific. Parts and support are almost non-existent here for all practical purposes.

lazlo
04-29-2008, 10:24 AM
The tutorial is mine.

Evan, we're not stupid. It's blatantly apparent when you cut-and-paste different writing styles, different terminology, and different application domains into a single paragraph. You do this all the time, but that post was particularly obvious.


You don't understand sampling theory. To reconstruct a sine wave sampled at the Nyquist frequency theoretically requires the summing of an infinite series of terms.

No, that's completely wrong. Fourier analysis breaks down a signal into a summation of "basis functions" -- which are sine waves. So if you do Fourier analysis on a sine wave, it has a single basis function, not an infinite series. That's why I picked a pure sine wave for my example, and why it's used to teach sampling theory to undergraduates.


A square wave is the product of an infinite series of sine wave harmonics. To reconstruct the sine wave fundamental from the square wave produced by sampling at the Nyquist frequency a deconvolution must be performed which is the inverse function of constructing a square wave from a series of sine waves.

Yes, that's true -- now you're quoting back to me what I said in my previous post, mixed with random Wikipedia cut-and-pastes about convolution kernels:


An infinite series occurs when you're sampling a complex signal with an infinite number of frequency components. Fourier analysis breaks each frequency component into a separate sine wave, such that the sum of all the sine waves == the original signal.

If you have a mathematical discontinuity, such as a square wave signal, it takes an infinite series of frequency components to represent.


What do you call approximating a circle then? Since pi is irrational you don't have any choice but to approximate it with a series of data points. It doesn't matter if you use splines or anything else, they can all be deconstructed to a series of straight line moves since the machine must be instructed to move in discrete steps.

Yes! Pi being irrational has absolutely nothing to do with interpolating a curve, but yes -- interpolating the unit sine wave in the challenge I presented to you is exactly the same thing as interpolating a unit circle in CNC -- it takes an infinite number of vectors to cut a smooth version of the curve.

In the case of the sine wave, Nyquist says that you can correctly sample the curve with 2 datapoints per cycle. With the case of a circle, which is sine + cosine, you can correctly sample the signal with 4 datapoints.

But as I've said many times in this thread, it will still take thousands of datapoints/vectors ((G01 moves) to interpolate a smooth curve, regardless of the Nyquist frequency of the curve.


they can all be deconstructed to a series of straight line moves since the machine must be instructed to move in discrete steps.

There is a name for that. It's called sampling.

No, that was exactly the issue that Swarf&Sparks pointed out at the beginning of the thread. Reproducing a curve with a series of straight lines is interpolating. It has nothing to do with Nyquist sampling theory.

lazlo
04-29-2008, 10:47 AM
No axe to grind, nor any connection, but Syil Australia are showing X4 ready to ship.

http://www.syil.com.au/product_X4Plus.php

Wow, A$6,450 ($6,014.46 US) is pretty steep. A Tormach is "only" $6,500 and a lot more machine (if you have the space).

By the way, Sir John posted here awhile ago that the pre-packaged X4 CNC machines were going to be re-sold by Smithy, but from reading through this month's HSM magazine, there are several other importers with better prices. It's funny because they all try to put funny shrouds on the head to disguise that it's a Sieg X3 :)

Evan
04-29-2008, 10:49 AM
Evan, we're not stupid. It's blatantly apparent when you cut-and-paste different writing styles, different terminology, and different application domains into a single paragraph. You do this all the time, but that post was particularly obvious.
I can forgive a lot of things Robert, but being called a liar isn't one of them. I always use quotes when I quote something. You just can't accept that I might know more than you do about something.

You will have to explain how I happen to have the high resolution copy of the drawing I posted. It pretty obvious that it isn't upsampled either. I always work in hi res and then downsample for posting.

You sir are an ass.

http://vts.bc.ca/pics3/samples.jpg

lazlo
04-29-2008, 10:56 AM
The specs at their site were interesting as they are similar to the Seig.

Isn't the X4 the converted X3 mill? The Tormach is a much bigger machine (1100 lbs versus 400 for the X3).


At US$6,800 (~ AU$7,300) it is still dearer than the X4 and has only 3 ports as opposed to the Seig 4.

Sorry Mick, I guess the price on the Tormach went up since the last time I looked at it...

oldtiffie
04-29-2008, 11:07 AM
Wow, A$6,450 ($6,014.46 US) is pretty steep. A Tormach is "only" $6,500 and a lot more machine (if you have the space).

By the way, Sir John posted here awhile ago that the pre-packaged X4 CNC machines were going to be re-sold by Smithy, but from reading through this month's HSM magazine, there are several other importers with better prices. It's funny because they all try to put funny shrouds on the head to disguise that it's a Sieg X3 :)

Thanks lazlo.

That's true.

But so far as I am aware, while "Tormach" is a real "go-er" in the US it isn't here in OZ for reasons given previously.

I accept absolutely that Tormach is a great machine - as is Seig - but the practicalities are different here in OZ.

If Tormach were at least equal to all aspects of buying it here in OZ I'd seriously consider buying it instead of the Seig.

I am not wedded to "China-made" but solely on performance, cost, support etc. on an "all-in" basis - nothing more and nothing less.

I daresay I'd go for Tormach if I were in the US - but it would be a close call.

If - repeat IF - I buy a Sieg it will be my last big buy as at 71 I am fast running out of time in my shop. I am a realist in that regard (and many others). But having said that I am not going to rush - or be rushed - into making a decision one way or the other before I'm ready.

Oops.

I just re-read this and I hasten to assure you that I don't think that you said that I am wedded to "China-made" nor that you think that I think you are "rushing" or "pushing" me in any way.

I just thought that in fairness I should make those points to you.

Its nice to have these machines and similar available.

lazlo
04-29-2008, 11:20 AM
I accept absolutely that Tormach is a great machine - as is Seig - but the practicalities are different here in OZ.

If Tormach were at least equal to all aspects of buying it here in OZ I'd seriously consider buying it instead of the Seig.

I am not wedded to "China-made" but solely on performance, cost, support etc. on an "all-in" basis - nothing more and nothing less.

Actually, the Tormach is "China-made" too :) The machine is made in China, and I read on CNCZone that it uses Chinese copies of the Gecko 201 stepper controllers (much to Marris' chagrin).

I'm surprised there's not a Tormach distributor in Australia -- shipping would surely be cheaper for you.

Swarf&Sparks
04-29-2008, 12:22 PM
Lazlo, I emailed tormach some time ago (12-18 months) re a demo machine in OZ somewhere.

The best they could come up with was in South Africa!

mochinist
04-29-2008, 02:49 PM
Amazing that you two have turned yet another thread into a who is smarter than who thread. Maybe exchange emails or sign up for yahoo messenger and blast each other away.

Swarf&Sparks
04-29-2008, 03:00 PM
"you two"
who two?
I think a lot of people are enjoying and learning from this thread.

OK, maybe I am the only people learning from this thread. :eek:

mochinist
04-29-2008, 04:16 PM
"you two"
who two?
I think a lot of people are enjoying and learning from this thread.

OK, maybe I am the only people learning from this thread. :eek:Evan and Lazlo, other than that it would be an enjoyable thread.

Swarf&Sparks
04-29-2008, 04:25 PM
Nah, lazlo's ok, he's on my side this time :D
(biased, me?)

mochinist
04-29-2008, 04:33 PM
Nah, lazlo's ok, he's on my side this time :D
(biased, me?)Evan and Lazlo are both fine and obviously smart guys, I just don't see how that discussion (now with large high res pics) is helping anyone here, homeshop or pro machinist. But hey I guess its just my opinion, hell keep at it maybe we can get another locked thread.:cool:

Swarf&Sparks
04-29-2008, 04:37 PM
Well, I learnt something.
Or several somethings.
Can't be all bad.
:p

tattoomike68
04-29-2008, 05:52 PM
I have autocad 2007. I never use it. I have a photographic memory so its a waste of time to make a detailed print.

Most everything I make I just scribble down some numbers and go with it.

machine work is all for fun now days so I can change anything and fling the prints in the trash anyway..

John Stevenson
04-29-2008, 06:51 PM
I'll try to clarify between the Sieg, the Syil X4 and the Tormach as far as I know.

Sieg started off making the small lathes and mills we are familar with X1, X2, X3 etc. They buy in their main casting from various foundries and machine up in house.

Othe companies also buy from the same foundries and copy, there isn't usually much difference in design but often a lot in quality.

Sieg has two arms, the export arm who indirectly we in the west all deal with and it's domestic arm who sell internally in China.

Syil started out buying manual machines from Sieg and converting them to CNC. Sieg were not happy with this as they blatently advertised them as Sieg and even said they were the CNC arm of Sieg which didn't go down a bundle.

The problem is the domestic arm has to sell to anyone internally because of politics.

Sieg has now come out with their own range of dedicated CNC's and refuses to sell to Syil because of their problems with support issues so Syil has approached a maker in Northern China to make the X4 for them using freely available parts from the copy manufacturers hence the remark about the old SX3 head which still has motor problems.

I can't comment on the X4 as I have not seen one or tried one, but the factory offered me one if I wanted it.
I have had one of the Syil conversions and I wasn't impressed with it's build or quality and this is not sour grapes.

I have seen and used a Tormach machine, in fact I installed the first pre - production model into the UK.
These are really impressive and are far bigger than they look, they are a desktop machine but you need a blooody good desk as they weight 1/2 a ton :D

Designed from the ground up as a CNC with pressure oiling, turcite slides etc they are very good value for the money and streets in front of the Sieg or Syil.

The drives are a copy of the Gecko but made in the US. I don't know the full story behind this but it wasn't a straight rip off and had something to do with an earlier drive that Marris and others worked on. Marris wasn't able to do anything about it so surfice to say they it must have been a grey area.

If you have the space and can afford it buy a Tormach, if you don't need a machine this size then pick one of the Sieg machines.

BTW both the KX1 and the KX3 both come with the 4th axis drivers and plug fitted as standard to accept any 4th axis with a 3 amp motor and XLR plug.

.

lazlo
04-29-2008, 06:57 PM
Syil has approached a maker in Northern China to make the X4 for them using freely available parts from the copy manufacturers hence the remark about the old SX3 head which still has motor problems.

BTW both the KX1 and the KX3 both come with the 4th axis drivers and plug fitted as standard to accept any 4th axis with a 3 amp motor and XLR plug.

Fascinating story John!

Just to clarify, the Syil X4 is the same basic machine as the X3 and the KX3, correct?

John Stevenson
04-29-2008, 07:16 PM
Robert,
I honestly don't know.
The head is off a SX3 but they advertise concrete polymer base and precision slides? whatever they are, linear rails ?

.

oldtiffie
04-29-2008, 07:24 PM
Thanks Lin, John and lazlo.

That lot has cooled my enthusiasm considerably.

I think I will just "sit it out" for a while yet as I am in no hurry and won't be hurried.

I retain my interest in "things CNC" never-the less in the meantime.

So it seems that for the time being at least in OZ:
- "Tormach" is a "non-starter" for reasons given by Lin;
- Seig is or may not be available in OZ;
- Seig will not sell to Syil;
- the Syil is an inferior copy of the later model than the Seig;
- the Seig/Cyil X3 head still has "motor" problems;
- Seig will not sell to Cyil due to inadequate support (in OZ).

A very murky picture that is most unsatisfactory to me.

Seems my "gut feeling" might have been right and might still be.

I will "step back" for a while until matters of concern are clarified.

I will check to see if there are real agents for Seig in OZ.

Any advice from members in NZ?

lazlo
04-29-2008, 07:30 PM
The head is off a SX3 but they advertise concrete polymer base and precision slides? whatever they are, linear rails ?

Barely any documentation on their web page, but like you say, it looks like they poured the base for linear rails, and used the rest of the X3 chassis/head.

Syil CNC X3 on the left, and the X4 on the right.

http://www.syil.com.au/images/X3/RgtSide.jpghttp://i164.photobucket.com/albums/u15/rtgeorge_album/x41.jpg

Edit: I notice that they took off the manual drill-press feed and the fine quill feed, so the X4 is a CNC-only machine, although the CNC X3 doesn't have handwheels on the table...

lazlo
04-29-2008, 07:31 PM
Any advice from members in NZ?

I'm not in NZ :), but you could buy an X3 locally, and buy a CNC retrofit kit from Syil, or John, or someone locally...

Swarf&Sparks
04-29-2008, 07:32 PM
You could check Hare & Forbes aka Herless machinery, Mick.
They do stock and support the X3.

Perhaps a certain pommie engineer would be interested in supplying CNC conversion?

nheng
04-29-2008, 07:44 PM
I think it's obvious that Evan used floating point wavelet decompression to enlarge his original image to the fine one shown here :D :D

When I think Nyquist and Shannon (in the real world), I generally think of a very finite number of sample points. If you were machining one cycle of a sinusoidal shape, only 2 points would be required to hit it for adequate (re)construction. But we all know that unless you keep hitting that sinusoid at different sets of 2 points each, it's gonna be pretty lumpy looking. From some years back, I recall that the standard "milliwatt" test tone used for telephone line and channel testing was 1004Hz. The reason was that the sampling rate was only 8kHz for a 4kHz voice channel. If you used 1000Hz (instead of the 4 hertz delta that "smoothed" the result) to test with, the 1000Hz would end up slowly beating with the carrier. The end result was that you would watch the analog level measurement slowly rise and fall on the (then) analog meters.

Here's some further food for thought for Evan and Lazlo: For a purely Gaussian response, the rise time of a typical system (like an oscilloscope) is approximated by:

Tr = 0.35 / Bandwidth

Now you guys can go argue over what bandwidth is required to CNC mill a pattern with "rises" or "falls" in the tool travel so that the steepest rise or fall is adequately represented :D :D

You guys who are interested in the interpolation aspect might also check out Bresenham's line and circle drawing (and I think curves too) algorithms. He was an old (IBM?) guy who had the task of coming up with an efficient plotting method for pen plotting. He did that and more. In the years since other alternate means of drawing have been devised but his methods might be of interest to those controlling steppers or servos and looking for the simplest path.

lazlo
04-29-2008, 08:02 PM
Funny that you mention Bresenham's algorithm Den -- many moons ago I wrote a vector graphics engine which implemented Bresenham in assembly code on an ancient 68000 VME board (this was before there were any VME graphics cards).

But Bresenham's algorithm just linearly interpolates a vector between two points, which I assume MACH is already doing in the G01 move. To interpolate a curve like we're discussing, you either have to decompose it into hundreds (or thousands) of G01's, or do a spline move, which would require that both axis' move at the same time.

oldtiffie
04-29-2008, 08:20 PM
Any advice from members in NZ?


I'm not in NZ :), but you could buy an X3 locally, and buy a CNC retrofit kit from Syil, or John, or someone locally...



You could check Hare & Forbes aka Herless machinery, Mick.
They do stock and support the X3.

Perhaps a certain pommie engineer would be interested in supplying CNC conversion?

Well I'll be damned!!!

Thanks lazlo and Lin.

I thought I knew the HAFCO catalogues backward - they seem to be my shopping list!!!

It sure does look like HAFCO have the X3 as Lin states.
http://www.machineryhouse.com.au/Products?stockCode=M155

http://www.machineryhouse.com.au/Mill-Drill-Machines

I reckon that if I've stood in front of that bl**dy machine once, I've done it a dozen times. I even tried it casually. I wish I'd paid more attention. I might just phone them or go and see it again.

But given John Stevenson's advice about the lesser quality/copy that Cyil is of the genuine Seig, I'd guess and just about be sure that it is the "cheaper" version/copy. But as said, I will inquire further.

If that is the case it sure does put a whopping premium on the Cyil CNC (retro?) fit - AU$4,000 plus!!

Lin, the
Perhaps a certain pommie engineer would be interested in supplying CNC conversion? would be John Stevenson I presume.

I think I'll stay out of the market here for a while yet.

Evan
04-29-2008, 08:21 PM
Now you guys can go argue over what bandwidth is required to CNC mill a pattern with "rises" or "falls" in the tool travel so that the steepest rise or fall is adequately represented

That's an interesting question. Mach 3 for instance dynamically allocates bandwidth to the axis that requires it. I was doing a picture today by the raster scan method. When a sudden change in Z is required the x axis is slowed sufficiently to allow the Z to make a reasonable approximation of a square wave. That's the low pass filter effect I wrote about. It doesn't stop the X axis entirely, only slows it. How much the approximation differs from a true square wave depends on the cutting speed. However, you can select absolute stop mode in which the x axis will be absolutely stopped and an exact approximation of a square wave will be cut.

That's a brick wall filter which allows only the frequencies <= the sample period to be expressed. It still doesn't make it possible to correctly reconstruct a curve from samples at a rate equal to the Nyquist frequency. It only prevents aliasing, which is what I said in the first place and several times since. I have also made it clear that I am speaking about reconstruction of samples as illustrated in my drawing.

I need to build a special tape drive to recover some of my early software I wrote. In particular I have a complete control system for a plotter that does linear interpolation in 2 axes and also controls variable reluctance steppers with full S curve acceleration profiles. I wrote that software in the 80's.

Swarf&Sparks
04-29-2008, 08:23 PM
I'll be out of the market meself, a while.
I've got an 8 x 4 router to build.
(and limited wherewithal)

jacampb2
04-29-2008, 08:56 PM
Well, I finally had enough time to sit down and read through the remaining pages of my thread. It started to run away on page 3 or so, and I didn't have time to read it all, so I just took the advice proffered in the first couple of pages and looked at that software.

I found (find) the ongoing debate interesting, although for the most part beyond my understanding. I have to say, by the end of it, I was starting to understand what you guys are trying to say.

Anyhow, tiffe, I much appreciate your trying to keep the discussion on track, and I am glad there were others interested in this subject as well. I will continue to read on for whatever new developments arise :D

Thanks,
Jason

oldtiffie
04-29-2008, 09:14 PM
Well, I finally had enough time to sit down and read through the remaining pages of my thread. It started to run away on page 3 or so, and I didn't have time to read it all, so I just took the advice proffered in the first couple of pages and looked at that software.

I found (find) the ongoing debate interesting, although for the most part beyond my understanding. I have to say, by the end of it, I was starting to understand what you guys are trying to say.

Anyhow, tiffe, I much appreciate your trying to keep the discussion on track, and I am glad there were others interested in this subject as well. I will continue to read on for whatever new developments arise :D

Thanks,
Jason

Thanks Jason.

Yes I am trying to keep the discussion "on-track" as you say. Its not difficult as there are 2 concurrent "streams" in this thread. Both are closely related.

The "lazlo and Evan" posts are a bit ethereal I suppose but need to be understood (sort of) to appreciate what the CAD and CAM software is required to do. If in any doubt, just ask them to clear a point or points up with and for you - I am sure they will. I, like Swarf&Sparks (Lin in OZ) have learned a lot from that discussion too.

In the end I hope we know what compromises and limitations there are that we need to address in getting a CAD drawing processed so that the job/work-piece approximates the CAD model within acceptable limits as regards size, finish and "all-in" cost as I expect that they will be the parameters/limits within we must operate in CNC.

There are a lot of "machine/practical" issues that have to be addressed as well - so there's plenty of "(s)mileage" in that aspect as well yet.

So, keep with us, keep posting, and tell us if we are going too far off-track, as the sooner (and more often??) that's done the better the thread performs.

jacampb2
04-29-2008, 09:27 PM
I think that I should mention, I am fairly happy with Acad for drawings. I will probably see if I can get a trial of solidworks again and see how it is for 3d modeling, but I don't have any immediate need to switch from Acad. The reason behind the title, of cad/cam is I know that some cam programs include their own CAD software, anywhere from robust to rudimentary. If there was an all-in-one sort of deal that people are making good use of, I would happily consider it.

It would be cool if there was a cam program that could read google sketchup. I have to say, that for a 3d design/cad (can it be called cad?) program, I found it chocked full of features and very intuitive. I was up and drawing in the free version inside 2 hours. I know that sketchup pro can export .DFX, but you are right back to a $500 software price tag for that gem. Great for professionals who are going to make that back in one job, not so great for the HSM guy who will be lucky if his work pays for his raw materials...

Anyhow, keep it coming.

Later,
Jason

J Tiers
04-29-2008, 11:29 PM
When I think Nyquist and Shannon (in the real world), I generally think of a very finite number of sample points. If you were machining one cycle of a sinusoidal shape, only 2 points would be required to hit it for adequate (re)construction. But we all know that unless you keep hitting that sinusoid at different sets of 2 points each, it's gonna be pretty lumpy looking.


What you have done is to describe the lack of amplitude data in a precise 2x sampling system.

Recall that the sampling must be FASTER than 2x, i.e. depending on ther rest of the system, it may be 2.01 x or whatever, but exactly 2x does not work.

Then also, the mere placement of a 'filter" at 0.49 x the sample frequency isn't good enough. You must not have significant signal at over that frequency, or you will alias on the "leakage".

In the case of the CAM system, the jaggies may not be sampling problems from input data, but a simple lack of resolution in output positioning. That will "re-quantize" the data, and may be looked at as somewhat equivalent.

lazlo
04-29-2008, 11:50 PM
In the case of the CAM system, the jaggies may not be sampling problems from input data, but a simple lack of resolution in output positioning.

Thank you Jerry. As you (and Swarf&Sparks) know, that's what I've been saying for 6 pages :)

Swarf&Sparks
04-30-2008, 12:10 AM
Fair call Rob

Mind you, in 1/8 step, the little router has a resolution of 0.0002"
backlash, what backlash?
Just wish it was stiff enough to repeat to anywhere near 0.1 MM :D

(whaddya mean, Young's modulus?)

nheng
04-30-2008, 12:33 AM
J.Tiers said "Then also, the mere placement of a 'filter" at 0.49 x the sample frequency isn't good enough. You must not have significant signal at over that frequency, or you will alias on the "leakage"."

Very true. This is one of the dangers of digital oscilloscopes with inexperienced users. Due to less than a brick wall response, you can easily alias back into the first Nyquist zone. I've been fooled occasionally by a near perfect sine wave only to realize (due to less than stable triggering on it) that it was 10s of times below the actual signal.

Swarf&Sparks
04-30-2008, 01:05 AM
I guess I'm lucky then.
I can't afford a DSO, I'm stuck with my old 10Mhz analogue :D

I wish I could be unlucky and have a 10Gs/s digital.

Evan
04-30-2008, 01:54 AM
In the case of the CAM system, the jaggies may not be sampling problems from input data, but a simple lack of resolution in output positioning.
That "simple lack of resolution" is producing mechanical aliasing. That is a perfect example of the reproduction frequency not being high enough to accurately reproduce the input. If the mechanical resolution is less than the data points resolution the aliasing occurs.

The limiting resolution of a device is the spatial equivalent of it's Nyquist frequency and sampling theory applies equally to it as it does to any other signal reconstruction.

In particular, the Shannon-Whittaker Interpolation formula describes these limits. It's a part of sampling theory.

Ken_Shea
04-30-2008, 02:31 AM
All this Nyquist , aliasing, reproduction frequency, Bresenham's algorithm, while impressive, who cares, over my head, interest and 99% of everyone else's as well, that is what a good cam program is for ;)

No jaggies on my system.

Swarf&Sparks
04-30-2008, 02:32 AM
with trumpeters who improvise
a full octave higher than the score

John Stevenson
04-30-2008, 03:35 AM
All this Nyquist , aliasing, reproduction frequency, Bresenham's algorithm, while impressive, who cares, over my head, interest and 99% of everyone else's as well, that is what a good cam program is for ;)

No jaggies on my system.

LOL,
Never a truer word. There must be billions of drivers out there who don't have a clue how a car works.

.

Ken_Shea
04-30-2008, 03:41 AM
LOL,
Never a truer word. There must be billions of drivers out there who don't have a clue how a car works.

.

:D
Having been in the auto repair business for over 35 years that statement is so true, truer then ever on todays mechanice I.E. parts replacers :D

Evan
04-30-2008, 05:20 AM
Never a truer word. There must be billions of drivers out there who don't have a clue how a car works.

Our insurance rates are proof of that.

oldtiffie
04-30-2008, 07:39 AM
Further to my post #103 at:
http://bbs.homeshopmachinist.net/showpost.php?p=353670&postcount=103

Wife and I went shopping today, had a nice lunch and we went to my main Machine/tools supplier (Hare and Forbes aka HAFCO) to see the so-called Seig-type "Super X3" non-CNC-ed mill.

For details see (as previously posted):

http://www.machineryhouse.com.au/Products?stockCode=M155

http://www.machineryhouse.com.au/Mill-Drill-Machines

I must say that I got quite pleasant surprise. While it had been re-badged as a HAFCO product, the master catalogue showed to be sourced from Seig. Perhaps Seig have varying levels of quality and perhaps this was not the top of the heap, but Seig it is and there are ample Seig spares.

I asked about spare motors as John Stevenson suggested a problem there. HAFCO said that the problem they have is the PCB in the head failing at about 5 months. But they have improved (no failures yet) PCB's in stock and are replaced on warranty, no questions asked.

I have to say that machine is as good as I could wish for. All ways are fine-ground. There is negligible lead-screw back-lash and it is as smooth as silk.

Very impressed.

HAFCO don't do the CNC fit and set-up but they did recommend one guy who is apparently excellent and uses all USA-made/sourced CNC parts. The set-up and set-to-work seems expensive from the "top of the head" advice. But I will wait until HAFCO emails me that chap's web and email addresses and I will see if and where I go from there.

I already have a "legit" and paid for and registered full copy of Mach3 and will also provide the computer.

I will keep the thread/forum posted.

I'd like the Seig X4 but if the X3 is all that I can get on good terms (cost, support, function etc), then I may "go for it".

This X3 has an MT3 taper which suits all my tooling for my lathe and HF-45 VCM.

Links to pics and comment follows.

This is the over-view from front right:
http://i200.photobucket.com/albums/aa294/oldtiffie/Seig_X3_mill/Seig_X3_1.jpg

This is the milling head from front right:
http://i200.photobucket.com/albums/aa294/oldtiffie/Seig_X3_mill/Seig_X3_2.jpg

This is the over-view from front-left:
http://i200.photobucket.com/albums/aa294/oldtiffie/Seig_X3_mill/Seig_X3_4.jpg

This is the capacity/capability plate:
http://i200.photobucket.com/albums/aa294/oldtiffie/Seig_X3_mill/Seig_X3_5.jpg

This is the front face of the milling head:
http://i200.photobucket.com/albums/aa294/oldtiffie/Seig_X3_mill/Seig_X3_6.jpg

The speed control is excellent. Speed set is registered in the green digital display at the top. The tactile controls are self-evident.

The quill is fitted with a DRO (converted height guage?).

The quill is excellent as regards fit, movement and clamping. The "coarse" ("drill") hand-feed is very good. The quill fine feed is excellent as well.

The "X" and "Y" movement is remarkably good.

Having a HF-45 VCM made me check the vertical ("Z") column feed. I need not have worried as it is as good as my surface and T&C grinder up/down feeds.

All-in-all, a very nice machine for the price - so far.

Even if I do "go ahead" with the purchase, I will still keep and use my HF-45 VCM as that machine has and can do some very good work. A lot of my gear will not fit on the "Seig" as it was bought for the HF-45. A lot of my smaller (grinder-sized) stuff will fit on the Seig easily.

I need a lot more advice and time yet to make an assessment as to whether it is OK to "go" or not.

I am much happier so far with this option that I was and am with Cyil.

Swarf&Sparks
04-30-2008, 07:55 AM
Garn Mick, grow old disgracefully.
Part with that pocketful of fivers you've been saving since Woomera.

It's a steep curve, but you will love CNC.

J Tiers
04-30-2008, 08:15 AM
That "simple lack of resolution" is producing mechanical aliasing. That is a perfect example of the reproduction frequency not being high enough to accurately reproduce the input. If the mechanical resolution is less than the data points resolution the aliasing occurs.


Evan

Whether or not this is interesting to Ken Shea..............

Please to quote entire statement, and not the first half only. That may allow you to re-state the second half and imagine that you are making a new statement,, but it isn't the same thing as correcting someone's lack of understanding..............


In the case of the CAM system, the jaggies may not be sampling problems from input data, but a simple lack of resolution in output positioning. That will "re-quantize" the data, and may be looked at as somewhat equivalent.

oldtiffie
04-30-2008, 08:22 AM
Garn Mick, grow old disgracefully.
Part with that pocketful of fivers you've been saving since Woomera.

It's a steep curve, but you will love CNC.

Yeah - you must be laughing your cock off at this tangle I'm in!!

But seriously, many thanks for the "pointer" and "heads up".

I have "E-machineshop" CAD as recommended by Steve (thanks Steve) its a "goodie" - no dimensions, just gets on with it and does the job in 3-D - and its free.

I have AutoCad 2004 but have just about forgotten how to use it!!.

I bought the IMSI "DesignCAD 3D Max" but have yet to install it and try it out. Reviews and published stuff looks good. I will get around to it.

I had to smile at your comment:

Garn Mick, grow old disgracefully.
Part with that pocketful of fivers you've been saving since Woomera.

I took the Missus out shopping and for a very nice fish lunch in a big Mall near here. I told her (afterward) that I had to go to HAFCO and that she could thank/blame you. She reckoned that it was all great - so you've done well, I've acted disgracefully (no regrets - at all!!!).

She not only thinks you are a nice bloke (yep she go that wrong too) but that if I wanted the machine to go and get it as she said we'd "get by".

You are a persuasive bast*rd aren't you?

And - as expected - you're not too ashamed either!! But I won't spoil it for her by telling her - no way - even I am not THAT blo*dy silly or stupid!!!

But seriously - we both had a very good day - thanks.

I don't know if my conscience is clear or not as I either lost it or gave it away more years ago than I want to remember!!.

Swarf&Sparks
04-30-2008, 08:30 AM
"conscience"

As long as you got the box it came in :D


Non, je ne regrette rien (a la Piaf)

gn3dr
04-30-2008, 08:36 AM
Barely any documentation on their web page, but like you say, it looks like they poured the base for linear rails, and used the rest of the X3 chassis/head.

Syil CNC X3 on the left, and the X4 on the right.

http://www.syil.com.au/images/X3/RgtSide.jpghttp://i164.photobucket.com/albums/u15/rtgeorge_album/x41.jpg

Edit: I notice that they took off the manual drill-press feed and the fine quill feed, so the X4 is a CNC-only machine, although the CNC X3 doesn't have handwheels on the table...

The X4+ has a port for an MPG and comes with an MPG so you can use it manually from this with the handwheel.

Swarf&Sparks
04-30-2008, 08:44 AM
If you believe the pretty pictures, it comes with a 4th axis too.

oldtiffie
04-30-2008, 08:45 AM
Thanks gn3dr.

Appreciated.

What is an MPG? Its probably a silly question and I will probably have a head-slapping session when I see your reply.

Swarf&Sparks
04-30-2008, 08:49 AM
Manual pulse generator
sorta elctronic handwheel Mick

gn3dr
04-30-2008, 08:54 AM
Thanks gn3dr.

Appreciated.

What is an MPG? Its probably a silly question and I will probably have a head-slapping session when I see your reply.

MPG is a manual pulse generator. Here's a picture of the X4 one

http://www.cnczone.com/forums/attachment.php?attachmentid=57775&stc=1&thumb=1&d=1208279584

Bigger pictures in the thread

http://www.cnczone.com/forums/showthread.php?t=55999&page=2

So you just switch to whichever axis you want - choose X1 for 0.0mm moves, X10 for 0.1mm moves or X100 for 1mm moves and then just turn the handwheel to move your axis. The X4 also has all the same buttons on the head.


I just bought an X4+ and while Ihaven't done anything with it yet I am happy with the overall quality of it. I went through all teh same puschasing decisions as you - do I buy an X3 and a conversion kit etc. In the end I choose the X4+ because it has some nice features - especially the MPG, compared to any of the conversions out there.

Also I decided I'd rather learn cnc with a ready made machine instead of spending time building a machine.

oldtiffie
04-30-2008, 09:04 AM
"conscience"

As long as you got the box it came in :D


Non, je ne regrette rien (a la Piaf)

Not "jig-a-jigging" likely!! When I got rid of it it stayed got rid of!!!

If I was too concerned (I wasn't) about what I stuck in what box in my youth I'd have died of worry, embarrassment or from that nagging conscience many times over many years ago!!!

lazlo
04-30-2008, 09:07 AM
The X4+ has a port for an MPG and comes with an MPG so you can use it manually from this with the handwheel.

Nice touch. It's hard to tell for sure from the pictures, but it looks like it has linear guides on both the X- and Y-axis?

Is there a quill feed, or is the Z-axis feed entirely from moving the head?

gn3dr
04-30-2008, 09:10 AM
Nice touch. It's hard to tell for sure from the pictures, but it looks like it has linear guides on both the X- and Y-axis?

Is there a quill feed, or is the Z-axis feed entirely from moving the head?

No linear guides - just standard dovetail type ways. It does have a one-shot oiling system built in though.

No quill feed - I don't see the need for one on a cnc system anyway.

oldtiffie
04-30-2008, 09:17 AM
Thanks gn3dr and Lin re the MPG.

I am pretty well stuck with getting a 3X and having it retro-fitted for CNC as I am not happy with the sole supplier (so far as I know - must check) in OZ for reasons previously given/posted.

So, unless there is another Seig dealer with the right credentials here in OZ I'm stuck with it - if I go ahead with it.

I only ever intended to use 2 ports mostly ("X" and "Y") and reserve the 3rd. port for my rotab or manual control as per MPG. I'd be quite happy to stop the CNC process to feed "Z" manually if needs be.

I am only just starting out on this journey which hopefully will come to a good conclusion but I will have no hesitation or regrets in terminating it mid-way if necessary.

lazlo
04-30-2008, 09:20 AM
In the case of the CAM system, the jaggies may not be sampling problems from input data, but a simple lack of resolution in output positioning.
Fair call Rob

Mind you, in 1/8 step, the little router has a resolution of 0.0002"

...and that's exactly what this thread should have asking: how well does 2.5D interpolate 3D compared with a true 3D system (i.e. commercial systems)?

I dropped an email to Art asking why he doesn't do simultaneous 3D control. PC's have the processing power, so I'm guessing the kinematics are pretty tough to solve.

Swarf&Sparks
04-30-2008, 09:21 AM
C'mon Mick
6 months on CNC and you'll be building your own Ikara (At least)
:p

lazlo
04-30-2008, 09:25 AM
No linear guides - just standard dovetail type ways.

Ah, that's very interesting. As John mentioned, the Syil page says "precision guides" so we were assuming linear guides:

"Solid concrete polymer on a steel base, precision guides"

So it seems like 80% of the standard X3 chassis, but with a custom polymer concrete base (nice!) and a CNC-optimized head (no quill feed).

That sounds like a pretty nice setup!

Swarf&Sparks
04-30-2008, 09:28 AM
I'd rather pour the epoxy and sand at this end.
Save freight :cool:

oldtiffie
04-30-2008, 09:47 AM
I notice at :
http://www.cnczone.com/forums/showpost.php?p=439752&postcount=19
that there is a "Manual" option/switch on the X4.

I do like the X4 but it seems that it might be out of my reach. So the X3 for me it is.

Lazlo, if the X4 ways are as good as the X3 I saw today, there should be no problems. "X" and "Y" dove-tails have parallel gibs with (I think) 4 adjusting screws which I prefer to tapered gibs. See my post at:
http://bbs.homeshopmachinist.net/showpost.php?p=353762&postcount=23

I just hope that there is a bottom not too far away in this quick-sand that I'm marking time in.

oldtiffie
04-30-2008, 10:49 AM
I fell that after all this trouble that I should at least re-visit the Cyil X4.

The Cyil web pages are at:
http://www.syil.com.au/product_X4Plus.php

http://www.syil.com.au/Products.php

This is a re-post/copy of my post at:
http://bbs.homeshopmachinist.net/showpost.php?p=353809&postcount=6



(from seperate thread)
Previous heading: Perhaps, maybe

-----------------------------------------------------------------

I have just re-visited the Syil web-site on OZ.
http://www.syil.com.au/product_X4Plus.php

http://www.syil.com.au/Products.php

I feel that I have to re-examine this option.

Setting aside then name (Syil) is this a re-badged or copy of the Seig X4 or not?

If it is a copy, is it as good as the Seig?

Price at AU$6,450 (Australia/OZ) ~ US$6,000 ~ UKP2,600

Any comments will be appreciated.


As stated - any comment or advice will be very much appreciated.

lazlo
04-30-2008, 10:57 AM
Hi Tiffie,


I notice at :
http://www.cnczone.com/forums/showpost.php?p=439752&postcount=19
that there is a "Manual" option/switch on the X4.

Cool -- long thread. That'll be a good time-waster while I'm feeding the baby :)


Lazlo, if the X4 ways are as good as the X3 I saw today, there should be no problems. "X" and "Y" dove-tails have parallel gibs with (I think) 4 adjusting screws which I prefer to tapered gibs.

I think it all depends on how good a job Syil did on the base. The "cutouts" on the front of the epoxy-granite casting sure look like box ways, but gn3dr says they're dovetails. If Syil did a good job on the ways, then it's probably a very nice machine. On the other hand, if Syil is buying the square Z-column and head from Sieg, and casting the base to save shipping weight (or maybe Sieg doesn't want to sell them the base re: John's explanation), then it could be a liability.

In any event, the X3 is a neat little machine. It's small enough to fit on a sturdy bench, but big enough, and rigid enough, to cut steel. The vast majority of home-built CNC's only cut wood and aluminum, because they're not rigid enough for steel.

So with a CNC'd X3, you can pretty much make any shaped steel or cast iron object that you could possible want in a model engineering setting.

My sabbatical (2 months vacation every 7 years) is coming up -- I may buy an X3 and CNC it for fun.

By the way, I hate pinned gibs -- they're so hard to get adjusted correctly, and really hard to keep them in place without Loctite. Why don't you like tapered gibs? They're so much easier to adjust.

oldtiffie
04-30-2008, 11:16 AM
Thanks lazlo - good response.

I really am in a dilemna now!!!

I will wait and see what the quote is for retro-fitting the X3 as to be honest I fancy the seperate head and quill down-feeds on the X3. It is sturdy enough and I have the manual "drill/mill" option.

I will take John's advise re the MPG at:
http://bbs.homeshopmachinist.net/showpost.php?p=353812&postcount=7


If anyone want a MPG then just buy a Contour Design Shuttle for about $40 - $50 and plug the USB plug into your computer and Mach already has a plugin to accept it.

Lets you jog 4 axis in steps of 1.000, 0.100, 0.010 and 0.001 units.




By the way, I hate pinned gibs -- they're so hard to get adjusted correctly, and really hard to keep them in place without Loctite. Why don't you like tapered gibs? They're so much easier to adjust.

I don't really mind them but there is too many doubts. Some have so much end-clearance with/on the adjusting screw that they "float/oscillate" axially as the "moving" part of the dove-tail operates.

See this - extracted from my:
http://bbs.homeshopmachinist.net/showpost.php?p=353762&postcount=23


There is a potential or real problem with machine gibs - particularly on a mill.

If the lateral thrust is taken on the side of the dove-tail that is "solid" ie the one without the gib, the results will be far better than if the lateral thrust was on the side with the gib and so on the gib itself.

First of all its a big ask to get a tapered gib to be a perfect fit so as the engage the "fixed" side of the dove-tail and the "moving side" of the dove-tail. Its all but impossible after some use - because wear it will. Just tightening it is no guarantee that it will bear evenly along its length. If it is worn at the ends the centre part will be "proud" and so the larger end clearances may well allow the table to "rock" laterally/side-ways. If the wear is in the centre part of the gib - perhaps due to excessive injudicious use of the clamps on the gib - the ends will be "proud" and may cause excessive "wedging/jambing" action in the dove-tail ways due to the 60 degree wedge on dove-tails.

If the gib is a "flat" type that is adjusted by several individual screws the the gib will probably wear at the position of the adjusting screws. Any lateral load transmitted to the gib will be via the relatively small screws which must transmit that load across the dove-tail to the "fixed part".

It is potentially worse if the gib is unknowingly bent into the form of a stiff curved spring.

Gibs and their problems just have to be "lived with" but compensating for errors in the unclamped state as is the case when under CNC control can be a real problem. This is less a problem when milling in non-CNC mode as the dove-tails and gibs are clamped in the non-milling way ("Y" when milling in "X").

mochinist
04-30-2008, 06:54 PM
I think that I should mention, I am fairly happy with Acad for drawings. I will probably see if I can get a trial of solidworks again and see how it is for 3d modeling, but I don't have any immediate need to switch from Acad. The reason behind the title, of cad/cam is I know that some cam programs include their own CAD software, anywhere from robust to rudimentary. If there was an all-in-one sort of deal that people are making good use of, I would happily consider it.I use the cad in Featurecam quite a bit, drawing in it is a little different though and a lot of people don't like it. Again it isn't a cheap program, I think for the basic 2.5d features you are looking at around 4000. Not sure how hard it is to pirate, but you can download the program to play around with from their website, it just wont spit out any g code.

oldtiffie
05-02-2008, 08:09 AM
I have responded and advised progress to date on items relevant to this thread at:
http://bbs.homeshopmachinist.net/showpost.php?p=354203&postcount=12

Its looking good!

Ken_Shea
05-02-2008, 08:34 AM
Its looking good!

Sure is oldtiff, I am getting excited for you :D
Are we having trouble sleeping yet ?

Ken

JimH
05-02-2008, 11:00 AM
Back to the original question of CAD/CAM.

Does anybody use the free version of Alibre design?

Any opinions on the Deluxe version (not the Professional) of TurboCAD and doing 3D?

Thanks,
Jim

lazlo
05-02-2008, 11:02 AM
I don't really mind them but there is too many doubts. Some have so much end-clearance with/on the adjusting screw that they "float/oscillate" axially as the "moving" part of the dove-tail operates.

Ah, that's a Chinese tapered gib :) Seriously, I had that problem on my Mill/Drill: it has a single-screw tapered gib, and the Mill/Drills have a very wide slot for the adjustment screw (manufacturing tolerance). So as you move the table, the gib tends to slip backwards into the screw slot.

It's a common fix on the Mill/Drills to either make a thick spacer, or add material to the slot (even JB Weld is fine) to take up the slack.

Also, "real" tapered gibs usually have adjustment screws on both ends, so the gib won't float.

chopperchuck
05-02-2008, 06:05 PM
i use alibre design much easer than auto cad 2006

oldtiffie
05-02-2008, 09:28 PM
What is the best "bang for buck" CAD and CAM separately and in combination with each other and works with Mach3?

I realise that this has been addressed several times in the not too distant past, but it has been re-addressed in the title of the OP to this thread:
"cad/cam, what are you guys using?".

So a "re-hash" to some extent seems topical and "in order".

Perhaps some who have not posted/commented before might like to post your advice/s here.

Is there anything newer of better "out there"?

Are you having any joy or disappointment with what you are using or have heard of from others with experience in using them?

Same applies to those who have posted before and might like to either restate or revise their ideas to perhaps tell us of something better for the HSM-er.

I suppose that cost/afford-ability, ease of learning and use in the Home Shop will be the main drivers.

jacampb2
05-02-2008, 09:55 PM
I spent a few hours the other night drawing a part in CamBam. I have to say that the machining operations interface seems to be fairly intuitive, however, the limited CAD capabilities really kind of sucks. It is easy enough to draw whatever you want, but you cannot specify start/end coordinates for lines and rectangles w/o first drawing your object, and then keying in the appropriate coordinates in the properties tree on the left. Long and short, it works, but seems very limited. I find it much easier to draw when I can specify points durring drawing and not have to move everything around once it is drawn. I found that setting a finer grid spacing than 1" helped a lot for the rough drawing, but orienting holes and arcs is more complex than it needs to be. I also could not get the "trim" command to work to save my life, and I have gotten very used to using that feature in Acad.

It took me about 2 hours to draw a 4x6 rectangle with a .5" slot 1.5"s from the +x side of the part and 3" long w/ a .25" radius at the end of the slot, and two .375" bolt holes. It took me less than an hour to assign the various machining operations. I could have easily drawn the part in Acad in 10 minutes or less. Obviously a lot of this was learning curve, and getting used to another new interface.

I have not yet tried to import a .DFX to it and assign operations, but I have to assume that as strictly CAM software, it will really shine, especially at it's price point :D

Hopefully at home this weekend I will get to try importing some .dfx files and see how it handles things.

Later,
Jason

oldtiffie
05-02-2008, 10:50 PM
Thanks Jason.

Any other general or over-all comment re CamBam?

I Googled it and this was the result:
http://www.google.com.au/search?hl=en&q=cambam&btnG=Google+Search&meta=

I went to the CamBam Home page
http://www.cambam.co.uk/
and a few others but I need more info from anyone that has used it - preferably by importing DXF files from a CAD system and then exporting the CNC code file/s.

Anyone with any comment on CAD and CAM generally and specifically?

I am totally new to CAM generally and various applications specifically.

Any comment and advice will be appreciated.

Mcruff
05-03-2008, 07:30 PM
For me Cadkey or (Key creator) beats anything on the market and has for years.
Heres why, when you draw on a piece of paper you simply draw a line and then reference from it. When you draw in virtually all other cad systems you are drawing on graph paper (so to speak) and must know the location of the point you want to work from. This causes you to have to remember lots of numbers and such. By drawing in Cadkey you don't need to know any points, only the reference that you want to work from, as in parallel to, perpendicular to, along this line so many inches and so on. There are no real world points that you need, you can if you want but thats also the hardest way to draw. Kind of like surveying your yard and saying I'm gonna start here and my back yard is 80' wide and 80' deep where as in most other cad systems you will need to know the coordinates of the 4 corners. Cadkey or Key creator(now) has been on the market since the mid 80's or so, is cheaper than autocad and will do way more for less money and do it alot faster.
They have a free download and you can find older versions on ebay quite often. I still run a version that was made in 98, still does everything I need for 2D and 3d drawing, I got Key creator V6 awhile back but I only bought it cause I got a good deal from a company that was going out of business.
I have used a bunch of cad sofwares over the years but always come back to Cadkey in the end!!

John Stevenson
05-03-2008, 07:51 PM
Trying to keep to the title CAD/CAM

OK there are three methods of doing this.

One is a CAD/CAM program that does both, you darw in one screen , that's CAD and process in the other screen, that's CAM.
Examples of this are Bobcad, Dolphin and a few more in advancing price.

Second way is to use a CAD package to draw your part and save as a DXF file [ Drawing Exchange File ] which can then be imported into:-

The third part, a CAM package, either stand alone or even one of the combined ones like Bobcad or Dolphin.

Recomending a package is a bit like religion and knurling :D
Many are set on what they can use, are familar with and probably got a host of drawings in that native format and no amount of recomendations will get them to change.

Mach3 does have a simple two tiered CAM system called Lazy Cam, the lower tier will bring in and process simple DXF shapes and drilling cycles, the paid for upper level allows you to use advanced operations and layers etc.

To be honest i haven't played with this as I have a CAM system but it's something I need to look at as people are asking about it for the turnkey CNC mills.
I aim to get to grips with this next week when Brian Barker of Artsoft comes over for the Harrogate show [ minus 4 days and counting .... ]

There are many free or very low cost CAD systems out there on the net, some very simple and some quite complex.
There are also free CAM systems out there, Cambam is one and the demo version of Flascut also spits out code that can be copy and pasted.

The bottom line is there is no one system that will suit everybody.
You need to download demos' and put some screen time in to first find a program you can draw in be it a combined CAD/CAM or a stand alone.

Without a drawing you are dead in the water.

.

Evan
05-03-2008, 09:22 PM
Jason,

The new version of CamBam is much improved. There are a lot of differences between it and the old one. It isn't released yet but I have been testing it for Andy for a month or so to date. There are still a number of bugs to work out and Andy tells me it is difficult to juggle the various demands on his time, something I am very familiar with when it comes to writing software. Writing software is like writing a feature novel and even the least interruption can entirely derail your train of thought.

oldtiffie
05-03-2008, 11:11 PM
Thanks Evan.

I have had a look at the CamBam web site.

The current version to my untrained eye seemed to be OK.

I notice that there are "free" and "pay-for" versions. That is not a problem.

I notice that it is DOS-based. Does that mean it requires a full DOS environment or will it run under the "DOS" utilities in the W2K and XP "Cmd" feature in a "Windows" environment?

Is CamBam's CNC code directly compatible with Mach3? If not,what additional work is required.

As I understand it, I would generate/export my CAD drawing in DXF format to CamBam - or equivalent - which will generate the e-Code (?) to/for Mach3 to "run" the CNC machining process/es.

Advice will be appreciated.

Evan
05-04-2008, 04:40 AM
CamBam is a Windows program, no DOS involved. I'm not sure why you got the idea it is DOS based. I see that Andy has moved to a public beta for CamBam Plus.

CamBam generates standard g-code and the output runs fine on Mach 3.

oldtiffie
05-04-2008, 08:38 AM
CamBam is a Windows program, no DOS involved. I'm not sure why you got the idea it is DOS based. I see that Andy has moved to a public beta for CamBam Plus.

CamBam generates standard g-code and the output runs fine on Mach 3.
Sorry Evan.

You are right!!

I must have plucked that one out of the air or from no -where!!!

I was (I thought!!) sure - seems not - I'd better be less sure and check better I guess.

I see that on the CamBam web site that CamBam plus is at Beta Release 0.9.2, Released 3 March 2008, build 2982.17276

http://www.cambam.co.uk

I presume that is the version you beta testing.

Could you give us a "run-down" on the "pro's and con's" of CamBam from a new users aspect in its application to/with Mach3 please in both the "free" and "paid (for)" versions please?

S_J_H
05-04-2008, 10:02 AM
This is a really great cam program and one of the best I have ever tried!
It even does continuous rotational milling.
Has an excellent wizard!! Simulation, all sorts of good stuff!
30 day free trial and then it stops creating G-code.
(shhhhh.....however there seems to be a loophole where if you use the wizard to generate a tool path it still generates the g-code after the trial period!)-
http://www.deskproto.com (http://www.deskproto.com/)

A very good free resource for parts drawings that you can import directly into your cam software in many different file types- http://www.3dcontentcentral.com/default.aspx

Steve

Evan
05-04-2008, 11:27 AM
Could you give us a "run-down" on the "pro's and con's" of CamBam from a new users aspect in its application to/with Mach3 please in both the "free" and "paid (for)" versions please?

I have no idea how to do that. I haven't ever been a "new user" as I started computer programming in 1963. The very first thing I do when trying out a "new to me" program is try all the common ways to make it die. Once I am reasonably certain it is stable then I go on to guess how it works. If I can't guess how it works then I might refer to the help files if I really need the program for some reason. In general though if I can't figure it out without help then I am not likely to like using the software.

Your best bet is to download and try it yourself. You don't need to have an actual machine to decide how you like using the program.

oldtiffie
05-04-2008, 07:27 PM
This is a really great cam program and one of the best I have ever tried!
It even does continuous rotational milling.
Has an excellent wizard!! Simulation, all sorts of good stuff!
30 day free trial and then it stops creating G-code.
(shhhhh.....however there seems to be a loophole where if you use the wizard to generate a tool path it still generates the g-code after the trial period!)-
http://www.deskproto.com (http://www.deskproto.com/)

A very good free resource for parts drawings that you can import directly into your cam software in many different file types- http://www.3dcontentcentral.com/default.aspx

Steve





Thanks Steve.

First of all, before going any further, I down-loaded and installed E-Machineshop as you advised on a separate thread as a good 3D CAD basic program. It is very good - particularly for someone who only needs a "parts only" (ie not "assembly") CAD. It generates a good drawing file - haven/t tried the export DXF function but I have no reason to doubt it.

http://www.google.com.au/search?hl=en&q=emachineshop&btnG=Google+Search&meta=

Setting the material and machine is OK as well.

The tutorial/help is very good indeed.

There are no dimensions on the drawing - but they are easily chased up and amended if needs be.

And its free!!

Many thanks.

Now onto "DeskProto":
Looks to be very comprehensive.
http://www.deskproto.com/

It sure is expensive - but if the "work-around" works as you say it will be OK. But I expect that if they find out about the "loop-hole" they will close it.

http://www.deskproto.com/order/priceinformation.htm

jacampb2
05-04-2008, 08:52 PM
I tried the new beta version of CamBam plus today. You are right Evan, it is much better. I may have to break down and license it... I still find it easier to generate my drawing in Acad, but for quick jobs, it is much improved.

Later,
Jason

Evan
05-05-2008, 08:56 AM
A few hints on using CamBam Plus;

To trim convert everything to a polyline. Then place trim fences using whatever line or shape you want. Highlight (select) the trimmer and trimmees and on the right click menu select edit and "break at intersections". Then you can remove whatever debris you like.

All translations and rotations are referenced to the origin. This is standard practice with 3D transformations. To make it easy when flipping things around place the entity you want to transform centered on the origin.

To copy and paste an object first do copy, then use the original as the copy and then paste. The paste will replace the original.

PaulT
05-05-2008, 01:01 PM
"What is the best "bang for buck" CAD and CAM separately and in combination with each other and works with Mach3?"

This might have been mentioned already in one of the many pages in this thread, so sorry if this is old news-

For 2.5D work, I think the best bang for your buck if $ is tight is the free 2D version of SolidEdge for CAD (search "SolidEdge 2D" and you'll hit it) and the SheetCAM package (www.sheetcam.com) for the CAM end. Both packages are solid and capability of serious 2.5D work, and the combo puts you out less than $200., definitely the king of "bang for your buck".

For 3D on tight funds, I use the "Basic" version of RhinoCAM with the Rhino 3D design package. The basic version will machine full 3D shapes but doesn't have as many machining methods as the full version, but for the prototyping and small run work I do it works fine. Demo versions are available from www.mecsoft.com . The combo package was around $1200. If you are doing detailed geometric type designs Rhino can be frustrating as a CAD tool, its strength is in free form 3D designs, but its a very powerful package for the money. For detailed work with it I do 2D views in SolidEdge and import them into Rhino to turn into full 3D designs. It works but a design change will often mean alot of tedious rework.

If I could justify the expense I'd probably buy the full 3D version of SolidEdge, the design flow it uses makes sense to me and the latest version looks really slick and would eliminate the "rework" issue I mentioned above with Rhino. Unfortunately right now I just don't do enough 3D design work to justify it, so I'll keep on with what I have for now.

I tried jumping on the Alibre bandwagon, but in the tutorial the amount of stuff you had to do just to get to the point where you can draw an object put me off. I didn't feel that way with SolidEdge.

Paul T.

S_J_H
05-05-2008, 07:44 PM
Some random thoughts before I begin my Cinco de Mayo celebration with some quality margarita's:D.

With the advent of home shop cnc, this hobby has progressed to the point where many guys need computers and software to go along with their lathes and mills.
I'm no computer guy by any means but even I (a dumb painter) know how to get ahold of free software in just about any category.
To name just a few methods and I'll admit to trying all of these mainly out of curiosity-
torrents,
keygens and cracks,
using a superior uninstall software along with time limited demos etc.. I'm sure the real geeks know many more methods!

You can sit behind a computer now and "steal" thousands of dollars worth of "product", be it software, music or video while munching on potato chips and having a cold one. And chances are mighty slim you'll ever get in any sort of trouble. But it's still stealing and dishonest IMHO.

So keeping it on track-
Most of the more expensive cadcam stuff I find way over my head and/or have no need for software that powerful. Seems geared for pros and production shops etc.
I am just a HSM guy and make or sell nothing for profit.
I paid for my versions of MACH3 and Sheetcam which I use primarily.
It's got to be pretty hard for software companies these days, especially the little guys who I will definitely always support!

Steve
http://www.vivacincodemayo.org/history.htm

Evan
05-05-2008, 10:55 PM
I have never had an opportunity to try a $3000 software package. I have always wondered if they are that much better than a $300 program.

I do know that as an example Paint Shop Pro at $150 retail is just as capable as Photoshop at 5 times the price.

I do have older versions of various CAD programs that I didn't pay for and would be considered "abandonware" because either the company that sold them no longer exists or no longer supports or distributes the product. That doesn't change the status of the copyright although in the past many companies never bothered to include intellectual property in the assets of the company when it folded. In some cases it means nobody owns the copyright.

Still, is a mega bucks program that much better? I sure can't justify paying anything close to what much of the software sell for and I am totally opposed to buying time limited licenses for anything. I believe I should only have to pay once for a product.

toastydeath
05-05-2008, 11:44 PM
I have never had an opportunity to try a $3000 software package. I have always wondered if they are that much better than a $300 program.

There's usually a difference. Whether it's a 2700 dollar difference, that's certainly in the eye of the beholder.

I've dorked around with Solidworks, and Pro/E. Pro/E, while far more expensive, will do a whole lot more stuff. If a person doesn't know what they're looking at, they look the same. But there's a huge difference in the small stuff - mostly blending surfaces together (fillets) and other complex curve manipulation. Most people will read that and say, "Well, whatever, I'm not making aerospace parts," but it does come into play even with simpler objects. Anytime I've had Solidworks flip out on me, Pro/E will do the same thing without a peep.

gellfex
05-05-2008, 11:47 PM
There's a school of thought regarding "pirating" expensive professional level software for personal use, that the vendor benefits from gaining market share even if it's unpaid. There's no way Photoshop would be as dominant if everyone didn't learn it on a pirated copy before going out and getting that job at a licensed seat. Some people think it's part of their business plan. There's simply no way I would ever pay for a Solidworks seat, but if I did use a pirated copy for my 1 man shop, that's one more person out there that might someday be hired to use the software at a larger firm. They lose no sale, but increase their user base.

For the record, I own the license for the DesignCad I've been using for the last 18 years, my needs aren't fancy. I use a pirated copy of Word, but I think making your living with unlicensed tools is wrong. I'm contemplating buying a SketchupPro license, it's a pretty useful program, with many of the features of Solidworks. If they ever get the bugs out of the free SketchyPhysics mechanics simulator plugin, it'll be killer. I'd also love to know if anyone has successfully exported a dxf or dwg file from Sketchup to CAM.

oldtiffie
05-05-2008, 11:52 PM
..................................
.............................
I'd also love to know if anyone has successfully exported a dxf or dwg file from Sketchup to CAM.

Thanks gellfex.

That is one damn fine topical question in the context of this thread/topics and that many will have or are/will/may consider getting even the "free" version of "Sketch-up"

Evan
05-06-2008, 03:46 AM
The main issue I have is that I am constantly pushing the limits of the software and beyond. Just in the last couple of days I have need to fill an 8" x 11" area with a dense pattern of points spaced .040 from each other. CamBam can't do it without choking and neither can any of the other programs I have. It seems like every other thing I try to do is too much for the software. The other day I wanted to plot a spiral that increased by .002" every 360 degrees out to 10" diameter. That is a lot of polyline segments and every program I tried it on decided to do some long term navel inspecting.

My Cad computer isn't the fastest but it sufficient with an Athlon XP 3000, 1.5 gigs of ram and a Nvidia gforce 6200 video card. I see the same problems on my dual core laptop also.

As a comparison I have been using Paint Shop Pro since version 2 when it was only a file format convertor program. I have every version since then. I commonly work in PSP at very high resolutions. It isn't at all uncommon for a single graphic sequence to add up to several hundred megabytes. Try that with the average low end CAD and it will be the last time it speaks to you.

Peter N
05-06-2008, 04:52 AM
I use Solidworks and have done for the past 6-7 years.
I started off with a pirate copy to evaluate it, then ended up buying it and staying on maintenance right up to this year.
Very, very useful, and definitely necessary since I use it for business. Additionally, it actually payed for itself with the fisrst 2 product design jobs we used it for.

It's got more poweful with more features every year (although much of these I don't use), but I agree with toastydeath, ProE takes it up another level.

However, ProE has a godawful IGES translator and we get no end of problems with IGES from this, and also from Catia. I had a moulding and tooling job in to quote the ther day, nothing remotely complicated at all except for lots of engraving on the part. The ProE IGES exported it with 7067 individual surfaces and the damn thing took 7.5 minutes to load into Solidworks.
We have now asked them to send it as a STEP file which will not only be a much smaller transfer but should bring it in as a single solid.

Peter

S_J_H
05-06-2008, 08:35 AM
There's a school of thought regarding "pirating" expensive professional level software for personal use, that the vendor benefits from gaining market share even if it's unpaid. There's no way Photoshop would be as dominant if everyone didn't learn it on a pirated copy before going out and getting that job at a licensed seat. Some people think it's part of their business plan. There's simply no way I would ever pay for a Solidworks seat, but if I did use a pirated copy for my 1 man shop, that's one more person out there that might someday be hired to use the software at a larger firm. They lose no sale, but increase their user base.

That's an interesting theory and I think may indeed be valid.
You won't have a very hard time finding software from the big guys like solidworks-
http://search.utorrent.com/search.php?q=solidworks&e=http%3a%2f%2fwww.mininova.org%2fsearch%2f%3futor rent%26search%3d&u=1
And it's such a huge and complicated install there will only be a tiny % of people with the skills to get it to run so it probably does not put the slightest dent into their sales. I'm just guessing though.
It seems the bigger or more popular and well known you get the more your software is likely to be pirated. Tried to find sheetcam pirates just as a test( which I already own a paid license) for example. Nope it's not out there. Small time operation.
Mach3 for example has gotten big enough now that there are cracks and torrents( I paid for my version). Seems a shame that people have to steal software like Mach that runs for under 200$:mad:.
Steve

Evan
05-06-2008, 09:51 AM
I'm not sure I subscribe to that theory and I know the BSA (Business Software Alliance) sure doesn't. They now offer rewards up to 1 million dollars to people who turn in software "pirates". They also don't just focus on large corporations. Another computer store in this town was audited by Microsoft on one occasion. By an incredible coincidence the night before the audit a fire broke out in the store. It destroyed all the sales records and then put itself out by melting the water line to the cooler which sprayed directly on it.:rolleyes:

I ran my computer business with one basic policy. If you want me to copy something for you, you must own the copyright, no exceptions.

The home user may be somewhat safer but you only need to look at what the RIAA/MPA are doing in terms of suing students and 12 year old girls for "stealing" music.

I will admit that I have on rare occasion used "pirated" software. I offer as nothing more than a rationalization that I have spent $1000s in total on purchased software personally. If I include the software I have sold with computers the figure is hundreds of thousands.

Although I have not tried to find "cracked" versions of the big name CAD/CAM programs I am well aware how and where to do it. I really think that the big name software companies would help their businesses if they would make their products available to the basement home shop user in some way. There isn't a ghost of a chance I will ever buy software that costs $1000 plus but I might be in a position to recommend it to somebody that will.

Stepside
05-06-2008, 10:58 AM
I used RhinoV4 to do Evans point problem and the same for his spiral problem.

An array of points .040 apart covering an 8 inch by 11 inch rectangle took less than 2 minutes to draw and array. It did not look like much until I "zoomed in" and could see the individual points. The spiral took me less than 3 minutes. In both cases the time included choosing the command and setting its parameters.

In both cases it was necessary to "Zoom In" to get the drawing to appear as something more than a black blob on the screen. This assumes, of course, that I understood what was being asked.

I realize this product approaches the $1000.00 window. One has a couple of choices. First would be to download the demo version for free. It will give you 25 "saves" before you can no longer save. If you wish to purchase, go on line and shop price as this usually gets you down to $875.00 give or take.

gellfex
05-06-2008, 12:23 PM
I really think that the big name software companies would help their businesses if they would make their products available to the basement home shop user in some way.

My point is that from their perspective that IS the current situation. They know that if you really want it, you can find it, but anyone professional enough to be an audit candidate won't do it. Status quo, without them having to explicitly give it away.

Evan
05-06-2008, 12:39 PM
$87.50 I can justify. $875.00 buys me a years worth of insurance for my truck. I need my truck more than I need Rhino.

In case anyone wants to know why I need the spiral it's a tool path for roughing out a glass mirror to a parabolic shape with a diamond cutter. It saves maybe 90 percent of the time in making a telescope mirror.


Status quo, without them having to explicitly give it away.

I really don't think marketing would be happy with that scenario. I also don't expect them to give it away. They can certainly come up with some restrictions that would be effective in preventing people from distributing the home shop version. Something like custom coding your copy to display your registration information such as address and phone number on the splash screen when it boots and including it in every file saved. Any decent hacker could remove that but that doesn't describe the majority of home shop machinists.

gellfex
05-06-2008, 02:55 PM
$87.50 I can justify. $875.00 buys me a years worth of insurance for my truck. I need my truck more than I need Rhino.

In case anyone wants to know why I need the spiral it's a tool path for roughing out a glass mirror to a parabolic shape with a diamond cutter. It saves maybe 90 percent of the time in making a telescope mirror.



I really don't think marketing would be happy with that scenario. I also don't expect them to give it away. They can certainly come up with some restrictions that would be effective in preventing people from distributing the home shop version. Something like custom coding your copy to display your registration information such as address and phone number on the splash screen when it boots and including it in every file saved. Any decent hacker could remove that but that doesn't describe the majority of home shop machinists.

You're right that they could easily add more security to prevent keygens and such, but they don't. They do just enough to "look" like they're serious about protecting their property. It's a legal fig leaf, so they can legitimately prosecute commercial illegal users.

BTW, why would the spiral path be that much better than concentric stairsteps? I've done all sorts of stairsteps for turning elliptical and other shapes.

lazlo
05-06-2008, 02:58 PM
$87.50 I can justify. $875.00 buys me a years worth of insurance for my truck. I need my truck more than I need Rhino.

I'm in the same situation -- I'd gladly spend up to $200 for a Solid Modeling CAD and/or CAM package, but $995 is out of the park for me. For that price, I can buy an X3 to CNC :)

I think Art priced MACH perfectly for it's capabilities and the target market.

Evan
05-06-2008, 03:23 PM
why would the spiral path be that much better than concentric stairsteps?

With a spiral tool path the mirror shape will represent the integral of all the calculated data points. With stepped circles only the values of the steps are cut with nothing in between. It makes the grinding more difficult since with stepped circles no part of the mirror has been pre cut to values in between the steps..

mochinist
05-06-2008, 07:23 PM
The main issue I have is that I am constantly pushing the limits of the software and beyond. Just in the last couple of days I have need to fill an 8" x 11" area with a dense pattern of points spaced .040 from each other. CamBam can't do it without choking and neither can any of the other programs I have. It seems like every other thing I try to do is too much for the software. The other day I wanted to plot a spiral that increased by .002" every 360 degrees out to 10" diameter. That is a lot of polyline segments and every program I tried it on decided to do some long term navel inspecting.Does mach 3 support parametric programming? I've never really got into it, but I would think with your programming background it would be pretty easy to pickup, and I think you could use that to generate your tool path you describe http://www.cncci.com/resources/parametric%20programming.htm


I have never had an opportunity to try a $3000 software package. I have always wondered if they are that much better than a $300 program.I haven't played with cambam much but I did fool around with bobcad 8 or 9 years ago and it was a nightmare compared to Featurecam. For me it is worth it, but my cam program is used 99% of the time to make money, if it was a hobby I would surely think differently. I don't want to copy/paste their sales page that describes some of their options, but you could kinda compare it to cambam with this link, if you feel like it.:cool:
http://www.featurecam.com/general/software/featuremill2d.asp

Evan
05-06-2008, 08:00 PM
Mach 3 is strictly a machine controller. It runs g-code and the only programming it supports is pre written "wizards" that are similar to conversational programming. It does come with a rather lame accessory called Lazycam which has a very few good features but can be brought to it's knees easily by a complex tool path.

The ability to handle complex drawings and the resulting tool paths seems to be the most difficult thing for the programmers to get a handle on. There are very few programs in the low end that do a good job.

There is a reason for that and it has to do with some not very obvious characteristics of the math used in calculating a 3D environment. It is very easy to get in a trap where the code that must be run grows at a rate that is proportional to the third power of the number of vertices/polygons etc that are in the drawing or model. This means that simple drawings and objects are fast to compute but as the complexity grows the time required for calculation increases almost ten times faster than the increase in the number of entities.

Working around this problem is possible but not easy. It requires some very sophisticated programming techniques. As I have said before these problems have been solved in the gaming industry but that doesn't mean those skills have been transfered to the CAD industry. It also doesn't help that the techniques used are usually a very closely guarded secret and also required a very high level of understanding of mathematics. A further problem for the CAD companies is that 3D software development in the game industry can be far more lucrative to a skilled programmer than boring CAD software.

John Stevenson
05-07-2008, 03:48 AM
One problem you haven't mention though Evan on the 3D gaming is that it can be fudged.

Does it really matter when Zork jumps 2 feet to miss the White Gronge he doesn't go 2 feet or actually passes thru part of the magic path ?

Matters a great deal to anyone flying in a Boeing, hence the spiraling maths.

However there are parts of games that can alter the ease of working. Watch a space very close to Mach3 in the near future.

Brian Barker of Artsoft is coming over this weekend for the UK show at Harrogate and will get his ear bent at the pub.


.

Evan
05-07-2008, 04:45 AM
Does it really matter when Zork jumps 2 feet to miss the White Gronge he doesn't go 2 feet or actually passes thru part of the magic path ?

It sure does to the hard core gamers. A game with sloppy collision detection won't last long in the market. I'm being serious too. These are the sort of problems that game programmers deal with all day and have direct application in controlling a machine. You don't want the machine to try and eat a clamp either regardless of the direction of approach.

Stepside
05-07-2008, 03:25 PM
Could we have a description of the mirror you wish to make. The curve that describes the cross section and the diameter. Is this a mirror with a hole in the center or is it solid?

Could you give us a quick overview of how you plan to proceed with construction? An idea of the tolerences required would be great.

Thanks

Pete

Evan
05-07-2008, 07:08 PM
This is a back burner project at present. I am however going to try milling one from aluminum using the same technique to see how close it comes. That will give me an idea about how long it will take to do the final grinding of a glass one.

These days that is how mirrors are made commercially. It isn't worth the time and effort to make one the old fashioned way unless you like whipping yourself for weeks. However, if I can cut the amount of hand work by 90% then it starts to make sense again, especially if I also build a semi automatic grinder. 9 inch binoculars would be nice....

The finished tolerance is an absolute minimum true parabolic shape to within 1/4 wavelength of green light. 1/8th is better. The closer I can come to that the less work there is to do. The theoretical minimum step size on my machine is about equal to one wavelength of deep red light.

Stepside
05-07-2008, 08:38 PM
Nice going Evan. Now I have to learn some more stuff. Sure beats watch "Dancing with the Stars". I guess what I was asking is "how much material is removed?" or "what is the maximum "Z" distance?"

Evan
05-07-2008, 09:48 PM
The "Z" distance on a mirror is called the Sagitta. That is the depth of curvature at the center. It determines the focal length of the mirror. The deeper the sagitta, the shorter the focal length, the greater the amount of material to be removed and the more difficult it is to generate the exact curve required.

Sagitta (s) = Mirror_Radius2 / ( 4 x Focal_Length ) = r2 / 4F

Astronowanabe
05-08-2008, 12:09 AM
Took a telescope making class 4 years ago from John Dobson.
I had never had a telescope or heard of the guy.
What a character. Unfortunatly some of his instructions left me wanting ...
for example ...
"Grind like a caveman till you can slip a penny and half a thin dime under the ruler"

Ok ... so before I had ever met the term saggita I wrote this up and sent it to the class.
then found out that most were more comfortable with Dobson's version

_____________________________________
rough guide to rough depth for rough grinding

assuming curve is paraboloid ...
and p is distance from vertex to focus point

for vertex at (h,k) parabola (x - h) * (x - h) == 4 * p * (y - k)
or rearranged (x - h) * (x - h) / (4 * p) + k == y

let the vertex (h,k) = (0,0)
let radius r = D / 2
let focal ratio F = p / D

since the depth of the mirror is the height of the parabola at the edge ...
solve for y at x = r for depth of concavity

r * r / (4.0 * F * D) == y

(D * D / 4) / (4 * F * D) == y

D / (16 * F) == y


so:

depth = Dia / (16 * Fnumber)
(translation to "thin dime" units left as an exercise)

Evan
05-08-2008, 02:11 AM
I think John Dobson is the only person currently living that has a telescope type named after him. He is responsible for the low cost- most bang for the buck- Dobsonian style of scope.

jacampb2
05-16-2008, 09:57 PM
So, to bring back my almost dead thread, how about DelCam Artcam?

I skimmed the thread again and didn't see anyone mention it, the software looks like it is mostly geared toward sign/wood working and artsy decorative crap. I am playing with a trial version. There is not a lot of written tutorial stuff that I am finding, but there are some good videos on the delcam web site. It is so incredibly far from anything I am used to, I could do pretty much nothing but load a .jpg on the first time playing with it.

I watched a few of the video tutorials, and found it is VERY simple to generate 3d engravings from bitmaps. A black and white logo for instance can easily be transformed to a 3d carving/engraving. The more colors you have, the more time it takes, but I am looking forward to fiddling with it more on my next days off. It also has some powerful tools for basically engraving photos. The demo is I believe time unlimited, but does not allow saves.

The bad is it is some ungodly price, like most of the powerful cam packages. Google turns up lots of shady links for the software free though...

Later,
Jason

John Stevenson
05-17-2008, 03:44 AM
Jason,
Take a look at V Carve then from Vectric.
Tony and Brian who run the company used to work for ArtCam.

Does a lot of what Art Cam does at a 1/10 of the price.

.

Swarf&Sparks
05-17-2008, 12:44 PM
I'll second sir John re Vcarve
It's a steal at the price and very friendly, helpful people to deal with.

Literally minutes after I had 3 axis test done on my little router, I was turning out results like this:

http://i23.photobucket.com/albums/b387/idgara_eng/smithy2phys.jpg


No axe to grind, no connection, just a more than satisfied user.