PDA

View Full Version : Casting a thing of the past ?



John Stevenson
06-12-2008, 08:02 PM
Or just easier to produce patterns for short runs?

Playing about today using Solid edge to get a .stl model of a hypothetical crankcase then move it into Vectric's Cut3D program to generate the code.

Direct link here but 5.6meg download

http://www.stevenson-engineers.co.uk/files/crankcase.wmv

.

Chipslinger
06-12-2008, 08:32 PM
Like watching grass grow. :)

I bet it was more fun doing it.

fasto
06-12-2008, 09:38 PM
Sir John, forgive my lack of knowledge, but how does one hold that part to the table? I presume it's got 4 tabs in the corners, no? When the cut out toolpath is run, then how is the part held? :confused:

toastydeath
06-12-2008, 10:48 PM
That's just a demo. Either the part isn't designed to be actually made, or they're going to make it out of a deeper block than the finished part.

Casting is still very much alive and well, as many of those toolpaths take hours (or days, in some cases) to run. Can't exactly make 100 parts a week if it takes you 30 hours to machine. What you're seeing is appropriate for prototype and similar work, where several zeros attached to the end of a quote is not an inhibition. It's also appropriate for making patterns and cores, ironically.

Edit:

Here's a cool video of a part like that getting made.

http://youtube.com/watch?v=pPCrTYwp2UY

lazlo
06-12-2008, 11:02 PM
Casting is still very much alive and well, as many of those toolpaths take hours (or days, in some cases) to run.

There's a neat How It's Made episode where they're making aluminum engine blocks, and it's all castings. But these aren't your GrandPa's castings: nearly everything on the line, from pouring the sand to tamping, to coring the sprue holes, is done robotically. There's not a human in sight, even for quality control, which is done with image recognition.

There was another slick casting episode about making pulleys. The casting sand is used for a mold, a set of pulleys are cast (these were cast iron), the sand casting is shredded and the pulleys separated, and the sand is immediately recycled for the next set of castings.

tattoomike68
06-13-2008, 02:11 AM
That's just a demo. Either the part isn't designed to be actually made, or they're going to make it out of a deeper block than the finished part.

Casting is still very much alive and well, as many of those toolpaths take hours (or days, in some cases) to run. Can't exactly make 100 parts a week if it takes you 30 hours to machine. What you're seeing is appropriate for prototype and similar work, where several zeros attached to the end of a quote is not an inhibition. It's also appropriate for making patterns and cores, ironically.

Edit:

Here's a cool video of a part like that getting made.

http://youtube.com/watch?v=pPCrTYwp2UY

I agree casting is alive and well. Unless the customer wants billet machined, LOL.

I would let the CNC cut a pattern oversized with shrink and warp factors added. Have that cast and let a cnc finnish cut the whole thing in lets say 4 minutes VS 10 minutes with all the roughing passes needed. Some guys may want to work thier machine to death and make chips for the chip conveyor, myself I want to make parts that are done quick if thats my bread and butter.

John Stevenson
06-13-2008, 03:53 AM
As TD has said it's just a demo but it will be cut to go on the example board that we take to the shows, probably cut two, one as the rough waterline part with the first tool and the finished item with the ball nose cutter.
It will be held with a couple of screws tapped in from underneath on a sacrificial plate so it can go all round to machine the shape out. The inside on the finished engine would be machined out later to accept the crank so that material where the screws fit is not needed.

Probably run this tonight after work whilst I'm having tea, just waiting for a parcel of cutters to arrive.

True you don't want to make 100's of these with these time restraints but a one off for the pattern is still well in on time compared to conventional pattern making.

For a prototype it's still quicker given the same time to make the pattern, travel to the foundry, leave it to be cast, travel back etc but the whole job is more for interest than use. We are always looking for examples that would take a lot of making by manual methods to work with.

.

ptjw7uk
06-13-2008, 04:58 AM
Im afraid this is just another indication of the march of progress in that in the olden days machining most shapes from the solid could not be done!

As a one off for a look see machining from solid is ok but not in the workplace where even traditional casting has given over to lost wax casting processes by reasons of complexity.

Still nice to watch it being machined!

Peter

Charles Ping
06-13-2008, 07:39 AM
John

I think that you're onto a good thing for non cored parts -and a bike crankcase is a great example. I can think of others where I've had to import timing case castings from Australia and, because they're "3rd generation" castings (ie made from a pattern that was copied from a part), the critical dimensions are slightly out and a lot of compromise is needed in the subsequent machining. It would be much easier to go your way if you have the machinery and associated software skills.

A friend of mine bought a Speedway Jap crankcase, cast in Scandanavia I think and machined in the Czech republic. Even then they weren't cheap. I wonder how yours would compare.

Charles

Swarf&Sparks
06-13-2008, 10:08 AM
John, have you tried any lost foam/wax patterns with Cut-3D?
I have Vcarve and I'm well impressed. Does Cut3D allow independent axis scaling for shrinkage etc?

John Stevenson
06-13-2008, 04:48 PM
Charles,
I know what you mean only too well.
Recently I was given a water jacket housing that had the holes and counterbore clearly marked on the casting, I was asked to drill and counterbore them and face the main face up which I did.

Later this was returned with a gasket and I had to slot the 4 outer holes as this casting was taken off an original and no allowance was made for shrinkage. This was on an old Alvis and there must be many of the old clubs out there having the same problems.

Swarf,
Yes Cut 3D will make allowances.

This is the original setup screen for that crankcase.

http://www.stevenson-engineers.co.uk/files/3d%20default.jpg

Just above the picture of the car model is a ticked window which reads , "Lock XYZ ratio"

In this picture I have unclicked it and changed X from 77.99mm to 120.0mm

http://www.stevenson-engineers.co.uk/files/3d%20stretch.jpg

So it is possible to draw an accurate model and have two files, one that produces the pattern with shrinkage and one that will machine the part to original size.

So far this is as far as I have got as my bag of cutters didn't get delivered today.



.

Chipslinger
06-13-2008, 05:25 PM
As an amature I need to ask....

say you have the measure ments for a copy and had to make the real thing to original size, which is about 1/8" larger.

Could you use a smaller ball end mill say from 7/16 to 3/8 to compensate?

Charles Ping
06-13-2008, 05:58 PM
OK John.
To do this on a decent scale (not in production speed terms but simply to be able to make real engine parts rather than model parts) what's the best approach to getting a reasonable machine .
Is it refurb a Bridgy Interact and add Mach 3 or would you start somewhere else. What sort of sterling cost is it to get set up with a machine, software and associated kit?

John Stevenson
06-13-2008, 07:04 PM
Charles,
It all depends on size of part. Guessing what you want it for then I'd say go as big as you can.
Pays off in getting rigidity and also, daft as it sounds it's cheaper.

A Bridgy BOSS 1 to 5 is an easy way to go, quite a few of them about, usually broken as the electronics were crap. They go from free to about 900.
Just looked on ebay but there are not listed but a good idea of what is available is this one for 500 notes.

http://cgi.ebay.co.uk/Supermax-CNC-manual-milling-machine-Anilam-m-control_W0QQitemZ190225326059

That will get you a CNC built machine as opposed to a conversion with ballscrews,motors, mounts and pressure oiling etc.

The electronics can be dumped except for the cabinet and transformer.
You then need and all in UK pounds.
3 drivers @ 70 each
Power cap at 25
Handfull of assorted electronic bits at say 25
Breakout board at 50
Computer, usually for free or a few quid.
Mach license at 80

You may or may not have to budget for an invertor to get use of the spindle motor or for use on single phase

That will get you a running machine.

3D CAD software to get the file, Alibre Express is free and can do the parts you need, Solid Edge, Solid works and Rhino do educational copies if you can get on a scheme for about 200 a pop.

Cut 3D costs 160

Add to this is fixtures and tooling as needed for the jobs so it's not too great an outlay if you do it yourself.

The alternative is to buy new, either the Sieg or the Tormach if it will fit your budget and size restraints.

.

J Tiers
06-13-2008, 10:16 PM
You machine those, and you'll have 70% + of the weight of the blank in expensive chips......

Someone once told me that the milling machine was invented purely as a matter of necessity because casting or forging alone wasn't good enough. Mills are too expensive to use unless there is absolutely no alternative. He had a point.

how come that fancy expensive simulator gizzie produces a part looking so coarse and crude? Can't it interpolate and smooth out the layers (I know it can)? Looks like a really poor quality "grown" part. Maybe no problem for the proto?

John Stevenson
06-14-2008, 05:09 AM
JT, You miss the point.
That job could be the pattern that goes to the foundry for a 10,000 off run so that 70% of sod all which would have been lost anyway by any manufacturing means is a drop in the pond.

For a one off it still pays as you would have no casting or forging anyway and still have to knife and fork it.

Probably the title was a bit misleading as this wasn't a post about doing it this way instead of getting castings more a heads up.


how come that fancy expensive simulator gizzie produces a part looking so coarse and crude?

Fancy? expensive ? at 160, you need to get out more and read a few more adverts.
It can clean it up and make it more lifelike just as the $10,000 programs can but it takes more passes and more time.
The roughing pass is just that, why waste time doing nice Z raster cuts when another cutter is going to come along and remove it ?

I know where Charles is coming from as he's heavily into vintage cars and the clubs that run the spares programs have access to foundries but not pattern makers at reasonable cost.

The Alvis club want some bell housings doing for the later big sixes but so far the best quote they have for a pattern is over 2,000 and the anticipated run of castings won't cover that.
The cost is due to the fact that the pattern needs to be in 5 pieces and one core print.

It maybe worth while to have another look at that job. I was going to machine them if they had the casting done anyway so the alternative is to have a simple casting done like a large diameter pot and then remove the excess metal away.
It would be slow but sod it the machine can earn it's keep whilst I'm in bed and they would only want about 6 or 8 a year.
The internal shape won't matter a hoot as long as there is clearance and the outside can be blasted anyway to get it looking original.

.

.RC.
06-14-2008, 06:25 AM
http://youtube.com/watch?v=pPCrTYwp2UY

If I chucked something like that and tried to turn it it would not be long before it would be thrown out of the chuck at Mach 2...

DR
06-14-2008, 07:53 AM
What am I missing here? What's the point of this thread?

If I wanted to make a single crankcase like that shown, there's a much easier way. Send the CAD file to an investment casting company.

They use a rapid prototyping machine to make a wax model of the part including the hollow backside. That model is then used to cast the part to fairly high precision. Of course the casting would still need flat mounting surfaces to be machined and holes to be bored, drilled, reamed, or tapped to exact size, but so would the Cut3D version.

Cut3D does appear to be a relatively inexpensive program. I don't understand it's machining strategy. Why the diagonal finishing pass, why not simply finishing incrementally across in the X or Y axis direction? The toolpath file would be smaller that way.

I wonder how useful Cut3D really is. It doesn't appear to be a substitute for a general purpose 3D CAM program. What about simple profiling, hole making, pocketing, etc.

J Tiers
06-14-2008, 08:17 AM
JT, You miss the point.


nope, wasn't talking to you with that.......

Although the rapid proto + casting method is going to be more like the actual part........

There is a difference in a casting vs hogging it out of "billet" :rolleyes: not only in appearance and machining time, but also in properties. For some applications a hogged-out part won't actually "imitate" the production part.



Fancy? expensive ? at 160, you need to get out more and read a few more adverts.

Why?

I don't have, nor particularly want (at the moment) a CNC.

And besides, that 160 is equal to about $5000 by now, isn't it? :D

John Stevenson
06-14-2008, 08:27 AM
And besides, that 160 is equal to about $5000 by now, isn't it? :D

Just about..........

.

mochinist
06-14-2008, 10:38 AM
What am I missing here? What's the point of this thread?

If I wanted to make a single crankcase like that shown, there's a much easier way. Send the CAD file to an investment casting company.

They use a rapid prototyping machine to make a wax model of the part including the hollow backside. That model is then used to cast the part to fairly high precision. Of course the casting would still need flat mounting surfaces to be machined and holes to be bored, drilled, reamed, or tapped to exact size, but so would the Cut3D version.And what would that cost and how long would it take? Not asking to be feisty I have just never had to do that before and would be curious. Most of my customers seem to want their parts yesterday, some of them can be talked into waiting a couple days:p


Cut3D does appear to be a relatively inexpensive program. I don't understand it's machining strategy. Why the diagonal finishing pass, why not simply finishing incrementally across in the X or Y axis direction? The toolpath file would be smaller that way.

I wonder how useful Cut3D really is. It doesn't appear to be a substitute for a general purpose 3D CAM program. What about simple profiling, hole making, pocketing, etc.I thought the tool path was kind of odd also but the program seems to be easy to use and the price is right, the last time I talked to my Featurecam sales rep, the 3d add on for cam program was something like $5000us.:(

lazlo
06-14-2008, 11:20 AM
I thought the tool path was kind of odd also but the program seems to be easy to use and the price is right

The toolpaths are not optimized like you get with MasterCAM (for 20x the price :) ).
In John's case (pardon the pun) Cut3D seems to want to cut the bottom perimeter, then jump to the upper left corner, and then continue the toolpath ???

Edit: I just realized what the software is doing: it's scanning the part left to right from Top Left to Bottom Right. If it runs out of workpiece on the bottom, it picks back up at the Top Left. The toolpath would be way more efficient if he followed the Z contour of the part instead, like MasterCAM does.

Does CamBam or Vcarve do any better job on the toolpaths?


If I chucked something like that and tried to turn it it would not be long before it would be thrown out of the chuck at Mach 2...

No kidding -- that's like a 5:1 stick-out. I think that's a power chuck though, and notice there's virtually no runout at the end of the weird vase thingy.

DR
06-14-2008, 12:32 PM
And what would that cost and how long would it take? Not asking to be feisty I have just never had to do that before and would be curious. Most of my customers seem to want their parts yesterday, some of them can be talked into waiting a couple days:p

I thought the tool path was kind of odd also but the program seems to be easy to use and the price is right, the last time I talked to my Featurecam sales rep, the 3d add on for cam program was something like $5000us.:(


How much for casting? Based on customer parts I've worked on my guess is between $500 to $1000. That would be about the same as having a shop do it as John showed (at normal CNC shop rates). Timing would be a week or so, probably the same as machining.

This part doesn't really show the benefits of casting since it has no areas that can't be reached by a cutter. If it had internal passages for fluid, then casting would be far, far better.

Generally speaking the rapid prototyping/investment casting process wouldn't exist unless it was more cost effective than machining.

Cut3D's cost? I'm not so sure it's that great a deal in comparison to a program like Vector CAD/CAM. Vector is (was?) available for around a thousand bucks and includes 3D CAD, plus about 100 more capabilities.

It's my experience that newbies to CNC get excited about 3D surfacing. When you come down to it very little, if any, of that is done in the vast majority of production CNC shops.

J Tiers
06-14-2008, 01:03 PM
It's my experience that newbies to CNC get excited about 3D surfacing. When you come down to it very little, if any, of that is done in the vast majority of production CNC shops.

Just about any part I would want would need a good-looking surface, if that is what you mean............... By "good-looking" I mean faired and smoothed similar to a typical casting.

DR
06-14-2008, 01:33 PM
Just about any part I would want would need a good-looking surface, if that is what you mean............... By "good-looking" I mean faired and smoothed similar to a typical casting.

The point I wanted to make is, for the newbies, don't run out and buy Cut 3D. For general CNC machining it just isn't that useful a program. Spend more and get a package that'll do all your needs.

Newsflash!!!!!! I just remembered you can get a free download of Mecsoft's Freemill that will do almost the same thing as this $300 software. It appears to do the same type 3D surfacing as Cut 3D. But, in both softwares that's all they do which leaves you needing another package for the vast majority of your CNC'ing.

lazlo
06-14-2008, 02:02 PM
Newsflash!!!!!! I just remembered you can get a free download of Mecsoft's Freemill that will do almost the same thing as this $300 software. It appears to do the same type 3D surfacing as Cut 3D.

Tha's pretty neat -- thanks for that DR! They're the VisualMill guys:

http://www.mecsoft.com/Mec/Products/FreeMill.shtml

http://www.mecsoft.com/Mec/Products/images/FreeMill.jpg

John Stevenson
06-14-2008, 03:04 PM
No wonder it's free, it's a joke.

Decided to compare it to see if it works the same and save some folk a bit of money.
DL it and install with no problems.
Import the crankcase file and it says it's too big for an inch file do I want to select metric ?
Yes please as I was going to click the metric box anyway but what is the limit on file size then ?

Ok loaded up, select the cutting face, depth etc very similar to Cut3D.

Select the tool in this case the roughing tool of 6mm, then go on to speeds and feeds which although I have selected metric are still in in / min.
So far I haven't seen anyway to limit the cut per pass but we go on.
Next tool for finishing ? Nope you only get one tool.

OK go back and select a 3mm cutter, still no depth per pass.
Next screen is simulation and it's very rough as there is only one tool doing the work.

Next screen is the post file screen, no post for Mach so select generic Fanuc post.

%
N1G20G90G99 [Select inch units ?????]
N2G40G17G80 [ cancel offsets as expected ]
N3T1M06 [ select tool 1 which was a 3mm cutter]
N4M37
N5G0X3.632Y-0.337S4000M03 [select speed and switch on spindle]
N6G43Z0.118H0 [ move tool 0.118" above the work ]
N7G1Z-0.756F1 [ Stuff a 3mm cutter 0.756" deep into the work at 1" per minute ]
N8Z-0.78 [ Stuff it a bit deeper ]
N9X-0.394 [ Wrench it sideways 0.394" until it breaks ]
N10Y-0.278 [ Wrench it the other way ]
N11X3.632 [ decide whether to laugh or cry ]

BTW the fed of 1" per minute was obviously drawn out of a hat as the displayed feed was 4" per min.

So forgetting the inch metric mars lander bit, you get one tool, no depth of cut per pass, feeds determined by a cut of the cards and a chance to become Hertel's best customer but hey wait, it is free after all.............

.

Michael Moore
06-14-2008, 03:32 PM
I'm not sure why MecSoft leaves that old and severely limited program available as it sure doesn't serve as much of a showcase for VisualMill/RhinoCAM. I think FreeMill is based on V4.0 and 6 is the current version. They do have a demo download of the current version available though I'm not sure what limitations it has.

V6.0 has lots of features I have yet to try out and what I've used so far seems "good enough" for what I've tried to do.

cheers,
Michael

Michael Moore
06-14-2008, 03:46 PM
As to the main topic, some casting alloys appear to have physical properties that are hard to match with wrought alloys (and vice versa).

Which process (machining or casting) that is picked will vary depending on the job and the available tools and how comfortable the people doing that job are with the different tools. If you want more than one complicated part then cutting a plug with the mill so you can make multiple castings makes a lot of sense.

Starting with a near-net casting and avoiding making hogsheads of metal chips also makes a lot of sense to me, as does letting the mill chew out an oversize part from machineable wax instead of spending a lot longer making the part from metal. The latter would seem especially true if you have a lightweight machine that can't quickly remove large quantities of metal but is able to zip through the machinable wax (or foam).

But then you still need to be sure that the casting alloys you have available to you are suitable for the job and that the foundry is able to turn out sound castings.

Horses for courses.

cheers,
Michael