PDA

View Full Version : A mill camera



S_J_H
02-04-2008, 03:41 PM
This has always looked like a neat idea but after seeing the article in the latest Digital machinist I decided to make one this weekend. I picked up a cheap logitech webcam for under 30$ and dug out some scrap hunks of aluminum. Made a new housing with setscrew adjusters to dial it into the spindle centerline.
Well it works pretty darn good. It can be used for measuring, inspection and edge and corner finding,punch mark location and bore location.
If the surface edge is sharp I found it reliable to under .0005" finding both edges of a square corner. Within a .001" on a bore. The sharper the surface edge the more accurate it is. Bevels on the edges make it more difficult but you can still get very close. Probably can find a lot of uses for it. Punch marks look like craters.
http://i109.photobucket.com/albums/n48/S_J_H/x3%20mill/millcam002.jpg
http://i109.photobucket.com/albums/n48/S_J_H/x3%20mill/millcam009.jpg
http://i109.photobucket.com/albums/n48/S_J_H/x3%20mill/millcam010.jpg
Steve

S_J_H
02-04-2008, 03:45 PM
these are 64th graduations. I was able to measure the graduation marks at .006" with the camera.
http://i109.photobucket.com/albums/n48/S_J_H/x3%20mill/millcam007.jpg

This is a #80 drillbit. It is .0135" in diameter.
http://i109.photobucket.com/albums/n48/S_J_H/x3%20mill/millcam008.jpg

Picking up a corner edge. This is Starret precision ground flat stock. Surface finish doesn't look all the great when magnified.
http://i109.photobucket.com/albums/n48/S_J_H/x3%20mill/millcam005.jpg

tattoomike68
02-04-2008, 03:53 PM
Thats trick , its a optical comparator now.

very cool.

miker
02-04-2008, 04:43 PM
Nice job Steve. Which software are you using?

The housing looks a bit like a Video Head Drum out of an old VCR.

Rgds

DICKEYBIRD
02-04-2008, 04:50 PM
Wow, that's awesome! I don't subscribe to DM so haven't seen the article....does the stock webcam focus that close or do you have to make your own macro lense? MORE DETAILS PLEASE! :)

S_J_H
02-04-2008, 04:57 PM
Mach3 has a video window for a camera and you can also use this software as a stand alone and as of right now it is free-
http://www.miketreth.mistral.co.uk/centrecam.htm

Yes, the camera focuses that close on it's own. I am not using any other lens with it. It will focus to just about a 1/8" distance. This one has a manual focus ring. Just a cheap webcam from the local dept store.:)

Steve

Mcgyver
02-04-2008, 05:00 PM
hey Steve thats really neat! to .001? terrific. now you need to make an adapter to use it in the tailstock for 4jaw centrering!

how did you align the camera with the spindle, and are cross hairs in the software?

IOWOLF
02-04-2008, 05:13 PM
IMHO, that Beats a laser all to hell.

Well done.

John Stevenson
02-04-2008, 05:25 PM
Some old links on the same subject from 2006.

http://bbs.homeshopmachinist.net/showthread.php?t=21773&highlight=webcam

http://bbs.homeshopmachinist.net/showthread.php?t=23544&highlight=webcam

These posts never got updated that after doing some work with Art Fenarty at Mach3 and Brain Barker a new OCX file was written to allow a better video window to be used in mach.

That's the one Steve is showing on his screen.

[Edit]
For anyone wanting to try this the demo version of Mach will run the videao window without registration.


.

S_J_H
02-04-2008, 06:24 PM
As John said the idea has been around for some time. The webcams seem to be getting better too. I tried this a while back with an old webcam and the video was just to fuzzy so I discarded the idea until I saw the pics in the Digital machinist mag which looked very clear. It's quite clear with this newer webcam.
The cross hairs are part of the software. Centrecam also lets you make circles. So you can locate in on any sized hole in your work.
To adjust the camera I just rotate the camera in the spindle while focusing on a cross mark on the table. Adjust the setscrews just as you would a 4 jaw chuck.
I have one of those lasers, yes this camera is a heck of a lot better and useful.
For critical work edge and bore finding I'll stick with conventional methods. But I can see this can have a lot of uses plus it is fun to see your work and tooling magnified.
I had a square edge on the table. I located a corner edge with the camera using the cross hairs. I removed the camera and installed a ground .375" pin in a collet. I moved the table exactly .1875" in each direction and the pin did not contact the part. However I could not fit a .0005" shim between the part and pin on either edge. That's pretty darn good edge finding IMHO without ever physically touching off the edges.
yeah I could see a few uses for this on a lathe as well!

Steve

BobWarfield
02-04-2008, 06:36 PM
I had seen these before, but S_J_H, your housing is so trick it's much cooler!

Best,

BW

gzig5
02-04-2008, 09:21 PM
Very Nice.
Unless you go wireless though, you might want to make sure the spindle motor is locked out while using that.:D

Evan
02-04-2008, 10:39 PM
Very cool. Gotta make one.

darryl
02-04-2008, 10:42 PM
Nah, gzig5, that's the quick disconnect function. :)

That's pretty cool Steve.

wierdscience
02-04-2008, 10:45 PM
Really cool Steve,definately come in hand for things not to easy to use traditional methods on.

BigBoy1
02-05-2008, 09:43 AM
I sure wish I was more versed in electronics. I have enough trouble keeping the computer up and running so I can visit this site. About the only electronic gadgets I have in my shop are the radio and DRO! The guys who work with electronics are really Wizards. They somehow know how to put the "Magic Blue Smoke" inside of the elecronic components. However, when I release the "Magic Blue Smoke" from the components, they stop working. I guess the "Magic Blue Smoke" is what makes the components work!

Bill

Evan
02-05-2008, 11:46 AM
They somehow know how to put the "Magic Blue Smoke" inside of the elecronic components.

Nah. That's the people that make the stuff. The best us elec techs can do is learn how to avoid releasing it too often. BTW, we used to call that smoke "brown smell". :D That's the smell you get when a transformer burns up.

QSIMDO
02-05-2008, 02:10 PM
Can anything like that be done on a manual machine?
Connect it to a computer?

Evan
02-05-2008, 02:18 PM
Certainly, no reason why not. All you need is a handy obsolete laptop.

aboard_epsilon
02-05-2008, 02:29 PM
would be nice if you had a set of adjustable parallel lines on the software ...that way you could put a lump of steel in the vice ..........and converge the lines to find centre without doing any measuring .

all the best.mark

John Stevenson
02-05-2008, 03:41 PM
would be nice if you had a set of adjustable parallel lines on the software ...that way you could put a lump of steel in the vice ..........and converge the lines to find centre without doing any measuring .

all the best.mark

Take a look at Mikes software list in post #6

For anyone that wants to try this Mach will run the web cam in demo mode.

.

aboard_epsilon
02-05-2008, 03:54 PM
what's the minimum comp power it needs

think i have about 80 MB of ram on my laptop
166 processor and windows 98 second addition

all the best...markj

Evan
02-05-2008, 04:14 PM
There are a variety of free on screen measuring tools that you can use in conjunction with a web cam image. They don't depend on the webcam software at all. Here is an example of me using 4 different tools to measure features of this webcam image of the front of one of my digital cameras at close range. I have even more tools but they can't be screen captured as they use that function.

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

Here are some free tools to get started with. Some will run on 98 but not all. If it says something about dot.net then it is xp or higher.

http://www.freewareplanet.net/measure.php

rotate
02-05-2008, 05:15 PM
Thanks Evan, those are excellent set of programs. I've been using Corel draw to do something like that but it's doesn't always work very well.

Smokedaddy
02-05-2008, 09:29 PM
This has always looked like a neat idea but after seeing the article in the latest Digital machinist I decided to make one this weekend.

Steve,

I saw this as well and was thinking about making one. Glad to see someone making a new body. How much time would you guess you have in your project.

-SD:

lazlo
02-05-2008, 11:04 PM
This has always looked like a neat idea but after seeing the article in the latest Digital machinist I decided to make one this weekend. I picked up a cheap logitech webcam for under 30$ and dug out some scrap hunks of aluminum. Made a new housing with setscrew adjusters to dial it into the spindle centerline.

Steve, that looks fantastic! Very professional.

I enjoyed Arnie Minear's article, but I wasn't really fond of the machining aesthetics (the hemisphere mount).

One comment: the Logitech QuickCam, and the Micro Innovations web cam in Arnie's article, has a QCIF (352 x 288) == 100,000 pixel sensor. Looking at the specs and the shape of the web cam, I think the Micro Innovations web cam is the OEM camera that Logitech puts their QuickCam name on.

The Intel QX3+ kiddy microscope that I've been using as a cheap optical comparator has the same resolution (QCIF = 352 x 288) == 100,000 pixels.

The latest webcams, including the Logitech Ultravison have "1.2 Megapixel" sensors, and glass (Karl Zeiss -- yeah, right :o) lenses. I'm sure a lot of that is marketing bullsh!t interpolated resolution, but I'm guessing the sensor is at least 4 times the resolution of the QuickCam:

http://www.logitech.com/index.cfm/webcam_communications/webcams/devices/241&cl=us,en

http://www.logitech.com/repository/268/jpg/1889.1.0.jpg

Assuming they didn't muck with the focal length, one of the newer web cams should make for a very high resolution optical comparator.

I'll check some prices...

lazlo
02-05-2008, 11:09 PM
I sure wish I was more versed in electronics. I have enough trouble keeping the computer up and running so I can visit this site.

You don't need to know anything about electronics for this project Bill. You're just cracking open the plastic case on a web cam, and mounting the guts in a holder that you mount in the mill spindle.

You could just as easily drilled a screw into the back of the case and put it in a drill chuck :D

lazlo
02-05-2008, 11:17 PM
So here's a 2.0 Megapixel web cam, the Creative Live! Cam Optia. $75 at Gateway:

http://www.everythingusb.com/creative-live-cam-optia-af-14195.html

http://www.everythingusb.com/images/list/creative-optia-af-webcam-prod.jpg

S_J_H
02-05-2008, 11:58 PM
would be nice if you had a set of adjustable parallel lines on the software ...that way you could put a lump of steel in the vice ..........and converge the lines to find centre without doing any measuring .

centrecam has that feature. This is a 123 block and the 2 adjustable lines have located the center between the holes.-
http://i109.photobucket.com/albums/n48/S_J_H/x3%20mill/millcam011.jpg

Here again is a 123 block and using the circle function I located in one of the holes in the 123 block.
http://i109.photobucket.com/albums/n48/S_J_H/x3%20mill/millcam012.jpg

This is cool, this is a hair I plucked from my head on a piece of paper,lol. I mic'd it at .0025". Note you can see the paper fibers as well.
http://i109.photobucket.com/albums/n48/S_J_H/x3%20mill/millcam013.jpg

Smokedaddy, I bought the camera on friday and had it working by sunday. No more than a full days work on the housing. I don't plan things out. I build on the fly from a vision. Now sometimes it works out and sometimes I make scrap. This time it worked out.:D
Steve

Rich Carlstedt
02-06-2008, 12:18 AM
OK Guys. how do you center the screen to match the concentricty of the holder when you first set it up?
Is it a mechanical adjustment, or electronic ?

Rich

S_J_H
02-06-2008, 12:28 AM
Rich, the camera is a 2 piece unit. 4 setscrews dial it in perfectly to the spindle center. Just install the camera unit in a collet, focus in on a cross hair mark on the table. Rotate and adjust the camera until it shows no runout. Exactly like adjusting a 4 jaw chuck. Takes but a minute or so.

Steve

Smokedaddy
02-06-2008, 12:29 AM
Thanks,

In any case, thanks for the posting and information. I just wanted an "idea" on how much time you spent on this cool project. Isn't the knurled piece another individual component? Do you have a lens cover in that piece to keep dust off the sensor? Just curious exactly how you made it.

Cheers,
-SD:

Dawai
02-06-2008, 01:01 AM
Excellent idea there, good application.

With a cnc thou, you no longer have to "center find" you can mount the camera on the mill head with a known offset, and step into exact center to drill or? . NOT have to chuck it up and move it around. A nifty lil sock to cover it up while not in use is nice tho. Vibration did not affect the logitech I have.

I put velcro on my sock.

I had removed mine to do some coolant cuts.. it's on the shelf still.
OK.. with a laser, since the camera is offset, angle the camera to center laser and you have a scanner. Simple triangulation. Also been done.

Someone said Art of Mach3 was going to incorporate it into his new software. Cool, last time I was in his forum on Yahoo he was trying to use a industrial distance measure and stepping about.

Your camera looks so much better than mine. much more resolution. I like.
Might be time for a upgrade here.

S_J_H
02-06-2008, 10:00 AM
David, this camera needs to be very close to the work or part to focus in sharply. Having it offset permanently would have issues. But an adjustable lowering/raising mount on the side of the mill head could easily be made so it would be out of the way while machining.
Hmmm...
To focus in on the Hair shown in the above pic it was only about 1/8" from the surface.
Steve

aboard_epsilon
02-06-2008, 12:13 PM
how about having ...sort of mounted on a precision half pipe section ....on a swinging arm ...that sort of half rapped its self around the spindle extension ....when brought in .....

if the cam was then mounted a few inches below where the collet went in ...

you could even use it with tooling in the machine

then all you have to do is swing it away to the side when not in use ...and its there and ready in an instant when you need it .

all the best...markj

Evan
02-06-2008, 12:34 PM
I was thinking more along the line of a low power spherical mirror mounted at a 45 degree angle so that it surrounds the spindle. It can be made of aluminum and having a central obstruction (the spindle) doesn't affect it's ability to image the area directly below the spindle any more than it affects a newtonian telescope or a Maksutov lens system.


I feel an experiment coming on... :D

TGTool
02-06-2008, 11:56 PM
I'm also much taken by the ideas and pictures of using a web cam this way. It occurs to me that this also could be one solution to the problem of raising the head on round column mill-drills while maintaining position. OTOH, maybe not without a telephoto to focus the same spot from a greater distance.

On thinking about the alignment issue I now have a doubt about the geometrical accuracy of the setup. Rotating the spindle and adjusting the camera position only guarantees centering accuracy at that specific distance, doesn't it? That is, if the camera were mounted at an angle to the centerline of the spindle, you would always be able to zero the crosshair offset, but raising or lowering the camera would require a re-zero unless both X-Y offset and centerline coincidence were accounted for.

Jan

Evan
02-07-2008, 12:56 AM
That is why my laser telescope collimator has two triads of set screws.

http://vts.bc.ca/pics/coll1.jpg

The collimator, or something very much like it can be used to do the same job as the web cam if fitted to the drawbar bore. Of course that means the tooling must be removed.

Every device that needs to be oriented in some particular way has six degrees of freedom to be concerned with. There is rotation in X, Y and Z (pitch, roll and yaw) and there is translation in X, Y and Z ( movement along the axis). The web cam can be off target in only 4 possible ways since both movement in Z and rotation around Z doesn't change the targeting unless it is off target in another axis.

With the laser collimator one triad of set screws is in the same axial position as the laser emitter so it is used to change centering, that is translation. The other three setscrews alter pointing (rotation) as the laser is free to pitch etc within the first set of screws. Of course there is interaction between the two adjustments so the adjustment requires an iterative approach in which some of the error at each end is nulled and then the process is repeated until the desired level of accuracy is reached.

S_J_H
02-07-2008, 12:58 AM
Jan,
I think you have a point that I have over looked! If the lens is at an angle It would be possible to offset with the adjuster screws such that it would appear to be dead on, but it's really not.
I guess the way to check this is to focus in at close range and dial in the camera real good on a cross hair on a 360 rotation. Then raise the camera and rotate it again 360' and see if it no longer is dead on with the cross hair.
If it is off between the 2 points, then I need to add tilt adjusters. That will not be hard to add to the camera housing.
Steve

Evan
02-07-2008, 01:01 AM
You only need to rotate the camera 90 degrees. Since it can only be off in X and Y that will take care of those possibilities.

Rich Carlstedt
02-07-2008, 01:30 AM
This is really a neat project....and thanks to all
for sharing great info

Rich

S_J_H
02-07-2008, 09:45 AM
You only need to rotate the camera 90 degrees. Since it can only be off in X and Y that will take care of those possibilities.

Yep you are right. I'll check this later and correct if needed. My guess is that any angle error is quite low as the camera only has a working distance of between 1" and 1/8" from the surface. I already verified it's edge finding on a corner to be under .0005" with a ground pin so there can't possibly be much of an error from any angle at the lens.

Steve

A.K. Boomer
02-07-2008, 10:35 AM
Good quality work, (and very nice knurl job too)

Someone brought up wireless as a possibility also? but before that I was wondering about a housing that just holds the lens guts --- and then running their wires to a separate circuit box just to keep whats mounted in the spindle small, I have no idea if these guts would let you do that or not,

I cannot believe how close you can get -- and I also cant believe the lens does not need any mods. Great work, If you wanted to you could crank them out and sell a Million as its very professionally done.

oldtiffie
02-08-2008, 07:27 AM
This really is a very impressive thread.

Thanks to the OP and all other posters.

I was that impressed that I went and bought a Logitech Pro 2 mega-pixel web cam today.

http://www.logitech.com/index.cfm/webcam_communications/webcams/&cl=us,en

http://www.logitech.com/index.cfm/webcam_communications/webcams/devices/3055&cl=us,en#

http://www.logitech.com/index.cfm/69/3231&cl=us,en

Its minimum focal length is 100mm (4").

I will have to wait until my XP computer if fixed (its has a "hardware problem" somewhere) in about a week or so. The (ie my "other") computer I am using now is a W2K machine - very good too. The specs for the camera require XP or Vista.

I wonder why it didn't specify W2K - anyone with any ideas?.

I wouldn't worry at all about the camera axis being at an angle and describing a conical section - ie a cone - and "problems" with varying heights above the work/table.

If it is adjusted in the "X" and "Y" planes - by rotating the mill quill spindle - so that it is centred at work level, a change of height should not matter as it can be re-adjusted. The cone/s will always be concentric with the quill spindle.

What it WILL do is point out whether the mill column on which the milling head rises or falls is out of square with the mill table.

I have seen no other easy method of identifying and assisting in this adjustment or verification.

No - I am not talking about "tram" of the mill quill with regard or relative to the mill table.

It is quite possible, say for example, for the mill column to be out of square with the table and still have the tram read as accurately as you can get it.

How so?

If the mill column is leaning say 1 degree to the left, the mill head can be pivoted clockwise 1 degree and apparently correct this "tram error".

Such is not the case.

All that has happened here is that the "column tilt" has been countered by the tram correction.

If a laser or web-cam were put in the quill spindle and the mill head moved up and down the column, the light/axis would move left-right etc. on the table as the apparently correctly trammed milling head was raised and lowered.

Sort of answers some of those vexing problems and questions about incorrect performance on an apparently correctly set up mill doesn't it?

If I can get the web cam to identify a 0.001" variation of centre on a mill table in a say 18" rise/fall of the milling head I'd be as pleased as ....................... (fill in your own rude words here). I'd be satisfied with an accuracy of 0.005" over that distance!!.

All of this benefit would be in addition to that as in previous posts.

Another thing I want to have a look at another time is just how square the "X" and "Y" travels are to each other. I have got most of the ideas on that too. It is always just taken for granted that it is 100% correct too - but is it? How do or would you know? How would you check or verify it? How would you correct or allow for it?

But as I said - that is for another time in the future.

This web cam has a lot of potential uses.

Many thanks to all.

Evan
02-08-2008, 09:11 AM
What it WILL do is point out whether the mill column on which the milling head rises or falls is out of square with the mill table.
An excellent point. This is something that occured to me but I didn't mention because I am not familiar with the methods or even the possibility of adjusting the average mill for this condition. When I designed my mill I provided for such adjustment. The vertical rod ways are captured top and bottom in large blocks of aluminum that are bolted to equally strong support features. Those blocks can be shifted once bolts are loosened by opposing pairs of setscrews, both sides, top and bottom, in order to adjust the ways for perpendicularity with the X-Y plane.



Another thing I want to have a look at another time is just how square the "X" and "Y" travels are to each other.

That I measured just a few days ago. I built an engraving sub table and scribed it with the mill in a grid of 1 cm squares. When done I checked the squareness with a swiss master precision level. A further check is to install a dial gauge in the spindle and simply run it along the sides of the level in each axis after aligning the level to one of the axes.

S_J_H
02-08-2008, 10:00 AM
Tiffie, looking forward to your results with that camera. I am not sure how well the higher end models focus up close though? The one I am using will focus at under .5cm as seen in the pics in this thread.
I am not going to bother with any angle error. It's not worth the time as it takes 30 seconds to adjust this thing pretty much dead on. I have already noticed that it needs a tiny adjustment when installing it each time on the mill if I want it dead nuts. Probably the import 3/4" R8 collet is not all that exact.

Steve

TGTool
02-08-2008, 10:55 AM
Tiffie,

You're going to have to be sure the camera angle is square with the table in order to verify machine Z axis squareness. How you gonna do dat? If the camera is tilted at all, moving in Z will show an apparent shift in X or Y (or both) and there's no easy way to distinguish whether the shift is caused by the mill column out of square or the camera out of square. One or the other has to be correct in order to verify the other.

X-Y squareness using the camera might be checked by scribing two lines with the mill travel. For instance, with a scriber in the spindle (no rotation required) drag a scratch line in X with Y locked. Then scribe a line in Y with X locked. Now rotate the plate on the mill table by 90 degrees and align the scribe mark with one axis again using the camera to verify coincidence at both ends. Now when you track the other axis you'll see the double angle error in the X-Y squareness of the machine. Same thing you might do with an indicator and a master square and perhaps a higher degree of accuracy.

lazlo
02-08-2008, 11:32 AM
I was that impressed that I went and bought a Logitech Pro 2 mega-pixel web cam today.

Cool! -- That's the web cam I recommended on page three. We can compare notes :)

I ordered the Logitech QuickCam Pro from Amazon when I posted that last Tuesday -- it was $81 with free shipping, and there's a $15 rebate from Amazon on top of that:

http://www.amazon.com/Logitech-QuickCam-Pro-for-Notebooks/dp/B000RZNI4S/ref=pd_bxgy_e_text_b


If I can get the web cam to identify a 0.001" variation of centre on a mill table in a say 18" rise/fall of the milling head I'd be as pleased as ....................... (fill in your own rude words here). I'd be satisfied with an accuracy of 0.005" over that distance!!.

Tiffie, these 2.0 Megapixel cameras have 20 times the resolution that Steve and Arnie Minnear's web cams have: 2 million pixels for the Logitech QuickCam Pro 9000 versus 100,000 pixels for the normal QuickCam:

Higher megapixel performance: A true two-megapixel sensor. Up to eight-megapixel photos (enhanced)*.

If Steve is picking up an edge to within 0.0005" with 100,000 pixel resolution, the 2.0 Megapixel web cams should be amazing.

Like I posted earlier in the thread, the QuickCam Pro's have Carl Zeiss glass lenses. I realize that's not a real Zeiss lens, but most web cams have plastic lenses, so you would hope the glass lens would improve the overall optics.

lazlo
02-08-2008, 11:40 AM
Tiffie, looking forward to your results with that camera. I am not sure how well the higher end models focus up close though? The one I am using will focus at under .5cm as seen in the pics in this thread.

That's a good question -- Logitech's description of the QuickCam Pro says this:

http://www.logitech.com/index.cfm/webcam_communications/webcams/devices/3055&cl=us,en

Carl Zeiss® optics and autofocus keep images razor-sharp, even in extreme close-ups.

I'm going to cobble-up a holder for the mill and make sure I can focus on fine detail before I crack open the case. Amazon is really good about returns, but they won't take a web cam in pieces :)

Evan
02-08-2008, 05:59 PM
I just tried one of my web cams. It's 1.3 true megapixel and made for the job as it comes in a pancake low profile case that unplugs from the stand. You then replug the cable directly into the camera minus stand. For a quick trial I glued a supermagnet to the back and stuck it to a tool in the spindle. When hooked to the computer the picture is excellent and it will focus to within a few inches. It also has a zoom function. While I was playing with this Mach 3 crashed hard, taking the system with it. I rebooted and tried again, this time using the built in viewer in Mach 3. After a minute or two the same thing happened. In fact, after some more dinking around with it it appears that the camera and Mach 3 cannot be in operation or even plugged in at the same time on the same computer.

The computer is totally stable including with web cams. I use it as my archive file server and it runs 24/7. It goes for months without a reboot and also is responsible for monitoring the back door webcam as well as serving the images over the LAN.

John Stevenson
02-08-2008, 06:24 PM
Evan,
Not sure what is happening here but we run a web cam at shows where we can't cut metal due to H&S Nazi's.

We set it up on a job like the spiral gears and run the machine with the web cam showing the tool path.

We can run for hours like this.

.

S_J_H
02-08-2008, 06:28 PM
But what about the magnification with the better auto focus webcams?
This is from one of Oldtiffies Links-

When the image sensor detects a loss of sharpness in objects within the viewing frame, it quickly and precisely readjusts the lens, ensuring a sharply defined image—even in extreme close-ups where the subject is only 10 centimeters from the lens.
In the picture of the single hair I was at around .3cm away.
I have no issues running this webcam and the mach3 video window at the same time. Works perfectly. I can run centrecam along with mach3 as well. I'm using an older desktop pc in the shop.

Steve

Evan
02-09-2008, 04:28 AM
I think I found the problem. The video card is an ATI Radeon 9200 and on one of the crashes it finally told me that there may be a driver issue. I switched it for the Microsoft ATI driver and it seems fine.

oldtiffie
02-09-2008, 06:09 AM
Tiffie, looking forward to your results with that camera. I am not sure how well the higher end models focus up close though? The one I am using will focus at under .5cm as seen in the pics in this thread.
I am not going to bother with any angle error. It's not worth the time as it takes 30 seconds to adjust this thing pretty much dead on. I have already noticed that it needs a tiny adjustment when installing it each time on the mill if I want it dead nuts. Probably the import 3/4" R8 collet is not all that exact.

Steve

Thanks Steve.

Thanks too for the concept. All I did was to extend it a little further as I could see the genesis of a solution for checking the verticality of a mill column with respect to the table.

I checked your links as well as the posts from lazlo and Evan.

It went on from there.

I am still shaking my head about the concepts and achievement in your CNC lathe and Evan's CNC self-designed and home-built fabulously accurate machines.


I have a small-ish problem though. I don't have a computer in my shop to allow me to set the web cam up on my mill in live time. I can see that I am being remorselessly and inevitable drawn into the black hole of putting good lap-top computer in the shop and installing some of the software mentioned by you, John Stevenson, lazlo and Evan.

I might even have to "get my finger out" and install the DRO's to my mill at least.

I will try installing the web-can software on my W2K computer and see if it works. The cam spec calls for either XP or Vista - I have no idea why.

Anyone able to tell me please?

I thought that XP was very similar to W2K.

My XP computer has a few problems that are being sorted out. I hope to have it back next week after which I have to re-install everything from scratch and set it to work.

Anyway, if/when I get the web-cam software installed I will try it on a stand near the computer inside the house.

My main criteria for the web-cam I bought was the Carl Zeiss glass lens as well as the improved resolution and a longer focal range and hopefully adequate optics.

I want to chase up and install the software you, Evan and lazlo use and recommend.

I am not convinced that the camera needs to be too closely set up on centre or with its axis parallel to that of the mill quill. I will address that in another post.

This is going to take some time.

oldtiffie
02-09-2008, 08:09 AM
Thanks TGT.

Good post with lots to think about.




Tiffie,
you're going to have to be sure the camera angle is square with the table in order to verify machine Z axis squareness. How you gonna do dat? If the camera is tilted at all, moving in Z will show an apparent shift in X or Y (or both) and there's no easy way to distinguish whether the shift is caused by the mill column out of square or the camera out of square. One or the other has to be correct in order to verify the other.

In the ideal state, I agree with you.

But as I am not in the ideal state I have to work around things a bit to get an acceptable and hopefully/probably a good/adequate result.

If the camera lens axis is both parallel to and concentric with the mill quill both would be identical and on the same centre-line.

Now bear with me here.

That being the case there should be no deviation from a static/"fixed" image as the mill quill is rotated through 360 deg (yep - by hand and not under power!!) and the mill head run through its high-low limits of travel.

The centre line then would describe an arc and therefore a radius of zero and a cone of zero diameter would be produced.

The concurrent centre of the camera lens and the image it produced would not "move" or be displaced on the computer monitor either by the quill rotation or the mill head vertical movement.

Now if the camera lens were parallel to but off-set from the quill centre, the lens would describe a circle and the lens centre would be projected to described a cylinder such that the image on the monitor when the quill/camera were rotated that would appear to be eccentric to a point on the "target". The centre of the the projected quill and lens would describe a similar eccentric circle on the target. This could be measured and allowed for and the alignment calculated and set/recorded.

If the camera lens was not parallel to the quill the optical centre of the lens when the quill was rotated would describe a cone. If the "point" of the cone were co-incident with the target, there would be no apparent nor real misalignment and the alignment could be accepted without further calculation or adjustment.

If the "point" of the "cone" were not coincident with the "target" - ie the "cone" was above or below the target - the lens centre would again describe a circle on the target. If the centre of the eccentricity were reduced to zero by moving the mill table and the target the centre of the cone would be coincident with the target. The off-set of the centre of the cone and the target can be aligned by equalising/balancing it by adjustment of the "X" and "Y" lead-screws while rotating the quill.

I realise that this is a "rough and dirty" solution but it should work.

In every case, the ultimate adjudicator is the human eye which can detect and eccentricity to a very high order of accuracy which is made easier by the highest resolution of the web-cam lens/optics.



X-Y squareness using the camera might be checked by scribing two lines with the mill travel. For instance, with a scriber in the spindle (no rotation required) drag a scratch line in X with Y locked. Then scribe a line in Y with X locked. Now rotate the plate on the mill table by 90 degrees and align the scribe mark with one axis again using the camera to verify coincidence at both ends. Now when you track the other axis you'll see the double angle error in the X-Y squareness of the machine. Same thing you might do with an indicator and a master square and perhaps a higher degree of accuracy

Evan has the correct answer for the checking of the mill "X"/"Y" "squareness" - a Precision Frame Level.
http://www.hareandforbes.com.au/sample_2/Catalogues/Metalworking/42.html

Mine is a/the 200mm model.

This level has 4 "scraped-in" sides flat and square to each other to an accuracy of 0.02/1,000mm which is 1:50,000 or 1 in 50,000 which is 0.004mm per 200mm which is 0.0002" (2 "tenths") over the 8" length of any side.

I would lay this on the table, preferably with no or minimal clamping, and align it to my say "X" axis with a very good Test Dial Indicator (TDI) to as near as I could get it after the "X" clamps were lightly applied. I would then set the TDI to the "side" faces of the Precision Frame Level and traverse the "Y" slide/lead-screw and check for deflection on the TDI. Any deflection recorded is equal to the "out of square" of the "Y" slide to the "X" slide.

TGTool
02-09-2008, 09:26 AM
Evan has the correct answer for the checking of the mill "X"/"Y" "squareness" - a Precision Frame Level.
http://www.hareandforbes.com.au/sample_2/Catalogues/Metalworking/42.html

Mine is a/the 200mm model.

This level has 4 "scraped-in" sides flat and square to each other to an accuracy of 0.02/1,000mm which is 1:50,000 or 1 in 50,000 which is 0.004mm per 200mm which is 0.0002" (2 "tenths") over the 8" length of any side.

I would lay this on the table, preferably with no or minimal clamping, and align it to my say "X" axis with a very good Test Dial Indicator (TDI) to as near as I could get it after the "X" clamps were lightly applied. I would then set the TDI to the "side" faces of the Precision Frame Level and traverse the "Y" slide/lead-screw and check for deflection on the TDI. Any deflection recorded is equal to the "out of square" of the "Y" slide to the "X" slide.

Ah, wonderful! If you have a frame level then you've got the solution to Z-axis checking as well. Same process. If the spindle and table tram correctly, then the same DTI test going vertically up the standing square will read any error in head, quill, or knee, whichever movement you're checking.

I'm still thinking about how you've described the cone/cylinder/translation for checking squareness but I'll sketch out the possible scenarios and see if I'm confident there's a solution there.

lazlo
02-09-2008, 11:02 AM
I will try installing the web-can software on my W2K computer and see if it works. The cam spec calls for either XP or Vista - I have no idea why.

I thought that XP was very similar to W2K.

It is, and the Logitech web cam should run on Win2K.

The driver framework is the same for Windows XP and Win 2K, but the Logitech is probably using the DirectX streaming audio and video stack for the video and audio capture (there's a microphone on the web cam).

I just checked, and the latest version of DirectX 9.0c specifically supports Win2K. I'd recommend you update to DirectX 9 before you install the Logitech drivers:

Supported Operating Systems: Windows 2000;...

http://www.microsoft.com/downloads/details.aspx?FamilyId=2DA43D38-DB71-4C1B-BC6A-9B6652CD92A3&displaylang=en

oldtiffie
02-09-2008, 09:05 PM
Evan has the correct answer for the checking of the mill "X"/"Y" "squareness" - a Precision Frame Level.
http://www.hareandforbes.com.au/samp...orking/42.html

Mine is a/the 200mm model.

This level has 4 "scraped-in" sides flat and square to each other to an accuracy of 0.02/1,000mm which is 1:50,000 or 1 in 50,000 which is 0.004mm per 200mm which is 0.0002" (2 "tenths") over the 8" length of any side.

I would lay this on the table, preferably with no or minimal clamping, and align it to my say "X" axis with a very good Test Dial Indicator (TDI) to as near as I could get it after the "X" clamps were lightly applied. I would then set the TDI to the "side" faces of the Precision Frame Level and traverse the "Y" slide/lead-screw and check for deflection on the TDI. Any deflection recorded is equal to the "out of square" of the "Y" slide to the "X" slide.




Ah, wonderful! If you have a frame level then you've got the solution to Z-axis checking as well. Same process. If the spindle and table tram correctly, then the same DTI test going vertically up the standing square will read any error in head, quill, or knee, whichever movement you're checking.

I'm still thinking about how you've described the cone/cylinder/translation for checking squareness but I'll sketch out the possible scenarios and see if I'm confident there's a solution there.

Thanks TGT - another really good post.

Yep. The reasons I bought the square was to check the verticality of the mill column as well as the squareness of the "X", "Y" and "Z" ways/slides.

The square only really provides for using one side at a time -I wanted to do both at the same time - mount 2 magnetic bases on my mill head and run it up and down and check it for squareness in the "Z" plane to the table in the "X" and "Y" planes.

A "natural " square is really the way to go as it will do it easily.

I just side-tracked myself with the web-cam and electronics I guess.

For those that might not be aware of it, a "natural square" is a hollow cylinder that is as near round, straight and parallel as is practical that has been accurately mounted in a lathe and the ends accurately faced off.

The ends are then for all intents and purposes both flat and square to the cylinder axis.

If the "square" is stood on or clamped to the mill table the face of the cylinder is perpendicular to the table to a very high order of accuracy.

It is ideal for checking "squareness" or verticality to the table both as reference.

For my purposes, precision ground steel tube - such as hydraulic actuator piston rod would be ideal. I would prefer say, 100mm OD x 300mm long (200mm and 100mm long as well).

It is also ideal for clamping to for machining purposes. This is far and away the best way to machine 2 faces on an angle plate to be square to each other - and similar uses - all with the mill in correct "tram" and with fly-cutters (my preferences).

Works really well on a surface or Tool & Cutter grinder as well.

But after this reality check, I have to say that on reflection that the "natural square" is looking better for setting/checking the verticality of the mill column.

Thanks for giving me a much-needed "jolt" and for getting me "back on track".

oldtiffie
02-09-2008, 09:13 PM
I will try installing the web-can software on my W2K computer and see if it works. The cam spec calls for either XP or Vista - I have no idea why.

I thought that XP was very similar to W2K.


It is, and the Logitech web cam should run on Win2K.

The driver framework is the same for Windows XP and Win 2K, but the Logitech is probably using the DirectX streaming audio and video stack for the video and audio capture (there's a microphone on the web cam).

I just checked, and the latest version of DirectX 9.0c specifically supports Win2K. I'd recommend you update to DirectX 9 before you install the Logitech drivers:

Supported Operating Systems: Windows 2000;...

http://www.microsoft.com/downloads/details.aspx?FamilyId=2DA43D38-DB71-4C1B-BC6A-9B6652CD92A3&displaylang=en

Thanks lazlo.

You really are an asset and a great help.

I will do that later - hopefully today - and will keep you posted.

Evan has a great very pertinent thread running at:
http://bbs.homeshopmachinist.net/showthread.php?t=27492
that "ties in" very well with this thread and topic.

trukr
02-09-2008, 10:47 PM
SUPER Thread !

Like your design, steve. Simple in construction and function. I went out and bought my web cam the day after reading the article in DM. Then downloaded Centrecam and started playing around. The trial version I have of Mach3 does not have all the goodies I see in the side bar of your screen shot? Is that only on the full version, or do I have an old one? I had been trying to come up with a design to make centering easy, was thinking of dovetail ways etc.(way too complicated!) but your idea is the solution - much better!! Although the point Jan raises still needs to be addressed. Still processing that one.

With the camera so close to the subject the camera blocks ambient light . Wonder if adding a built in light source, LED's or fiber optics would be worth the effort.

Thanks all - good conversation.

Evan
02-09-2008, 11:00 PM
The problem with my web cams and Mach 3 is definitely the ATI video card/driver. It even crashed with the Microsoft driver. I ripped out the Radeon 9200 card and put in a Gainward Nvidia G-force MX-4. Works great.

oldtiffie
02-10-2008, 06:52 AM
Links to pics of my frame level and web cam are at:
http://i200.photobucket.com/albums/aa294/oldtiffie/Frame-level1.jpg
http://i200.photobucket.com/albums/aa294/oldtiffie/Frame-level2.jpg
http://i200.photobucket.com/albums/aa294/oldtiffie/Web-cam1.jpg
http://i200.photobucket.com/albums/aa294/oldtiffie/Web-cam2.jpg

The function and accuracy of the frame level is self-evident - see previous posts.

The web-cam is a high resolution item. I am reluctant to remove the "innards" of the web-cam from its cover/body. I will just have to think of some way to mount it in the mill to be adjustable to the quill centre. Spring mounting is the first method I'd thought of - but needs more thought.

oldtiffie
02-10-2008, 07:08 AM
But what about the magnification with the better auto focus webcams?
This is from one of Oldtiffies Links-

When the image sensor detects a loss of sharpness in objects within the viewing frame, it quickly and precisely readjusts the lens, ensuring a sharply defined image—even in extreme close-ups where the subject is only 10 centimeters from the lens.
In the picture of the single hair I was at around .3cm away.
I have no issues running this webcam and the mach3 video window at the same time. Works perfectly. I can run centrecam along with mach3 as well. I'm using an older desktop pc in the shop.

Steve

Thanks Steve.

That very matter was of concern to me too.

My minimum focal distance is 100mm where-as yours is less than 4mm.

That is a 25:1 - or better/larger - ratio.

I may get better resolution but I am almost certain that I will not get same level of magnification that you have.

As I don't have computer in the shop the best I can do - when I get the software installed - is to try it on the office desk near this computer.

Just as a matter of interest, I thought that I'd use a drafting scale rule - say a 1:100 or 1:1 with graduations at 1mm.

These scale rules are engraved very accurately. All lines etc. are very "sharp". The black figures etc. on a white back-ground are the best black-white contrast about.

Evan seems to have hit on the lighting solution at his thread at:
http://bbs.homeshopmachinist.net/showthread.php?t=27492

Evan
02-10-2008, 08:16 AM
The solution is simple and effective. I don't have pics to post yet but all one needs is to add a pinhole lens to the camera and enable the low light mode. I have tested it with a cheap 1.3 megapixel web cam and it is able to keep in focus from about 2 centimetres to infinity. I used some adhesive aluminum tape so it probably spoiled the image a bit with traces of glue on the pinhole diffracting the image. I just don't have time to test and play with all these things at once. As it is I only get a few hours sleep per night.

A pinhole lens is easy and cheap to make and has the added benefit of protecting the camera lens. If one wanted to really go the full yardage then removing the camera lens would provide even better results, using only the pinhole. When using only a pinhole lens the magnification ratio is dependent on the distance the pinhole is from the sensor. It's linear and if the pinhole is the same distance as the sensor is wide then the ratio is 1:1.

lazlo
02-10-2008, 11:00 AM
My minimum focal distance is 100mm where-as yours is less than 4mm.

That is a 25:1 - or better/larger - ratio.

I may get better resolution but I am almost certain that I will not get same level of magnification that you have.

Mick, are you measuring that with the Logitech Quick Cam Pro? Or are you reading the specs somewhere?

My package didn't arrive from Amazon on Friday :mad:, but I just opened this week's Best Buy ad, and they have this Logitech Quick Cam Pro (the 2.0 Megapixel camera we're talking about) for $79.99 plus a $15 rebate from LogiTech.

I'm going to run by Best Buy this morning and get a second camera...

http://www.bestbuy.com/site/olspage.jsp?skuId=8380566&st=8380566&lp=1&type=product&cp=1&id=1179165373839

lazlo
02-10-2008, 11:31 AM
OK, here's a deal that's hard to pass up, even if you don't want to make a mill cam :)

Buy.com has the "Logitech QuickCam Deluxe for Laptops," which is the 1.3 Megapixel version of the Web Cam that Mick and I just bought, for $10 after rebate, with free shipping. That model also has the glass Carl Zeiss lens:

http://www.buy.com/prod/logitech-quickcam-deluxe-for-notebooks-webcam-silver-black-cmos-usb/q/loc/111/204247388.html

rotate
02-10-2008, 11:35 AM
You have to be very careful with the resolution on webcams. Often 1.2 megapixels turns out to be "effective" number of pixels. Try to determine what the naitive resolution is. Manufacturers often do not disclose this information for a good reason.

Evan
02-10-2008, 11:40 AM
That seems expensive. I picked up the Microsoft Lifecam recently for $30. It's 1.3 MP I believe but has an exceptionally quiet (no noise) sensor. It will focus to the front of the camera face. Example, the camera pressed to the screen.

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

lazlo
02-10-2008, 12:26 PM
You have to be very careful with the resolution on webcams. Often 1.2 megapixels turns out to be "effective" number of pixels. Try to determine what the naitive resolution is.

I already addressed that issue 2 pages back:


Tiffie, these 2.0 Megapixel cameras have 20 times the resolution that Steve and Arnie Minnear's web cams have: 2 million pixels for the Logitech QuickCam Pro 9000 versus 100,000 pixels for the normal QuickCam:

Higher megapixel performance: A true two-megapixel sensor. Up to eight-megapixel photos (enhanced)*.

lazlo
02-10-2008, 12:37 PM
That seems expensive. I picked up the Microsoft Lifecam recently for $30.

See my post above yours Evan -- you can get 1.3 Megapixel cameras for $10 after rebate.

There's a price premium for the new 2.0 Megapixel cameras. But I figure the time it takes me to make the housing is worth more than the $50 difference in price between the 1.3 and 2.0 Megapixel web cams.

lazlo
02-10-2008, 05:37 PM
I went web cam shopping this morning, and I spent the last two hours jury-rigging "mill cams" on my desk :D

The four web cam I tested, with their actual (not interpolated) resolution:


Logitech Quick Cam Pro for Laptops...........2.0 Megapixel
Logitech Quick Cam Deluxe for Notebooks. 1.6 Megapixel
Logitech Quick Cam Communicate Deluxe. 1.6 Megapixel
Microsoft LifeCam VX-6000.......................1.6 Megapixel


A couple of key points:


All the name-brand web came (Logitech, Microsoft, HP,...) come in desktop and laptop versions. The desktop Web Cams seem to have much better optics, even though they appear to have the same CMOS sensor.

In order to fit the sensor, the electronics, and the lens in such a small space (about the size of a lighter), the focal length is really crushed. All the laptop web cams appear to be optimized for 6 - 36" focal length.

The 2.0 Megapixel Logitech I posted on the 2nd page of this thread is the desktop version. The version that Tiffie posted above is the laptop version of the same camera. Same specs (same resolution), same software. BUT, the laptop web cams are really tiny. Look at the Quick Cam laptop cam next to a quarter:

http://i164.photobucket.com/albums/u15/rtgeorge_album/QCPro9000.jpg

...as opposed to it's big brother:

http://i164.photobucket.com/albums/u15/rtgeorge_album/QCCommPlus2.jpg


You really want a manual focus for a mill cam. All of the remaining 3 web cams I tried were manual focus, and could be manually adjusted to < 1/8" focal length. The web cams that were "auto-focus" (no manual focus ring) seemed to be optimized 6 - 36" focal length.


I haven't found a 2.0 Megapixel web cam with manual focus.


The product names are infuriatingly similar, so read the box carefully before you buy.

The difference between the Logitech QuickCam Deluxe and QuickCam Pro is 640 x 480 versus 1280 x 960 resolution (1.3 Megapixel). The Quick Cam Pro 9000 is 2.0 Megapixel.


The CentreCam software is lousy.

I spent 45 minutes trying to figure out why all four web cams looked largely the same, before I realized that it can't handle more than VGA (640 x 480) resolution. CentreCam also freezes when you remove the web cam from your computer (I tried it on my laptop and it does the same thing), it advertises as free for non-commerical use, but the demo version pastes ("Demo Version") on your images, and there's no way I could find to take a screenshot :rolleyes:

lazlo
02-10-2008, 05:50 PM
OldTiffie bought the 2 Megapixel Logitech Web Cam Pro for Laptops (the first picture in the previous post). Although the image was very nice for a web cam, I couldn't get it to focus well at less than 4 - 6", which doesn't make it very useful for a mill cam.

Although, if you look at how small it is compared to a quarter, you could probably fit it in a quill ring adapter, and leave it connected to the quill. I don't know how long the plastic shield in front of the lens would survive the chips, but it would make a fun quill cam :)

Of the remaining three, the Logitech QuickCam Deluxe for laptops was the 1.3 Megapixel version of the Logitech Web Cam Pro for Laptops, but it does have manual focus. Images were good, but not great:

http://i164.photobucket.com/albums/u15/rtgeorge_album/QCCommPlus.jpg

The remaining two web cams were the Microsoft LifeCam VX-6000, and the Logitech QuickCam Communicate Deluxe. Both 1.3 Megapixel. The manual focus on the Logitech was more solid, and I had a much easier time getting it focused. The colors on the Microsoft LifeCam were better saturated (which is actually a big deal, because it's tough to illuminate the workpiece when you have the camera < 1/8" off the surface), and as Evan mentioned, there seemed to be less sensor noise.

Here's a rule similar to the one Steve posted on the first page. Those are 1/32" tick marks, and the '8' is the major (inch) division:

Logitech QuickCam Communicate Deluxe

http://i164.photobucket.com/albums/u15/rtgeorge_album/LTQCCommunicatePlus.jpg

Microsoft LifeCam VX-6000

http://i164.photobucket.com/albums/u15/rtgeorge_album/MSvx6000.jpg

I'm still trying to decide between the Microsoft VX-6000 and the Logitech Communicate Deluxe. I think the next step would be to dismember the cases and improve the manual focus threading -- the lens' and the focus rings were clearly not designed as microscopes, but with a little tweaking, I think either of these two would make really great mill cams.

Cheers,

Robert

oldtiffie
02-11-2008, 12:15 AM
Thanks lazlo.

The previous 2 posts of yours regarding web-cams are as good a report as any I've seen in many a long day.

Very focused (sorry), concise and technically informative.

I suspected that my 2mega-pixel cam would be as you describe but I think it will do the job never-the-less.

Evan came up with what are 2 brilliant (sorry - again) solutions to 2 problems:
- illumination and exposure (flooding with light); and
- improving the focal lengths from a very small figure to infinity - by use of a "pin-hole" on the camera lens.

I suggested to Steve in a recent post on this thread that using a quality drafting scale (engraved back on white - for clarity, "sharpness" and contrast) would help as well.

Your posts have just about "pulled it all together".

You really did put a lot of time and effort as well as a very high odder of professional competence into this which I, for one, really do appreciate.

Many thanks.

I was very interested in the resolution and clarity in your pics - particularly this one:
http://i164.photobucket.com/albums/u15/rtgeorge_album/MSvx6000.jpg

One my computer screen the "28" and the "4" are about 70mm (say 0.280") each side of the "inch" (ie 32/0) graduations which at the "8" is 4mm wide - say 0.120". I could easily centre my mouse cursor - an by inference - a line - within 10 thou (0.010") of centre on the line at the "8" - ie within 0.055" on one side of the line and 0.065" on the other.

The effective magnification on my screen of the "8" to "4" (ie 4/32" = 1/8" = 0.125") at a screen distance of 70mm (say 2.800") is 2.800/0.125 = 22.4

Hence my "5 thou" offset on the screen is 0.005/22.4 = 0.00022"

That was not hard at all to get within 0.22 thou and could get much better than that.

I am sure that an accuracy of 0.0005" (ie "half a thou" is quite achievable.

Please check my assumptions and figures and let me know.

Again many thanks.

This really is the HSM forum at its best.

I suspect and hope that his thread has a lot of "life" in it yet.

And no - I still have not got the software installed yet - too busy really - but it will be done.

lazlo
02-11-2008, 12:38 AM
I suspected that my 2mega-pixel cam would be as you describe but I think it will do the job never-the-less.

Evan came up with what are 2 brilliant (sorry - again) solutions to 2 problems:
- illumination and exposure (flooding with light); and
- improving the focal lengths from a very small figure to infinity - by use of a "pin-hole" on the camera lens.

There's another, more elegant solution: the focal length is fixed on the camera without a manual focus ring. But there's not reason why you could add a manual adjust. And in fact that picture of mine that you posted is from the little Logitech Quick Cam Pro -- the 1.2 MPixel predecessor to your 2.0 Megapixel, fixed focus camera. So you should be able to crack open the front, and make a threaded capture tube that holds the front lense -- effectively retro-fitting it with a manual lense.


I was very interested in the resolution and clarity in your pics - particularly this one:
http://i164.photobucket.com/albums/u15/rtgeorge_album/MSvx6000.jpg

One my computer screen the "28" and the "4" are about 70mm (say 0.280") each side of the "inch" (ie 32/0) graduations which at the "8" is 4mm wide - say 0.120". I could easily centre my mouse cursor - an by inference - a line - within 10 thou (0.010") of centre on the line at the "8" - ie within 0.055" on one side of the line and 0.065" on the other.

The effective magnification on my screen of the "8" to "4" (ie 4/32" = 1/8" = 0.125") at a screen distance of 70mm (say 2.800") is 2.800/0.125 = 22.4

Hence my "5 thou" offset on the screen is 0.005/22.4 = 0.00022"

That was not hard at all to get within 0.22 thou and could get much better than that.

You could definitely do better than that. In most cases, I had a Rube Goldberg contraption of zip ties, center punches, and a Mituotoyo magnetic indicator holder serving as the camera holder. In some cases I was holding the web cam in my left hand while I was reaching for the "Take Photo" button :)

By the way, the 8 in that picture is 2.1 mm wide by 3 mm tall, if that helps.

I'm sure that an accuracy of 0.0005" is quite achievable.

lazlo
02-11-2008, 12:45 AM
BY the way, another issue that I ran into was getting the part evenly illuminated, when the web cam lens was 1/8" over the item of interest.

This is Mike Treth (the guy who wrote CentreCam, and the 2006 Model Engineer article)'s build:

http://www.miketreth.mistral.co.uk/images/photo%2001LowRes.JPGhttp://www.miketreth.mistral.co.uk/images/photo%2003%20LowRes.JPG

http://www.miketreth.mistral.co.uk/images/photo%2004%20Low%20Res.JPGhttp://www.miketreth.mistral.co.uk/images/photo%2006%20Low%20Res.JPG

He noticed, as I did, that it's critical to get a uniform light source onto the item of interest, or you get strange glare and reflections:

http://www.miketreth.mistral.co.uk/Adding%20LEDs.htm

"Initial results were encouraging but the image was very susceptible to changing light levels. This was in the main, caused by the close focus and there close proximity of the CentreCam body to the work piece. This necessitated a long angle light source which gave rise to glints and did not illuminate the work piece that well."

lazlo
02-11-2008, 12:49 AM
So his solution to the uneven light was to build a ring of white leds into the ring around the manual focus:

http://www.miketreth.mistral.co.uk/images/photo%2017%20Low%20Res.JPG
http://www.miketreth.mistral.co.uk/images/photo%2019%20Low%20Res.JPG

...and this was his result:

http://www.miketreth.mistral.co.uk/images/photo%2021%20Centre%20pop%20on%20blacked%20brass.j pg

I'm thinking a strand of electroluminescent wire wound in a ring around the manual focus ring would generate a much more even lighting.

BobWarfield
02-11-2008, 01:10 AM
Lazlo, you need to diffuse the light. Consider a plexi ring that's frosted or something similar. Otherwise you'll just get nasty reflections back. I've seen that on some other web cam articles.

Best,

BW

oldtiffie
02-11-2008, 06:07 AM
Lazlo, you need to diffuse the light. Consider a plexi ring that's frosted or something similar. Otherwise you'll just get nasty reflections back. I've seen that on some other web cam articles.

Best,

BW

Thanks Bob.

That was my thought as well.

oldtiffie
02-11-2008, 06:29 AM
There's another, more elegant solution: the focal length is fixed on the camera without a manual focus ring. But there's not reason why you could add a manual adjust. And in fact that picture of mine that you posted is from the little Logitech Quick Cam Pro -- the 1.2 MPixel predecessor to your 2.0 Megapixel, fixed focus camera. So you should be able to crack open the front, and make a threaded capture tube that holds the front lense -- effectively retro-fitting it with a manual lense.



You could definitely do better than that. In most cases, I had a Rube Goldberg contraption of zip ties, center punches, and a Mituotoyo magnetic indicator holder serving as the camera holder. In some cases I was holding the web cam in my left hand while I was reaching for the "Take Photo" button :)

By the way, the 8 in that picture is 2.1 mm wide by 3 mm tall, if that helps.

I'm sure that an accuracy of 0.0005" is quite achievable.

Thanks again lazlo for the the stirling effort.

On my screen the "8" is 50mm wide x 65mm high which is a magnification factor of about 50/2.1 ~ 24 and 65/3.0 ~ 21.6 - say and average of about 22:1 - so each "real" "thou" on the job is about 22 "thou" on the screen.

How far away from the rule were you?

This sort of re-inforces what I said earlier that the axis of the lens does not have to be all that close to the axis of the mill head quill not does it matter if the lens axis is tilted. If the axis of the lens is say with 5 or 10 or so thou at the target all will be well. If the quill/camera are rotated the image will seem to rotate and it will not be hard at all to adjust the mill table "X" and "Y" so that the circle formed by the rotating camera was co-incident/concentric any required position on the target.

I would want the quill to be at the position that I wanted it to be in when I was machining so a minimum focal distance of 4 to 6" would not be of real concern.

While admittedly, the increased distance would reduce the size of the image on the screen, I expect that (my?) increased resolution should reduce that disadvantage somewhat.

This combined with the software that Steve and Evan used to overlay "graticules" on the image seem to have huge potential.

Steve made and excellent point with the edgers/surfaces getting "blurry" and less distinct at the higher image size.

I guess that means another "trade-off' or compromise.

Evan's solution of using a "pin-hole" seems to be "the way to go".

For Evan: what size aperture would be required for the "pin-hole" please?

This is great stuff - Steve, lazlo, Evan, Bob et al - please do keep it coming!!!

Evan
02-11-2008, 10:45 AM
I haven't done any more experimenting with the pinhole yet. Too many other things to do.

I did take apart the Microsoft Lifecam and on the basis of ease of disassembly alone I would recommend it. Take out one single phillips screw in the back and the front pops off with an assist by a thin bladed screwdriver prying in the slot either side of the top button. That's it.

Two small screws hidden under black rubber plugs hold the stand/yoke on. The camera works can easily be left mounted in the back half of the case for installation in a custom housing. The front focus ring comes off with the front housing but the actual focusing part of the optics is still on the optical housing and can still be easily adjusted. An entirely different manual focus adjuster can be made and take advantage of the extra space it would give between the lens and the subject. This is possible because the lens in the MScam is set back a good 1.5 cm from the front of the original focus ring.

Incidentally, I did some more experimenting last night on the low light aspect of the MScam. The performance I am getting is phenomenal. It's like a starlight scope. Unfortunately there is no starlight to try it on. It's able to image clearly my entire living room and dining room on the light of a cigarette lighter.

lazlo
02-11-2008, 01:57 PM
I did some more experimenting last night on the low light aspect of the MScam. The performance I am getting is phenomenal. It's like a starlight scope.

The CMOS sensors used in most cheap web cams are very sensitive to infra red. So much so that there's often a glass IR filter inside the web cam lens element (that can be removed).

http://www.hoagieshouse.com/IR/Lens_Unmodified.jpg

More detail on how to remove the IR filter on a web cam:

How to make a webcam work in infra red
http://www.hoagieshouse.com/IR/

Evan
02-11-2008, 07:53 PM
I already removed it. I have a 35mm film can full of them. I use web cams for astronomy. It does produce some interesting color shifts without the filter.

DENedbalek
02-11-2008, 09:43 PM
Gentlemen:

I've been following this thread with great interest. I'd already begun the building of a simple laser crosshair (with a plastic magnifying lense to focus the lines to a much higher degree). Now with this webcam project I've abandoned the laser and jumped into this project. I have the web cam (an older Logitech QuickCam with manual focus) that does a good job of grabbing an image as close as about a 1/2" away.

My problem is building the holder for the camera. How are you guys getting the camera lense centered in the housing? I understand the need to able to adjust the angle (90 deg apart) to get it pointing vertical, but how are you centering it to the axis of your housing? Are allowing for adjustment there also? If so, what means are you using?

Thanks for the great discussion on this.
Dwayne

oldtiffie
02-11-2008, 10:33 PM
Gentlemen:

I've been following this thread with great interest. I'd already begun the building of a simple laser crosshair (with a plastic magnifying lense to focus the lines to a much higher degree). Now with this webcam project I've abandoned the laser and jumped into this project. I have the web cam (an older Logitech QuickCam with manual focus) that does a good job of grabbing an image as close as about a 1/2" away.

My problem is building the holder for the camera. How are you guys getting the camera lense centered in the housing? I understand the need to able to adjust the angle (90 deg apart) to get it pointing vertical, but how are you centering it to the axis of your housing? Are allowing for adjustment there also? If so, what means are you using?

Thanks for the great discussion on this.
Dwayne

Thanks Dwayne.

The original poster (OP) to this thread - S-J-H (Steve) at:
http://bbs.homeshopmachinist.net/showpost.php?p=331694&postcount=1

got it right first off.

Check his set-up and software.

Evan's software is first class as well.

All that had happened since Steve's first post was a process of exploring improvements and options and various web-cams.

Steve's set-up is still the standard - as Steve said, it might need some refinement - but perhaps not.

As I said earlier, it may not be necessary to have the lens centre/axis too close to that of the quill in terms of parallel or concentricity at all - even though it will remain desirable - perhaps just not necessary.

If the lens is pointed a the target sufficiently accurately that the eccentricity can be accommodated, measured and allowed for, all will be well. All that will be required is to rotate the quill with the web-cam mounted.

I am reluctant to open up and/or dismantle the camera lens from its body - so I have to find another way of installing/connecting the camera to the quill.

I might even try "Blue tack" - the blue removable "putty/Plasticine" stuff that is used ti fix photos etc to walls.

Its well worth a try.

The computer screen and the lens will do the job in conjunction with the software.

This thread and project has a long way to go before the discussion is complete, the prototypes developed and a working model achieved.

This thread is the HSM forum at its best and what helping each other is all about.

This thread is not and was not intended to be restricted to any persons or group - at all.

Your background thus far can only be of assistance to us all.

Please do join in and contribute.

S_J_H
02-12-2008, 12:41 AM
Well I'm really glad to see so much interest in this subject!
When I started this thread I sort of thought it would be of little interest to most.
I have used mine now a few times on real projects and not much else I can say other than it sure is a great new piece of tooling at my disposal. Thanks to all who have taken things a step further and tested some of the better webcams. I can see in the images Lazlo posted that they are sharper than mine. I'm not sure it would warrant me running out and buying a better webcam as so far mine will do anything I ask of it. I'm not finding to much trouble with lighting. I think I have enough light in my shop that it works well even very close up.
Steve

oldtiffie
02-12-2008, 01:39 AM
Well Steve,
you sure did start something good - real good - many thanks.

I'm not so sure that the better cams will do it any better.

The specs on mine say it will focus down to 100mm where-as yours will focus down to less that 3mm - that is a ratio of 33:1 in your favour that the increase in resolution of mine will not address.

I've thought about the "tilt" on your set up and came up with the "3-screw" adjustment as on the alidade on a surveyors level or theodolite - perhaps with springs on the screws to take up or counter any back-lash. Once you have the method of using these under your belt it will be second nature and is very quick and accurate.

I have down-loaded and installed the "CentreCam" software you showed in your early posts. I will give it a run shortly and will register it - I am sure it will be OK.

I have installed and registered Mach3 so I will try the video application within Mach3.

Please do continue with your critique', input and discussion.

oldtiffie
02-12-2008, 02:31 AM
It is, and the Logitech web cam should run on Win2K.

The driver framework is the same for Windows XP and Win 2K, but the Logitech is probably using the DirectX streaming audio and video stack for the video and audio capture (there's a microphone on the web cam).

I just checked, and the latest version of DirectX 9.0c specifically supports Win2K. I'd recommend you update to DirectX 9 before you install the Logitech drivers:

Supported Operating Systems: Windows 2000;...

http://www.microsoft.com/downloads/details.aspx?FamilyId=2DA43D38-DB71-4C1B-BC6A-9B6652CD92A3&displaylang=en

Thanks lazlo.

I just installed DirectX for W2K as you advised.

I will install the web-cam software either this evening or tomorrow and will give it a try on my office desk. I don't have a computer in the shop to use it with the mill - yet.

Sparky_NY
02-12-2008, 09:37 AM
Well Steve,
you sure did start something good - real good - many thanks.

I'm not so sure that the better cams will do it any better.

The specs on mine say it will focus down to 100mm where-as yours will focus down to less that 3mm - that is a ratio of 33:1 in your favour that the increase in resolution of mine will not address.

I've thought about the "tilt" on your set up and came up with the "3-screw" adjustment as on the alidade on a surveyors level or theodolite - perhaps with springs on the screws to take up or counter any back-lash. Once you have the method of using these under your belt it will be second nature and is very quick and accurate.

I have down-loaded and installed the "CentreCam" software you showed in your early posts. I will give it a run shortly and will register it - I am sure it will be OK.

I have installed and registered Mach3 so I will try the video application within Mach3.

Please do continue with your critique', input and discussion.

I have seen Mach3's video interface mentioned a few times in this thread and being a user that caught my attention. The centrecam software seems very useful with its various crosshairs, circles, centering lines etc. and I wonder if the Mach3 interface has similar features or is just a plain camera window?

Screen shots from anyone playing with the Mach3 interface would be real interesting to see. I am sure now that a cam will be a project in the near future after following this thread.

lazlo
02-12-2008, 09:56 AM
I'm not so sure that the better cams will do it any better.

The specs on mine say it will focus down to 100mm where-as yours will focus down to less that 3mm - that is a ratio of 33:1 in your favour that the increase in resolution of mine will not address.

Tiffie, look at the pictures I posted. Most web cams with a manual focus ring can focus at 1/8" or below.

Many of the tiny laptop web cams, including the "Logitech QuickCam for Laptops" that you bought, don't have a manual focus ring, and therefore can't focus less than 4 - 6".

So just get any high resolution web cam with a manual focus ring, and you're set.

oldtiffie
02-12-2008, 10:15 AM
Tiffie, look at the pictures I posted. Most web cams with a manual focus ring can focus at 1/8" or below.

Many of the tiny laptop web cams, including the "Logitech QuickCam for Laptops" that you bought, don't have a manual focus ring, and therefore can't focus less than 4 - 6".

So just get any high resolution web cam with a manual focus ring, and you're set.

Thanks lazlo.

I almost got the bloody webcam installed - but no go. Perhaps its because its a lap-top webcam - make any sense?

It was a bigger PITA uninstalling it.

I will swap it over this week for the normal webcam.

While it does have an auto-focus, it does have a utility under "Settings" to control the focus "manually" via software which was interesting. But I couldn't try it to see how close it would go as I couldn't get it to install.

I will have another look at more webcams.

I am quite convinced that this is the way to go.

Perhaps this one will install under XP - perhaps I might wait until I get my XP computer back later this week.

Yeah - I sure do wish I'd seen that report and pics that you posted - might have saved a lot of time, heart-ache and VERY bad language here.

I'm sure too that "tilt" and eccentricity of the camera with regard to the mill quill (within reasonable limits of course) will not be a problem.

The "tilting" base using the Surveyor tribrach adjustment or a variant of the principle will help a lot too. I just need to think it out a bit more yet though.

Again, thanks for the help as it is very much appreciated.

I will keep this thread posted as regards progress - or the lack of it!!

lazlo
02-12-2008, 10:27 AM
I almost got the bloody webcam installed - but no go. Perhaps its because its a lap-top webcam - make any sense?

You have to install the QuickCam V11.1 software first, before you plug the camera in. Did you do that?

Otherwise, it was an uneventful install for me...


While it does have an auto-focus, it does have a utility under "Settings" to control the focus "manually" via software which was interesting.

It says "autofocus" on the box, but its really fixed-focus. I think that focus menu is just cropping the image. I haven't returned that web cam yet -- I can try it this afternoon when I get home from work, but I dont have great expectations.

Cheers,

Robert

Evan
02-12-2008, 11:36 AM
I have been experimenting with the MS cam some more. I have it producing a usable image at 0.02 lux illumination. That is the illumination provided by a standard wax candle at a distance of about 20 feet. When I find some time from other projects I will write it up as well as do some more test on the pinhole lens idea. A proper pinhole lens may be able to provide better focus than the standard lens, especially if used in conjuction with a monochromatic filter. This may make it possible to image the features of interest at the resolution limit of the sensor.

Speaking of the resolution limit, the maximum accuracy of a digital optical system is hard limited by the sensor resolution. It is governed by the image scale and the size of the sensor together with the number of pixels in the X and Y axis of the sensor per unit length of the projected image features.

This can be calculated but the calculation must be made for each individual optical system and sensor. It is further complicated by color sensors as they only have a physical resolving ability one third of the pixel count. This is because the final color image is the result of interpolating the triads of RGB pixels into a single three component color value that is assumed to occupy the same position in the image space. In reality it does not as the three RGB pixels on the imager plane are in different physical locations. The resulting color pixel value is then composed of a "smeared" location in the original focused image that corresponds approximately to the center of the three RGB sensors that comprise a pixel triad.

This is further complicated by the use of different software and firmware algorithms used in the actual camera circuitry for combining the pixels to arrive at the standard True Color bitmap. This is then even further complicated by the use of compression schemes to transmit this information to the computer for display.

There are a variety of different schemes in use at the camera firmware level as well as different approaches to the actual design of the imaging sensor. Pixel arrays may be linear RGB, interlaced and staggered linear RGB, hexagonal RGB or alternate line RGB. Each has a different effect on how the final image corresponds to the imaging of small features that are near the resolution limit of the imager.

The bottom line is that the actual display accuracy of fine features may only be assigned a real +- precision value if these factors are known and accounted for. Empirical testing will not reveal these factors.

lazlo
02-12-2008, 12:32 PM
I haven't found a 2.0 Megapixel web cam with manual focus.



I found the Holy Grail: a 2.0 Megapixel web cam with manual focus: the HP EW099AA#ABA. True 2.0 Megapixel CMOS sensor, manual focus ring. Cheap! ($37):

http://www.techexcess.net/hp-2-megapixel-webcam-ew099aa-aba.aspx

http://www.techexcess.net/images/products/hp/hp-webcam-EW099AA-ABA.jpg

The only problem: the reviews say the picture quality is spectacular, but the video capture drivers suck. These web cams are all made by the same two or three Taiwanese companies, so I might be able to get this HP web cam to run with the Logitech drivers, or the generic Windows web cam drivers...

S_J_H
02-12-2008, 05:44 PM
I could not even load the software that came with my logitech webcam. I have a older non registered version of XP In my shop PC and I have not ever been able to upgrade to SP2 from Microsoft because of this. The logitech software says I need to upgrade to sp2 to be able to run and I can't do it.
So bottom line I I did not install anything yet the camera works in mach3 and centrecam as seen in all my pics. I guess I am just using whatever drivers are built into XP?

Steve

lazlo
02-12-2008, 06:23 PM
Right, there is a default, generic USB web cam driver that Windows XP (at least) will install if you plug the web cam in before you install the manufacturer's drivers. That's why there are loud warning stickers taped to the end of the Web Cam's USB connector when you take it out of the box: if the OS installs the generic web cam drivers, the Logitech software won't "see" the new web cam, and won't have the chance to install the "correct" drivers.

At that point, you can either delete the web cam from the hardware tree, or right click -> Update Driver, and point to the directory with the {Logitech, HP, Microsoft) drivers.

I'm not sure how that works for the high-resolution web cams though. Without doing the calculation, I think for > 1.3 Megapixels you need to use USB 2.0. I don't know what happens if you plug a high-resolution web cam into a (slow) USB 1.0 port, which can't handle all the data from the high resolution image.

Ordinarily you'd think it would default to VGA (640x480) resolution, but that would mean that the camera would have to down-rez the image, which I doubt they do.

I don't know if I have any machines left in the house with USB 1.0 -- I'll have to check when I get home from work...

oldtiffie
02-12-2008, 06:35 PM
I could not even load the software that came with my logitech webcam. I have a older non registered version of XP In my shop PC and I have not ever been able to upgrade to SP2 from Microsoft because of this. The logitech software says I need to upgrade to sp2 to be able to run and I can't do it.
So bottom line I I did not install anything yet the camera works in mach3 and centrecam as seen in all my pics. I guess I am just using whatever drivers are built into XP?

Steve

Thanks Steve.

That just about validates what I thought about my Logitech and the problems I am having with my fully registered and up-to-date W2K computer.

Its odd because the box the webcam came in said it required XP but the CD in the box said W2K. The software did not tell me I needed XP nor did it say that W@K wouldn't do either. Chasing what it said was the name of the "missing" driver and getting that installed was a PITA - didn't work - but it was a bigger PITA getting it un-installed.

I think I will replace the Logitech ands see what else there is that will suit W2K.

My XP computer is due back this week (hopefully) but I want the webcam to work on both W2K and XP.

I will look for a manual focus webcam even if at lower resolution. I am sure that is the best solution.

But I'll keep looking.

If the webcam was as good as my 40/40 (yep twice as good as 20/20) hind-sight it would be marvelous.

I can see that I am going to have to get another computer with XP and a parallel port (for Mach3) for the work-shop and the webcam. I've been stalling off with that but it seems inevitable.

This is getting to be like the HSM equivalent of the "Stations of the Cross" on the way/road to Calvary!! I suspect that the vinegar-soaked bun at the end of it is going to taste sh*t-house in this event as well.

I will keep you posted.

oldtiffie
02-12-2008, 06:43 PM
I found the Holy Grail: a 2.0 Megapixel web cam with manual focus: the HP EW099AA#ABA. True 2.0 Megapixel CMOS sensor, manual focus ring. Cheap! ($37):

http://www.techexcess.net/hp-2-megapixel-webcam-ew099aa-aba.aspx

http://www.techexcess.net/images/products/hp/hp-webcam-EW099AA-ABA.jpg

The only problem: the reviews say the picture quality is spectacular, but the video capture drivers suck. These web cams are all made by the same two or three Taiwanese companies, so I might be able to get this HP web cam to run with the Logitech drivers, or the generic Windows web cam drivers...
Thanks lazlo.

That is a good result.

I have read your latest post as well - many thanks - it all makes sense.

I have replied to Steve's post/s as well.

I will get what I can that is available locally to suit both my XP and W2K machines (each is a back-up/"Plan B" for the other and are net-worked).

I am not interested in the audio or vidoe aspects of thewebcam as I will never use it - so same goes for streaming.

I just need moving frames and pic captures perhaps for mill/shop use.

I checked the link you gave for the HP webcam and it requires XP SP2 - which won't siut me as I require W2K compatability as well.


System requirements: USB 2.0 (high-speed) port, CD drive, Microsoft Windows XP Service Pack 2, Intel Pentium 4 with 1.75GHz or faster processor or equivalent AMD processor, 512MB system memory

Bugger it!!

lazlo
02-12-2008, 07:01 PM
That just about validates what I thought about my Logitech and the problems I am having with my fully registered and up-to-date W2K computer.

Its odd because the box the webcam came in said it required XP but the CD in the box said W2K. The software did not tell me I needed XP nor did it say that W@K wouldn't do either.

There were several critical USB 2.0 hotfixes in Service Pack 2, especially 812308 (http://support.microsoft.com/?kbid=812308) "You cannot use your USB 2.0 isynchronous Webcam to capture video on your Windows XP-based computer."

I know it sounds like you don't care about streaming video, but a screen shot from a web cam is actually a single frame from the streaming video that the web cam is sending. So the web cam video has to work in order to get screen shots.

But, you should still be in good shape: the Windows XP Service Pack 2 USB 2.0 hotfixes were also included in W2K Service Pack 4. So there's no reason the high-resolution web cams shouldn't work with a Win2K install, but you have to have a USB 2.0 host bridge (USB 2.0 hardware support on your motherboard). USB 1.0 can't handle the high data rates from the high resolution web cams.

To check if your system supports USB 2.0, open up your hardware tree, expand the "Universal Serial Bus Controllers" tab, and look for one or more entries that say "USB2 Enhanced Host Controller." If you don't have that, your system doesn't support USB 2.0, and you won't be able to use a high-def web cam:

http://www.softwarepatch.com/pics/devman.jpg

By the way Mick, ordinarily I'd say call me and I'll walk you through the install, but since you're on the opposite side of the planet, the call would be expensive :)

oldtiffie
02-12-2008, 08:18 PM
There were several critical USB 2.0 hotfixes in Service Pack 2, especially 812308 (http://support.microsoft.com/?kbid=812308) "You cannot use your USB 2.0 isynchronous Webcam to capture video on your Windows XP-based computer."

I know it sounds like you don't care about streaming video, but a screen shot from a web cam is actually a single frame from the streaming video that the web cam is sending. So the web cam video has to work in order to get screen shots.

But, you should still be in good shape: the Windows XP Service Pack 2 USB 2.0 hotfixes were also included in W2K Service Pack 4. So there's no reason the high-resolution web cams shouldn't work with a Win2K install, but you have to have a USB 2.0 host bridge (USB 2.0 hardware support on your motherboard). USB 1.0 can't handle the high data rates from the high resolution web cams.

To check if your system supports USB 2.0, open up your hardware tree, expand the "Universal Serial Bus Controllers" tab, and look for one or more entries that say "USB2 Enhanced Host Controller." If you don't have that, your system doesn't support USB 2.0, and you won't be able to use a high-def web cam:

http://www.softwarepatch.com/pics/devman.jpg

By the way Mick, ordinarily I'd say call me and I'll walk you through the install, but since you're on the opposite side of the planet, the call would be expensive :)


Thanks lazlo.

Its there OK - "NEC USB2 Enhanced Host Controller."

see link:
http://i200.photobucket.com/albums/aa294/oldtiffie/capture.jpg

You wouldn't want to be walking through what I'm up to my ears in!!

Thanks for the very generous thought and offer though.

lazlo
02-12-2008, 08:42 PM
Its there OK - "NEC USB2 Enhanced Host Controller."

see link:
http://i200.photobucket.com/albums/aa294/oldtiffie/capture.jpg

Yep, you have USB 2.0. problem I see is that you have a yellow exclamation point on the Logitech QuickCam, which means there's a problem with the device driver associated with it.

Try this:


In the Device Manager, expand "Imaging Devices." Either the Logitech QuickCam should be there, or a generic "USB Web Camera" should be there.

Right click on it, and select "Uninstall".

Remove the QuickCam (unplug it).

Now install the Logitech QuickCam 11.5 software.

Reboot.

Run the QuickCam software -- you'll get an animated picture of a hand plugging in a USB cable. Plug the web cam back in.


You should get a Windows traybar popup that a USB device has been recognized, then that a Logitech QuickCam Pro for laptops has been detected, then the Logitech QuickCam software should suddenly pop into video capture mode.

oldtiffie
02-12-2008, 09:02 PM
Sh*t - that WAS quick lazlo - many thanks again.

OK, I will hang on to the Logitech web cam and will try as you advise - I am sure it will work.

I noticed the yellow mark as well but didn't realise what it meant at the time.

We have to go out for the day but I will give it a go.

I'm going to get on top of this if it kills me - which it might!!

I will keep you posted.

I hope that this might be of assistance to others who are reading this.

S_J_H
02-13-2008, 12:56 AM
hmm, found some work arounds for pirated XP versions so I can update to sp2. That was easy to find.
Maybe this weekend if I get bored I'll update the old pc so I can install the logitech software. I'm well aware to install the software first. But again, no sp2 and no way to install the logitech software. I doubt it would make much difference as this camera is only 640x480, but in the future I might want to try one of the better webcams.

Steve

oldtiffie
02-13-2008, 06:40 AM
Well,

I looked all over the posts on this thread to get the best option for a new web-cam for my W2K computer at a reasonable resolution, manual focusing and adequate lighting and software.

I also wanted a good glass lens.

I wanted to avoid the problems I had with installing and un-installing (on my W2K computer) the Logitech software (which is meant for XP but so far as I could see, should have run on W2K but didn't/wouldn't - yet).

For lazlo: I haven't used the remedies you advised, but I will shortly - I wanted to get my progress reported first.

Anyway.

The best I could do at a very good price was a generic/"no name"/(re?)"badged" Chinese web cam that is 1.3 mega-pixel, (VGA = 640 x 480 at 30 frames/sec) has manual focus, is made for W2K, has in-built lighting (LCD's) around the lens, does not have a glass lens. The installation and software look to be pretty good.

So around the circle I went and arrived back pretty much where Steve started.

It is also pretty much in line with lazlo and Evan's advice as well.

I can live with VGA (640 x 480) and 30 frames/sec if the lighting and image clarity are OK. Steve's posts and pics made that very clear and plain - particularly the manual close focusing.

This cam also has a ball-joint on the end of the stalk which should make life a whole lot easier for "tilt" adjustment which would work with Steve's set-up (which has forward-back and left-right adjustments).

I bought it here in OZ at a nation-wide store called "Dick Smith Electronics". I guess they will be in NZ as well. I have seen some of the "Dick Smith" stuff in the local "Tandy" stores but not vice-versa. I don't know if there is an "arrangement" there or not.

Anyway (again).

Pics follow:
this is a pic of the box the web-cam cam in. It pretty well sums it up.
http://i200.photobucket.com/albums/aa294/oldtiffie/Web-cam6.jpg

this is the "Generic hand-book" specs etc.
http://i200.photobucket.com/albums/aa294/oldtiffie/Web-cam5.jpg

This is a "top-right" view of the whole cam including the lens, LCD lighting, body and the "stalk" with the ball-joint on the end of it.
http://i200.photobucket.com/albums/aa294/oldtiffie/Web-cam3.jpg

and this is a "lower-right" view.
http://i200.photobucket.com/albums/aa294/oldtiffie/Web-cam4.jpg

Sorry about the artifacts in a couple of the pics but the photo set-up left a lot to be desired and had to be corrected (for).

I hope I/we are getting somewhere now.

My sincere thanks to all who have contributed to me at least getting this far.

Its not finished yet, but I am sure we are well on the way.

Any assistance, advice, comment or observations will be very welcome and appreciated.

oldtiffie
02-14-2008, 08:51 AM
Yep, you have USB 2.0. problem I see is that you have a yellow exclamation point on the Logitech QuickCam, which means there's a problem with the device driver associated with it.

Try this:


In the Device Manager, expand "Imaging Devices." Either the Logitech QuickCam should be there, or a generic "USB Web Camera" should be there.

Right click on it, and select "Uninstall".

Remove the QuickCam (unplug it).

Now install the Logitech QuickCam 11.5 software.

Reboot.

Run the QuickCam software -- you'll get an animated picture of a hand plugging in a USB cable. Plug the web cam back in.


You should get a Windows traybar popup that a USB device has been recognized, then that a Logitech QuickCam Pro for laptops has been detected, then the Logitech QuickCam software should suddenly pop into video capture mode.

Thanks lazlo - lots.

I did as you said - or tried to - all seems to be OK. Is it?

See first pic.

Previous situation (second pic) seems to have resolved itself.

Has it?

http://i200.photobucket.com/albums/aa294/oldtiffie/capture3.jpg

http://i200.photobucket.com/albums/aa294/oldtiffie/capture.jpg

t geist
02-17-2008, 09:08 AM
Sir, I have mach3 that I use for cnc router, how or where can i you it to get to the : video window that i see you using, is that part of mach3 or is that an add on to it?

lazlo
02-17-2008, 10:20 AM
Hi Mick,


Thanks lazlo - lots.

I did as you said - or tried to - all seems to be OK. Is it?


Sorry, I somehow missed this post. You've got the same problem as before: the Logictech QuickCam Pro is installed as a USB device without a device driver, so it has the yellow exclamation next to it.

When everything is installed correctly, the Logitech QuickCam will show up in the "Imaging Devices" section. Right now you only have one device there: the Epson Stylus Photo.

So try this again:


While everything is plugged in, right-click on the yellow exclamation "Logitech QuickCam Pro" and select "Delete."

When it disappears from the list of USB devices , unplug the QuickCam.

Now install the QuickCam Pro software.

Reboot your machine, and run the QuickCam Pro software. It should ask you to connect the web cam.

Now plug-in the web cam (but not before!).

oldtiffie
02-17-2008, 07:11 PM
Thanks lazlo.

Advice appreciated. I will go through that process in the next day or so. I have been "busy" on other threads as you will have noticed.

I have also de-fragged my disks and checked them as well as doing some much-needed house-work on the computer before I "Ghosted" 2 images of my "C" drive. I don't like to push my luck too far with the computer as regards "crashing", data loss and restoration.

I haven't forgotten the "web-cam" project, nor is it on the "back-burner".

I will install the new web-cam and give it a try out on the office desk in the next day or so and will post the results.

zarzul
03-04-2008, 12:03 PM
T Geist,

Just reading through the messages and saw your question. Mach3 has the video window built into it. I believe it is one of the top menu items. It changed from version to version if you have the latest version it is a plugin.

Arnie Minear
aka "zarzul"

.RC.
07-07-2008, 05:14 AM
Does anyone have any updates for this thread???

oldtiffie
07-07-2008, 06:30 AM
Thanks for resurrecting this thread Ringer - appreciated.

Not from this end really.

I got diverted with the CNC retro-fit for my new "Sieg" X3 mill and that stopped too. The mill is on a trolley and has been run but is still in its grease preservative -- just as it came out of the box. I am being pissed about by the supplier of the CNC kit - but I will persevere.

I bought the "Shuttle Express" that I intend to use with Mach3 so that I can "step" any axis at varying increments of "step" in any axis. I was going to use the camera on the mill with CNC control and Mach3 to do the camera-cam measurements under digital rather than manual control.

I am certain of the validity of the process but the "proof of the pudding" (concept) will be the actual implementation of the process.

I don't just want the cam\era for measuring but for positioning as well -ie to"pick up" features or details for location prior to milling.

I also want to eventually try it out as a digital comparator as well for checking things such as centre distances, angles such as on threading/screwing tools and correctness of surfaces produced by form tools. I will have a CNC-ed rotab as my 4th. axis.

The camera on the mill is a fundamental part of this process. I guess that the final accuracy will be limited to or by the accuracy of the magnification and resolution of the camera as well as the minimum stepping increment of the CNC system.

I may well put a 3-axis DRO (from Hare & Forbes) on the mill as well for use with the camera for use of the "metrology" purposes.

I have a HF-45 vertical mill as well that I have got a 3-axis DRO for but have still not installed on it yet. I'd like the camera on the mill set-up to be able to be fitted and used on that mill as well. That mill will not be CNC-ed.

I have a few issues here that I am dealing with that require most of my concentration and time and the camera needs more time and concentration than I can give it at present and will be that way for a while yet.

I am very pleased that this thread has been resurrected as I need to go back over it.

There have been some excellent posts on this thread by some very generous and competent and knowledgeable people here for which I am truly grateful.

John Stevenson
07-07-2008, 10:05 AM
Does anyone have any updates for this thread???

Yes.......

.

Alistair Hosie
07-07-2008, 05:19 PM
If I had one of them I'd have it trained on my beautiful face all day.Oh lod it's haaaaaaard to be humbeeeeeel wan yoo gets better lookin each daaaaaaaay.:DAlistair

John Stevenson
07-07-2008, 07:41 PM
After the initial work on the camera set in a pot chuck that goes in the spindle I realised that it had shortcomings.
The main one being that it wanted to be where the tool also wanted to be.

In a seach for a smaller camera I found some pinhole USB camera's when I was over in China last, here's one.

http://www.stevenson-engineers.co.uk/files/mill_camera.jpg

15mm diameter and about 40mm long so very compact without having to gut them and alter them.

Focus is adjustable and quite good [ no machine on site to show resolution and fitting ]

I have these permanently setup on the side of the head pointing down. No need to move these they can stay fixed out of the way.

To set them up initially I made sure they were vertical by raising and lowing the head so they kept focused on one point, that takes care of the Z.
Then I spotted a piece of metal in the vise and moved it to the side so I could pick the spot mark with the camera. Once on mark I noted the X and Y positions and put a new button on the setup screen called Zero Camera. This runs a simple VB script that carries the X and Y values.

So now to use the camera I move the work under the camera [ forget the spindle or tool ] until it lines up, zero X and Y at this point then press the Zero Camera button this them moves the work the preset X and Y values under the spindle and resets the X and Y values to zero.

Very quick, no having to remember any settings and no having to change the camera for a tool.

.

oldtiffie
02-23-2010, 07:23 AM
I have had an interest in things web-cam and measurement.

It lapsed when I had some problems installing one on my then W2K computer.

I had some on-going problems with my W2K computer which eventually $**** itself in the last 6 months (but as I'd had it since ~2001 I couldn't complain - but I did!!).

I had a total upgrade W2K>XP Pro but had stability issues there and it got wiped out with Norton - no surprise there - so it was a re-start. It had a transient intermittent problem with my "C" drive and it took some finding - but it is going very well now.

I had a another new computer built for W7 and it had "issues" but has settled down very well.

I have each computer virtually as a clone of the other and linked - I can switch between them using a Key-board - Monitor - Mouse (KVM) switch so as to save changing the KVM elements. I have a duplicate of each main-board so that I can be up and running on one computer while the replacement main board is fitted - if needs be.

I wanted to be able to use a good web-cam with my computer but with the web-cam fitted onto my mills or a good "X"-"Y" table - with a 6" "Vertex" rotary table so that I could use them as a comparator to measure distances and angle reasonably accurately.

I realised that I would need a web-cam with a very good resolution as well as a very good computer screen with a good resolution.

I realise that Mach3 has a "video" capability and that I would need a good "jogger" (4 -axis).

My Sieg X3 mill is to be CNC-ed but my HF-45 and my Sieg Super X3 are not to be CNC-ed.

My lathe cross-slide is made for a tool-plate (includes tee-slots) - so the web-cam can be used there as well to check on details and measurements (the rotary table fits nicely too).

I want the web-cam to be used on a good portable drill press that I have on a good "X"-"Y" slide that I also have on my desk - complete with the rotary table.

It all got a jog along when I saw a thread with a link recently relating to Steve's web-cam (in this thread) - thanks Steve.

I have "held off" as I was not satisfied with what was "out there" as regards web-cams.

I dropped into one of our larger computer (and lots of other "stuff") here in OZ last week.

There were two that met my requirements in different ways.

A cheap one had 6 LED lamps as well as a 10X digital zoom and 1.3 megapixel. (About 640 X 480). I was not too impressed with it but I installed it on my XP computer. I will post some pics of and give some details of it in a later post in this thread. I bought it for the LED's because of the reports in various places - here included - of "lighting difficulties".

The next one was excellent - and not cheap. It is a Microsoft
"LifeCam Cinema" - and it is excellent. I am certainly glad that I waited and got as lucky as I did. It has a video resolution of 1,280 X 720 in 16:9 format and a photo resolution of 5 megapixels. I was hugely impressed as it supplies its own light and focuses down to about 12mm (~1/2") and has a 8>9X magnification at that distance. A very good engraved draftsman's metric scale 10mm marks were about 89>90mm on my screen - and dead sharp - in *.jpg format. I will post some pics later.

A report of the LifeCam is here (looks like it was written by MicroSoft):
http://benchmarkreviews.com/index.php?option=com_content&task=view&id=443&Itemid=73

http://www.google.com.au/#hl=en&source=hp&q=microsoft+lifecam+cinema&meta=&aq=2s&oq=microsft+life&fp=2bcf0a553a95b437

http://www.microsoft.com/hardware/digitalcommunication/productdetails.aspx?pid=008

I will post some pics later.

I intend to use it for comparator/measuring purposes with the "Center Cam" software that I think Evan posted here.

I am not interested in using the web-cam as a sort of edge-finder - come- "3D Taster" as my real "3D-Taster" can do all that.

I had a very interesting email from LittleMachineShop.com today about this "Tormach" CNC scanner - looks really good - but it is expensive - and seems to need Tormach software as well as Mach3:
http://www.littlemachineshop.com/products/product_view.php?ProductID=3838&category=1963256893

I am also looking at this "jogger" (4 axis - "X", "Y", "Z" plus my rotary table plus a "spare") for use with Mach3:
http://www.littlemachineshop.com/3414

I have installed (IMSI) DesignCad 3D Max Ver 19 so I will have to get my finger out as regards getting back up to speed with CAD etc.

I have my Sieg X3 mill all ready for the "turn-key" installation of the CNC kits and Mach3 - including a new dedicated computer. I have paid for it and I have to get a few other things sorted out so that it can be installed. So it will take a few months yet.

davebaldwin
02-24-2010, 08:17 AM
Tormach recently introduced a product that may be of interest to readers of this thread. It uses a USB microscope instead of a webcam. I have no connection with Tormach other than a satisfied customer.

I haven't got one of their CNC scanners but I one day I may get around to making one.

Dave.


http://www.tormach.com/Product_CNC_Scanner.html
http://www.tormach.com/document_library/TechnicalDocuments/TD31454_Tormach_CNC_Scanner_Design_Analysis.pdf

nheng
02-24-2010, 12:23 PM
Kinda missed this thread, very nice job S_J_H !

Just a couple of quick notes on webcams that might help someone.

Many webcams are now UVC compliant. UVC (universal video class) compatible cameras, also referred to as "driverless" by some, operate over USB to a defined standard. Windows XP service pack 2 and higher have UVC support built in. They are also used by Linux and Mac people.

Logitech has a published list of UVC compatible webcams as an aid to camera developers:
http://www.quickcamteam.net/devices/logitech_uvc_device_list.pdf

The typical Logitech UVC compatible also comes with Logitech drivers which may provide additional functionality.

For computer types, there is also a Microsoft utility called "uvcview". It's a .exe that is an enhanced version of "usbview". It shows all USB devices on your system and gives a list of USB descriptors. If you want to see the actual camera resolutions made visible/available to your PC, they are in the descriptor list for the device ... a hex calculator would be handy ;)

If you have access to or want to spring for a C-mount USB camera, they open other possibliities, with a C-mount to RMS thread adapter, along with some C-mount male-female extension tube, you can toss a microscope objective on there for higher magnification and resolution. You can easily get into the couple of micron range ... great for engraving serial numbers onto gnat's arses :D I think Evan had a thread kinda related to this some weeks back.

Den

fireguy
03-15-2010, 03:08 PM
looking forward to seeing the pics you post oldtiffie.

with the new upgrades in cameras looks like an interesting project.

Please keep us informed.

KJ

cnc-joe
01-03-2012, 07:08 PM
Does anyone have a set of drawings or sketches for this ?

At least it would be a good place for me to start.

If I put something together - I will post it here.

flylo
01-03-2012, 09:32 PM
I have a dental camera & moniter that will be perfrct for this. It will focus from 0 to 80+ feet clear as can be. It's also very small & has a 90 degree mirror attachment. I'm glad I found this thread.

cnc-joe
03-01-2012, 09:03 AM
Ok.. I've doodled up something in CATIA - and want to share this..but can't seem to find the place where you attach an image to the message.....

?

Arcane
03-01-2012, 10:33 AM
cnc-joe, you can't attach an image in here. You'll have to upload your image to a site like photobucket.com or similar and then you can link to the image just by posting the url of the image or you can post the image itself here by using vB code. Example: photo url

cnc-joe
03-01-2012, 11:09 AM
cnc-joe, you can't attach an image in here. You'll have to upload your image to a site like photobucket.com or similar and then you can link to the image just by posting the url of the image or you can post the image itself here by using vB code. Example: photo url

Boy that really sucks.

flylo
03-01-2012, 01:27 PM
Where can I buy this copy of the magazine with this article? I have a very high resolution dentist camera with all the hardware color monitor etc that will focus crytal clear fron touching to 80'. It has the 45 & 900 degree mirrors & is about 1/2" in dia. I've got to try this.

George Bulliss
03-01-2012, 01:53 PM
Where can I buy this copy of the magazine with this article?

That was a popular article, featured in the Winter 2007 issue of Digital Machinist and written by Arnie Minear. This issue should still be available for purchase by calling 800-447-7367.

George

cnc-joe
03-02-2012, 03:43 PM
Ok, I have been putting together an experimental website.
You can find my CATIA sketch of the web cam here:

http://cnc-joe.com/projects_millwebcam.html

George Brown
03-14-2012, 01:23 PM
Was wondering, has anyone tried a usb microscope like this one

http://www.bhphotovideo.com/c/product/761329-REG/Celestron_44302_A_Deluxe_Handheld_Digital_Microsco pe.html

It has a built in light and magnification.

George

cnc-joe
03-14-2012, 01:47 PM
Yes - actually - Tormach has done this:

http://www.tormach.com/product_cnc_scanner.html

They even have some software to create a CAD file, too.
Looks nifty... just a little pricey for my use/need/roi.

Keep the chips flying,

Joe

George Brown
03-14-2012, 02:17 PM
Thanks Joe, I'm going to put one together. At under $50 for the usb microscope, you can't go wrong.

George

DICKEYBIRD
03-14-2012, 02:53 PM
Here's the one I bought: http://www.ebay.com/itm/2MP-USB-Digital-Microscope-Endoscope-Video-Camera-Magnifier-20x-to-200x-w-Driver-/110747638374?pt=LH_DefaultDomain_0&hash=item19c912da66

$27.98, free shipping. You may want to compare the specs. but I suspect they're the same unit.

George Brown
03-14-2012, 03:13 PM
Shoot, already ordered it

http://www.bhphotovideo.com/c/product/763006-REG/veho_VMS_001_VMS_001_200x_USB_Microscope.html

Almost identical specs, anyway, I'll have it tomorrow.

George

DICKEYBIRD
03-14-2012, 04:46 PM
Sorry, should'a kept my mouth shut. I hate it when that happens to me.

George Brown
03-14-2012, 08:30 PM
Don't sweat it, price diff won't kill me. At least I know I'm on the right track and it will work.

George

George Brown
03-16-2012, 11:08 AM
Just an update.....

The usb microscope
http://www.bhphotovideo.com/c/product/763006-REG/veho_VMS_001_VMS_001_200x_USB_Microscope.html

works perfectly. I can be almost 2 inches away from the material and still align to a pin point. I used the logitech quickcam V11.1 driver, since it was already on the computer to run the microscope, and it works well with that driver.

Now all I have to do is figure out how to mount it onto the cnc machine.

George