Webcam lens position tolerances

More
28 Mar 2011 11:58 #8194 by turbospeedskater
Hi.....

One question about webcams:

I have the AXIS Tab for webcams running, but all of my webcams have the problem, that the position tolerance between the lens and the CCD is too big for being really useful.
That means: if I touch the lens (or the machine vibrates during work), the lens is moving inside the lens-mount, and the picture is jumping for about 5-10% of the picture resolution.

So it is impossible to use the camera for adjusting the machine to holes or material edges....

Any idea?

Many thanks for any hint how to make it work,

Martin.

Please Log in or Create an account to join the conversation.

More
28 Mar 2011 12:35 #8198 by andypugh
It sounds like the ideal solution would be to stop the lens moving in the mount...

Is the camera mounted in the spindle? If so, then the pixel that doesn't move when you turn the spindle back and forth (by hand) is the one aligned exactly with the axis. That will always be true, regardless of how the lens moves.

Please Log in or Create an account to join the conversation.

More
29 Mar 2011 06:26 #8244 by turbospeedskater
Hi Andy,

thanks for your reply.

Up to now the camera is not fixed to the machine, I tried different positions when I encountered this problem.

What I would like to do is, to mount the camera on the Z-Axis, close to the spindle, and have the picture being adjusted when the Z-Axis in uppermost position. Then I have a known displacement between Spindle-axis and optical axis. With this solution, it would be easy to fix the lens with hot glue and everything would be fine. But the picture from that distance (about 400mm on my machine) covers an area of about 15-20cm length for 640pixels and that is much too bad for being used for precise positioning.

Another solution would be to fix the lens with hot glue with sharpness adjusted for a distance of e.g. 20mm. Then I only have to make a camera mount so that the optical axis is aligned to the spindle axis of rotation (what is easy, I just have done it for a laser) to be independent of rotating angle of the spindle/camera/laser and have exactly one "center pixle" that keeps in place at every rotation angle of the spindle. Disadvantage is, that I always have to remove the tool from the spindle and mount the camera whenever I change the workpiece.

I also thought of making my own lens-mount out of some metal (throwing away the plastic piece in front of the CCD that is in the camera), but I fear, then I will find the glass lenses not being aligned to center axis of the thread of the lens mount ( it's all only plastic... ). Ok, I can also make the complete lens system out of metal, but I don't want to reinvent a camera.....

I have seen some cameras on industrial robots and pick and place machines, but they are always "professional" cameras made of metal and they are much too big for being used on my machine.

So I still wonder if there is a webcam with good mechanical parts or any trick to solve this problem (maybe using some springs to push the lens system to a center position...).

How do the other guys here in the forum use their webcams on the machines?
Have they had the same problem?

Any idea?

Thanks for any hint!!!

Martin.

BTW: is there a maximum time for being logged into this forum? Every time I write something here (and that takes often more than an hour, because I have to answer some phone calls during writing), the forum software says "wrong code" when I submit the text.

Please Log in or Create an account to join the conversation.

More
29 Mar 2011 09:59 #8249 by andypugh
turbospeedskater wrote:

I also thought of making my own lens-mount out of some metal (throwing away the plastic piece in front of the CCD that is in the camera),

Is this an auto-focussing camera? Areyou talking of removing the auto-focussing components?

Are you sure that there is a real problem if you always make sure that the distance from the camera to the work is the same? It sounds like your problem might be paralax.
My intention is to hold my camera in a special toolholder so that it is exactly aligned with the Z axis. I assume this is not an option in your case?

BTW: is there a maximum time for being logged into this forum? Every time I write something here (and that takes often more than an hour, because I have to answer some phone calls during writing), the forum software says "wrong code" when I submit the text.

There is a timeout, but I think the "wrong code" is some other problem. It is wise on this forum (which we only have partial control of) to always select-all and copy the message before posting it in case it doesn't get posted.

Please Log in or Create an account to join the conversation.

More
29 Mar 2011 12:01 #8254 by turbospeedskater
Hi,

No, it is not auto-focus.
Focus is adjusted by adjusting the distance between the lens system (which is located inside a plastic tube with a thread on the outside) and the CCD.
For this, there is a plastic holder with a matching inside thread on top of the CCD. So the distance can be adjusted by turning the lens-system.

I am sure the two threads are not matching 100%, because you can feel the lens-system moving inside the CCD-thread when you push it from side to side.

I already have complete disassembled all of my cameras (which all have the same effect) and made several tests by glueing the PCBs with the CCD direct to the machine housing to avoid any movement of camera housing parts or PCBs inside the housing etc. So the only point left moving is the lens system inside the thread.

As I already said: I can for sure glue everything together (including lens), adjust the sharpness for a short distance and mount the camera in the spindle tool holder.
But then I have to exchange camera and tool every time I change the workpiece.

I would like to have a system where I can just measure out two points on my material and then automatic calculate position and angle displacement as a correction factor for X and Y positions for the next workpiece before I start the next milling/drilling job. And that only works in a short time if the camera is mounted to the machine and not inside the spindle tool holder.

I am not an optical expert ( I am electronic engineer with a lot of mechanical experience ), So I don't have any idea how to give a webcam another zoom-factor so that from Z-Axis top to material (distance about 400mm) I have a picture size of e.g. 10mm diagonal. With that I could glue everything together and can use the resolution for good adjustment.

Please Log in or Create an account to join the conversation.

More
29 Mar 2011 12:51 #8255 by andypugh
Why not mount it on Z-axis bottom, nearer the work? It could be quite a long way from the spindle, and pointing straight down. Use a camera-tool-offset in X or Y to allow for the offset. (You would need to calibrate the offset with a conventional edge-finder, feller gauges or whatever, but only once)

Mounting the camera on the moving part of the Z sounds sensible to me, then you can focus with Z-jog to different thicknesses. If the camera is vertical then there is no parallax error so distance matters less.

To focus on a small spot from 400mm away will require different optics. You almost need a USB telecope for that combination of spot size and distance.
www.instructables.com/id/40-USB-super-te...to-make-sees-crater/

Please Log in or Create an account to join the conversation.

More
30 Mar 2011 06:15 #8274 by turbospeedskater
As far as I can see this is the only solution:
Fix the lens system with hot glue and mount the camera somehow close to the workpiece.
Thanks for the hint with the telescope. I already thought about something like this, but lens systems that give me a resolution like the one showed in the link are too big to fit into my machine.
All space I have for a permanent mounted camera is about 30x30x30mm. Of course bigger ones are mountable but I fear I will break them during work. Or they get damaged by the cooling fluid or metal parts flying around.

I already had a new idea: Using two line-lasers that cross 90 degree exactly under the spindle and use the webcam only to get an enlarged picture of the crossing area onto the screen.
For this I can simply hold the webcam by hand somewhere close to the crossing point, or mount is somewhere close to the machine looking to the crossing point in some angle (means not vertical).
Then I can use a big-sized lens without any problems, because the camera is far away....

I have to think about this.....
Maybe I can find some cheap line-lasers somewhere to make some tests.

Many thanks for your input,
Martin.

Please Log in or Create an account to join the conversation.

More
30 Mar 2011 12:39 #8282 by andypugh
I asked a friend:
"If keeping the original optics then you need a "front-converter" type
teleconverter; i.e. an afocal lens system stuck on the front of your
existing optics. There are plenty of these about, usually x2 or x3,
sold for your average compact digital camera. Nikon sell their own
brand TC-E2 & TC-E3 for instance. I managed to get hold of a TC-E2 and
suitable adapters from that eBay to make it fit my Panasonic LX-2. But
that only gets you x3 at most which I doubt is enough for a spot 400mm
away.

The other option would be to throw away all the optics and get a
C-mount tele-photo lens from the aforementioned internet tat bazaar
and stick that in front of the webcam's CCD. You'd need to know the
diagonal of the CCD and the diameter of the area you want to be able
to see. The ratio of those two needs to be the same as the ratio of
lens focal length to 400mm.

I'd go for the second option if I had to do it."

Please Log in or Create an account to join the conversation.

More
30 Mar 2011 13:10 #8284 by BigJohnT
I made an adapter for a web cam where I stole the insides from the web cam and mounted it on a spotting scope for sighting in rifles. Just a simple bushing looking thing... Worked pretty well at 300 yards.

John

Please Log in or Create an account to join the conversation.

More
31 Mar 2011 05:20 #8308 by turbospeedskater
Hi BigJohnT,

using a sighting telescope seems to be a good idea.
I have one here and maybe I can make an adapter for the webcam. Should be easy, the webcam is already disassembled and I also have made some adapters for astronomical telescopes.

I will give it a try as soon as possible.
The only thing I am not sure about is: can I adjust the picture sharpness for a distance of 400mm...

But we will see.

I will keep you informed about the result.

best regards,
and many thanks for that hint,

Martin.

Please Log in or Create an account to join the conversation.

Time to create page: 0.163 seconds
Powered by Kunena Forum