New Zero Camera

I had read in a Make Magazine post on an efficient and reliable way to zero your CNC. They used a USB microscope camera. I thought it was a great idea and here is how I implemented it.

Here is the camera:

Here is the bracket I designed to fit it as close to the spindle as possible.

And here is the finished product:

The camera has LED lights on the tip and they work great.

I milled with two materials. Teak on the top one and a plastic called Starboard on the bottom. I wanted to use a hard wood because I loved the look but don’t think it is the right material for this design because it will crack after too much pressure. The plastic is hard enough but will take the pressure of the bolt to tighten everything into place. I think two are needed as to keep the cameras vertical alignment as solid as possible. Ideally I would have milled this out of aluminum but don’t think I know how to handle that yet. If someone is interested in doing this also but want to do it in aluminum, please let me know. I’d gladly share the files and would be interested in the aluminum version. I also have no drill press so I couldn’t countersink the screw and tap a thread. Right now the design has a 1/4-20 thread and cap screw as the final idea.

As a proof of concept I think it works great. I need to see if the current setup will drift over time with the vibrations of normal operation. It would be ideal for the camera to be attached to the spindle head somehow into one of it square sides as opposed to the cylinder to solve the possibility of drifting.

To view the image of the camera I’m using a demo version of a software called Scopebox:

It allows me to display a center mark overlay. I think it’s only for the Mac. The camera comes with software for the PC that I think already does the job but I can’t test it as I don’t have a PC. The hole you see on the center of the screen is from the smallest tool I have, a .032 square mill #122. I marked the material and then moved the camera over to find it. This gives me the offset numbers.

Let me know what you think.


@patofoto, I think that is way cool. Thank you for the complete description.


Awesome! Do you have a link to the Make article as well?


Love that article, love this attachment!

I was thinking about his solution for Z-zeroing on his machine, and wondering if a manual way of doing the same thing would work for us. Can someone please tell me, would it work to wire a bit of copper plate, a button battery, an alligator clip, and an LED together, clip the alligator to the cutter, put the copper plate on the work surface, and lower the spindle until it touched the plate, thus creating a circuit and making the light come on? Then manually subtract the width of the plate in the zeroing dialogue?

Does this even make sense? Would it work? Would it be accurate? Would it damage the circuits in the nomad?

EDIT: just to clarify, the idea would be that as soon as the light came on, you would know you were touching, instead of having to do it with rolling papers, by feel…

Here’s a rudimentary sketch of my idea for a zeroing device. Please forgive my complete lack of drawing and electrical engineering skill!

My questions:

  1. Would this work? Would it be precise?
  2. Would this potentially send electricity into the spindle and damage it?
  3. Could you use a multimeter set to test for continuity, and achieve the same results?

@MrHume, the usual thing to use is a piece of single-sided PCB where you show the copper plate. This removes the possibilty of shorting through the workpiece into the machine frame if you have a metal workpiece and aren’t using a wood/MDF spoilboard. It also removes the possibility of sending electricity back through the drive electronics. But an LED is milliamps so that would not be much of a problem in any case.

You just compensate for the thickness of the PCB (which you have previously measured) when you set the Z zero.

I think a multimeter set for continuity (annoying but useful beep) would work just as well.

And your skeching looks pretty darn good to me.

1 Like

If I understand what you’re describing, the formal name for that is a touch plate — I did a simple one for my SO3:

@patofoto I was revisiting this concept because I plan to try it myself. It’s been a while, have you noticed any issues with the microscope-camera drifting?

Also, do you have a way of calibrating/checking the alignment of the camera with the spindle? I was thinking it would be simple to just have the Nomad cut a couple straight channels, using the smallest cutter I have, and then check to make sure the microscope is lined up when the cutter is in the channel. But if there’s a better way, I’d love to know about it.

Thanks again for posting this cool device.

This is great stuff thanks

Unfortunately I encountered something with the camera I cannot fix. The camera is housed inside this pen-like cylinder. There is a knob at the top that moves the camera up and down to focus. The camera shifts within this cylinder when you focus and won’t keep it’s position. It can easily be moved, loosing it’s registration. One could re-check the offset every time used and it would be very accurate but it can be easily bumped and a re-check would be necessary. I described my way of checking the offset above. Hope that helps. I’ve been meaning to write to the people in the article where I read about this idea to see how they overcame this problem. If anyone is interested in the files for the brackets, I’d be glad to make them public.

Do you mean that it shifts left and right inside of the tube when the focus is changed? If so is it possible that the microscope is defective? If it’s not that, how does changing the focus change the offset? I’m sorry if this is a stupid question…

If the issue is the knob, would it be possible to just tape it (or even glue it) so it wouldn’t shift?

As for the offer to make the files public, I’d be very glad to have them.

I made a similar microscope mount for the Dewalt 611 using a Supereyes 'scope which was much cheaper than the Adafruit one linked in the above article, $28 versus $150.

I found it much easier to 3D print the mount and I can confirm that the Supereyes really needs an upper and lower clamp as it is not a cylinder. If you clamp it in one place, it will wobble. The differences in the diameter can be seen in the attached picture and on the Amazon page.

Do you have any issues with losing registration due to the focus knob being bumped? I certainly like the idea of a much cheaper camera! I don’t have a 3d printer but I’m confident I can make something work with just the Nomad.

The camera I have moves when focused and also when the USB cable coming out of the cylinder gets bumped. Not sure how much of an issue this as you can surely remeasure the offset. When the camera moves the angle where it points changes slightly. So far I haven’t pursued this any further since I’m still just doing simple projects to get used to my CNC.

I’ve circled that part that moves.

I’ve also been looking at that particular microscope from Amazon. Have you considered using a couple of dabs from a hot glue gun to secure the focus knob? I image this would be easy to remove and reapply if adjustments were necessary in the future.

Here is the STL file for the bracket in the pictures. The screw in the design is a 1/4-20 cap screw 1/2" in length. I have no way of tapping this with accuracy so I just drilled a hole and used a regular screw and nut.

Zero Camer Bracket.stl (265.3 KB)

Hope it helps someone. I design in Fusion 360 so if there is another export that might help anyone better, just let me know.

1 Like

True. Have not tried it. Will do. Thanks.

I was walking up the stairs today when it suddenly occurred to me that you could probably mount the camera tube inside of a piece of PVC pipe very easily. A few drops of hot glue plus a carefully drilled cutout at the top would serve to secure the USB cord from bumps, and I suppose if you only covered the top part of the camera it you wouldn’t have to worry about changing the brackets, though I don’t know if there’s clearance for this in the existing design.

Anyway just a random idea that popped into my head. I will likely implement something similar myself…

IMHO the camera moves too much inside it’s own housing to make it a viable solution now. I just did a quick measuring of the offset five consecutive times and I got five different results with the Y offset only moving about one hundredth of an inch but the X moves around two tenths. Too much for precision work. I will study a way to secure the camera. Will try to get in touch with the manufacturer to see if they suggest anything. Will keep you posted.