Computer Numerical Considerations

Not the full-fledged CNC project deserves but the brain-dump @esot.eric asked for

Similar projects worth following
Since @esot.eric put his CNC project (#7535) on hold I figured in order to maintain the quantity (not sure if I can match the quality, certainly not the extent) of new CNC related project logs I'd pick up the topic and share my ideas and frustrations with building this small CNC mill. Hopefully this will also motivate me to finally finish it (after being 2.5 years in the making).

My apologies for the bad pun/reference in the tagline ;)

Oh... where to start...

One, and probably the major, reason why I started building this machine is that I wanted to make PCBs. Just one-offs that could be made in hour(s) instead of having to wait weeks for the boards to be delivered. Obviously I could also etch the boards and while @CNLohr makes it look so easy, to me the process is just not appealing mainly because waste disposal is a whole nother issue. So PCB mill it is (or was). Plus, it might come in handy for milling other things as well and maybe I can 'slap' an extruder on there, too, but that's probably another 2.5 years down the line ;)


  • Minimizing backlash

    Stefan Lochbrunner11/28/2015 at 22:53 2 comments

    So far backlash hasn't really been a concern but the other day I had this idea that should help reducing it. It's similar to what was done in this machine which was recently featured on the blog where two nuts were pressed in opposite directions by some metal strips.

    It came to me when I had this threaded rod with those two standoffs laying on my bench along with a small box of misc parts among which were some springs. I was just sort of playing around with these parts wanting to see how far I could compress this quite stiff spring but then something interesting happened: The spring locked the two standoffs together pretty much eliminating the backlash, although it depends on how tight the spring is compressed. Of course, as it is the standoffs aren't completely locked, especially when the rod is turning fast but that's an easy fix.

    It's nothing revelatory but the way I stumbled upon it is somewhat amusing. As evidenced by the above example the general concept is a commonly known technique but the option of adjusting the pressure at any time makes for a nice feature.

  • Some links

    Stefan Lochbrunner11/12/2015 at 13:35 0 comments

  • Mechanical 'design'*

    Stefan Lochbrunner11/11/2015 at 16:54 1 comment

    *not really design, more like doing something and seeing what works/sticks. Problem: Metal doesn't stick.

    First attempt:

    I think you see what I was going for here:

    Read more »

  • 'Spindle'

    Stefan Lochbrunner11/11/2015 at 16:39 7 comments

    Thought about using a rotary tool but I'm not sure if the machine will be able to handle it. For now the plan is to use a DC motor with one of those cheap self-tightening chucks (although that's not too different from a regular rotary tool):

    Update 2015-11-24:

    Some thoughts on how I'm going to mount it:

    The motor is attached to an angle and will also be supported further on the left. Probably with some metal or acrylic pieces and a hose clamp to pull it tight against the base plate. The 2nd angle on the right will be a support/guide for the bit. I don't have any 1/8" inner diameter bearings (or drill bits for that matter) that could act as low friction guide so I guess a hole will have to do; it remains to be seen how I'm going to make that hole.

    Update 2015-12-23:

    I'm still waiting for the v-groove bearings and since I'd like to fabricate most of the next batch of parts in one go I haven't gotten too much done on this building wise. However, I did some tests on reducing wobbling of the bit. Initially the bit was off by about .5mm such that the very tip would move in a circle with a diameter of 1mm. I tried reducing it with a makeshift bushing but this had too much friction, causing the motor to slow down or even stall.

    Pushing on the chuck such that the tip of the bit would go towards the center approximately halved the wobble and by rotating the bit inside the the chuck got it to a point where the wobble is negligible. The problem is that this chuck holds the bit only on a slim ring where its jaws close and the end of the bit that is inside the chuck has some space to move around. In hindsight this is pretty obvious but I just couldn't resist when I saw this chuck on eBay. I guess this will have to do in order to finally get the machine running for the first time and I'll make some improvements once I'm there.

    Update 2016-03-03:

    As I mentioned in the other log I had to reorder the bearings but now that the second order arrived I finally got back the motivation to continue work on this.

    I pretty much just implemented the ideas outlined above so this is how the spindle mount looks like now:

    I made some mistakes on the acrylic base plate since I rushed it a bit to get it done and see if this concept would work but those are primarily cosmetic so I could just make a new plate if I feel like it. I also still have to shorten the brass right angle pieces since they limit the machines range of motion a bit. Assembly of the spindle-mount/z-axis (shown below) was a bit of a pain because a couple of screws are obstructed - the ones under the motor on the metal angle, for example - but once done it seemed to work quite well. The optical flaws aside, the motor is mounted pretty sturdily to the z-axis so if I can get rid of the wobble of the bit it should work nicely.

  • Endstops

    Stefan Lochbrunner11/11/2015 at 16:32 0 comments

    They are not integrated yet but these are some ideas:

    One of the ideas was to use buttons as endstops and mount them to the aluminium something like this:

    But since I also salvaged some of these optical interrupt switches from the printers it'd probably be a good idea to use these instead:

    Read more »

  • Electronics

    Stefan Lochbrunner11/11/2015 at 15:33 0 comments

    The stepper motors seen in the main project image are salvaged from old printers. Maybe I'll write about them in more detail some other time.

    To control the motors I'm using StepSticks which are really cheap on eBay and easy to use and are in turn controlled by an Uno.

    As you can probalby guess from the image the Uno is running GRBL so nothing fancy here.

View all 6 project logs

Enjoy this project?



0xF wrote 03/03/2016 at 22:26 point

If I may comment your design looks good, but I don´t see alumin(i)um holding up to the hard steel bearings long before your machine starts developing slack. I'm no mechanical engineer , just a guy with an angle grinder but IMHO i'd make the slide rails out of steel (readily available) or incorporate some method for adjusting the bearing as the rail wear out.

I've got to start searching around here for some of those V bearings tough, they look promising as linear sliders and i'v always wanted to build somesuch devices.


  Are you sure? yes | no

Stefan Lochbrunner wrote 03/03/2016 at 23:23 point

Interesting point, thanks for pointing that out! In the current design the x-axis is adjustable and I was planning on doing the same for the z-axis but the rails are probably not going to wear out evenly so replacing them with steel is a good call. Covering them with a steel angle might also work.

I think I'll keep it this way until the machine is actually working and then observe the issue. If it becomes a problem the square profiles are easily switched out for steel ones.

FWIW these are the bearings I bought:

Thanks again for the comment, this might save me some troubleshooting in the future :)

  Are you sure? yes | no

0xF wrote 03/04/2016 at 18:22 point

You're welcome, those are damn cheap on ebay, i might order a few (or a few dozens) thanks for the link !!

  Are you sure? yes | no

Eric Hertz wrote 11/12/2015 at 04:05 point

Awesome! Nice use of a CD/DVD part, as well!

Never seen those chucks before, reminds me of this cool old "mechanical" pencil I inheritted... Opto-hacking with onboard resistors: nice job.

Definitely getting some ideas, here. And thank you for finally putting into words why I hesitate to etch boards... waste-disposal. Duh.

  Are you sure? yes | no

Stefan Lochbrunner wrote 11/12/2015 at 13:13 point

Thanks for all these kind words!

Another DVD part I might use is of course the laser diode but I should probably finish the mechanical parts first. After burning a couple of diodes do you still think that LM317 driver is the way to go?

These chucks come in various sizes/diameters as they can't just hold any size bit. We'll have to see how it works out, I might have to stabilize the bit with a bearing though.

If you're set up to do etching it's probably easier but if you're set up to have a dusty workshop a mill is the way to go ;)

  Are you sure? yes | no

Eric Hertz wrote 11/12/2015 at 17:43 point

I'm not sure about the LM317 driver... The latest diode (and the first) seem to be OK, so maybe I just overpowered the others... it's also "the" way that most laser-hackers allegedly do their lasering. 

The chuck thing, was just thinking about it... Most rotary-tool bits are 1/8th inch diameter, if I recall... so *making* such a chuck would be relatively easy with a 1/8th drill bit, and another that matches the motor's shaft, a threading-tool, and a dowel of some sort... a couple screws for tightening. Using a DC motor rather'n a rotary-tool is a good idea; much smaller, speed-controllable, and leaves the rotary-tool for other uses. I'm working out details as we speak. Good call on bearings, I'll keep that in mind.

Mills are more versatile, too... PCBs, of course, but also wood-pieces and more. If I understand correctly, etch-tanks should be well-sized for the boards, so as not to waste etchant, which means lots of different tank sizes. Blah. The cat's just gonna have to get used to mill-dust, or I'll enclose the thing in a clear box. Thanks for that reminder, as well!

  Are you sure? yes | no

Eric Hertz wrote 11/13/2015 at 09:14 point

"Making such a chuck would be relatively easy..." my a**... Sorry I even suggested it. At the end of the bit  I'm getting between 1/16th and 1/8th inch of "wobble," way too much for a PCB :/ It's definitely largely-due to my chucks and not having exactly the right-sized drill-bits. Wee! Redesign!

  Are you sure? yes | no

Stefan Lochbrunner wrote 11/13/2015 at 10:41 point

In principle it's pretty straight forward but drilling exactly through the center of some cylinder is the kind of thing where you either need the proper tools (I don't think a drill press will cut it) or try until you get it right. Have you checked out the spindle link in the 'some links' log? It's all off the shelf parts. I got around 1mm of wobble with that chuck, hence the bearing for stabilization. The chuck is mostly a coupling at this point.

I think you can reuse the etchant, then it doesn't matter if the etch-tank or whatever container you use is bigger. I assume etch time are going to vary though. 

  Are you sure? yes | no

Eric Hertz wrote 11/14/2015 at 09:04 point

Good info, thanks. still experimenting :)

  Are you sure? yes | no

Eric Hertz wrote 11/12/2015 at 19:03 point

Good to know, but China doesn't ship to my neck of the woods (a land called Brokeasheck) ;)

  Are you sure? yes | no

Similar Projects

Does this project spark your interest?

Become a member to follow this project and never miss any updates