Let's Make Robots!

CNC Machine V2 with RepRap Controller

I finally finished my Fireball V90 CNC build.  In May, I got the frame, and I planned to use the motors and electronics I purchased for upgrading the Valkyrie-clone CNC.  So I made some NEMA-17 to NEMA-23 adapter plates from CD-R disks, and got it all up and running.  It's running off 12V for now, with 24V planned in the future.


I use the latest Arduino version of the RepRap CNC controller, minus all the thermal controls (I don't have a thermoplastic extruder).  To send the GCODE, I still use Chris Meighan's "GCODE for RepRap" Java program.

22 Nov 2010 -- I had some problems surfacing the table -- the Dremel prefers to cut the material when it is to the left of the cutting direction.  In the other direction, the bit tries to vibrate itself loose (and often succeeds).  Perhaps as a result of my bungled surfacing, the PCBs were not flat enough to etch properly -- I had to choose between not cutting through the copper, or going so deep that it erased the traces.  I've re-surfaced the table and plan to see whether I can get my SOIC-16 IC's and SOT-23 MOSFETs traced.  I'm very happy with the accuracy, it's just the cutting bit that I'm unhappy with.  A proper PCB spindle and good bits are on my Christmas list this year.  For now though, I'm just praying to the gods of sharp steel and static coefficients of friction that my remaining bits stay sharp enough to slice through copper, and that the boards stick to the table.  Update: The surfacing did a good job of removing most of the gradient on the table.  The V-shaped etching bit is still VERY sensitive to small gradients, and I have enough tilt that the boards are not fabbing correctly, but the results are much better than before.  I think I'll get a SOIC if I keep all the small traces in one area.  I have that much more respect for LPKF now that I've seen how difficult it is to make a good board. 


DSCN2658.JPG175 KB
DSCN2661.JPG193.06 KB

Comment viewing options

Select your preferred way to display the comments and click "Save settings" to activate your changes.
  1. is this(http://letsmakerobots.com/files/stepper-motor-driver.brd) the current and correct design?

Thank you for your extensive documentation. I like the simplicity of the old rep-rap drivers. it seems they work fairly reliably, and the main problem you had was mechanical. I'm planning on starting to build a small CNC from a small copier frame. It needs a bit of stiffening, but i think it will work well(mostly foam cutting).


That's the board design I'm using -- it checks out against what's on the machine right now.  These were adapted from the original RepRap version 1.2 driver board to use parts I had easily available, but I can't seem to locate the non-wiki RepRap website I originally downloaded them from.  Maybe they're on the Subversion repository? 


I have these Rep Rap drivers working on a bigger CNC as shown, using NEMA 17 motors at 12V.

Any electrical problem with my Valkyrie-clone CNC were with a different version of the drivers -- the RepRap drivers have not had any problem except with heat.  By pure dumb luck, I installed big blue high-efficiency CPU heat sinks on two of the drivers, but only had scrap heat sinks from an ATX power supply on the remaining two drivers.  I found that the drivers stayed quite a bit cooler with the CPU heat sinks (though in Alaska, you could just run it outside, right?).  To keep the drivers happy I put a 12V fan (from my scrap box) on the drivers that ran hot.  Although I'm running 12V right now (I'm still using it only for PCBs and don't need the torque), they do fine with 24V (of course make sure your voltage regulator is specced to go to 24V).

I'd love to see what you come up with using the copier frame.


smd board cut with Fireball V90 and Dremel

This board is an SMT board -- it uses 805 (2015 metric) resistors and 1206 (3216 metric) 0-ohm jumpers, and a SOIC-16 shift register.  There's also some SOT-23 and (I think) SOT-223 MOSFETs.  The holes are 2.54mm spacing.  You can see the rough cut from the Dremel/engraving tip, but it looks pretty good to me.

Nice to see you back!

The Firewall seems a very good build, I guess that's "Bye home brew frame."

I'm fighting the same war over here, while I have very good X-Y accuracy I have major issues with the depth due to the table not being ... flat. I hope to fix it this evening, I'm planning to replace the whole X-axis. 

One thing I have found is that using double sided tape to hold down the material works a lot better then screwing it in, so maybe it would work better for you if you just find a real flat sheet of something and tape the pcb to it.

Unfortunately, yes, I've given up for now on the homebrew frame.  The Valkyrie frame was brilliant, in that I could (and did) make it with a battery-powered drill and a battery-powered saw.  I definitely learned a lot by building it myself, and I feel a lot more competant in using my tools now that I built something that worked.  I was not looking forward to building the next frame or upgrading the Y-Z axis for better accuracy with the same tools, and chose to get something that someone else made.

Thank you for the tip on table flatness -- you've tried surfacing, and it still doesn't work?  I use my largest end-mill with a flat bottom for the surfacing and it almost polishes the table, it's so smooth.  I plan to see today whether this has worked.  The only thing that might go wrong is that all my boards seem to have a slight bend to them, and I don't have a very good clamping system (I need a "T-slot" table and suitable fasteners).  I may try the double-sided tape, if my clamping system does not work.



... the whole thing is crooked including the rails which follow the nice curves ... so surfacing would only get me a flat surface relatively to the curves on the rails :/ I should never have trusted and the apparent sturdiness of LDF plates or their cheapness. I'm going double this time ... hopefully this one lasts longer. This is the last point of wild play ... once I have this one covered I think I'm done for the moment.