Description and what I learned

HOME

Work has begun on my 3 axis pcb milling machine. Thanks to John Kleinbauer, I was inspired to build a modified version of his design. I must say that his idea is one of the most impressive I have seen in a while. If you have not visited his site, do so now! http://www.geocities.com/crankorgan/index.html The $35 for the plans is well worth it if you are going to attempt this project.

Preface

The simple construction and material availability make this a great project for the intermediate to advanced cnc hobbyist. I did make several modifications to John's design based on my needs, but the idea is the same. I substituted the .25" PVC for .5" Polyethylene since the local surplus shop has loads of this for $2 per pound. Also, since I have welding equipment and a small lathe, I was able to machine all of the stepper motor to lead screw connectors as well as the bearing end plates. I also designed and built an anti-backlash nut to remove inherent slop in the all-thread. This should improve the accuracy of the machine tremendously. I mounted my slides on the ways a little different than John. This seemed simpler to build and more accurate. I will have some pictures available soon so you can see just how well this went together. For the controller, I am using John's design as well. It will run from my breadboard initially, until I can mill the board and make it permanent.

Timeline

11-15-01 After looking over the pics on John's website and making some initial calculations, I started investigating where to get some of the raw materials. I was initially concerned that I would not be able to find the store front extrusions, but after a visit to the local glass shop, I realized this would not be a problem. The guy took me in the back yard and he literally had piles of small pieces of this stuff, waiting to be recycled. I noticed a nice green section that looked to be in good shape and decided on that. The glass guy said that they had just done a car rental store front and they had to have this green frame. I had him cut me two 16" pieces for the x and y and one 12" piece for the z axis'. The price: $5.00, cut and ready to go.

11-16-01 The next concern was locating the pvc sheet for the end caps and bearing supports. I looked on the web and found the prices a bit too high for my liking, so I went in search of some surplus. I scored! The local industrial surplus shop has tons of white polyethylene plate for a nominal $2.00/lb. I loaded up on this because I knew there would be some trial and error involved in machining these pieces to meet my needs.

11-18-01 Time to begin figuring out how to assemble this thing. I thought about the slop in the all-thread and nuts and decided to machine a custom anti-backlash adjustable nut that would help tighten things up. Thanks to the HB 7x10 lathe, the job was quite easy. These things turned out better than I thought they would and they work extremely well.

11-19-01 I have the first axis completed. It took me a whole day to figure out an easy way to mount the slide on the track, but it works great and is very simple. The table saw that I have is nice for making true cuts and machining the poly plate for the end caps and slide. I botched a few nice pieces, but eventually, I got a tight fit with minimal friction and excellent alignment.

11-21-01 The x and y axis' are complete. This has taken several days to make all the parts. It was difficult to duplicate the parts because of the crude tooling I have, but alas, I have x and y done. Z will have to wait for a rainy day.

11-22-01 Took a trip to HSC in Sacramento to get a few steppers and some components. I scored again. $6.00 a piece for 1.5A all ball bearing stepper motors. This place is great. I can always find something I need there. I also ordered all the Tip120 Transistors from Mouser.com.

11-23-01 Tried several versions of the z-axis support arm. Tried the plumbing pipe initially, but it didn't seem rigid enough. I ended up welding together a monstrosity and decided it was rigid, but too ugly! Reverted back to the plumbing pipe but used 1.25" pipe to beef things up a bit. Getting close to a complete machine.

11-27-01 Finished assembling the driver on my breadboard. What a mess! 6 ic's and 12 transistors. A wiring nightmare. I couldn't believe it but I actually got the software to talk to the motors. What a miracle! I smoked a few wall-wart power supplies and eventually hacked together an old transformer power supply. This thing is a beast and should provide plenty of current for the motors. Initially, the motors were running very choppy and I couldn't figure out why. After reverse engineering the circuit down to the logic level to make sure it was doing what I thought it was doing(I have all the truth tables if anyone is interested!), I then realized that my hacked power supply lacked a cap across the output. I had forgotten that the wave coming off the full-wave bridge is very choppy. A nice big cap made the motors run smooth as silk. I still don't have the z axis up and running yet, so a pen clamped in the Z arm is being used. Next week, I'll pick up a Dremel and mount it to the z-slide.

11.28.01 Last night I got the controller to talk to the machine for the first time. Since I am using DAKCNCv3.0, I had to write a few Perl scripts to convert the Eagle Gerber output to something that DAK could understand. Using a pen as the tool, I plotted a simple circuit board that had 2 14pin dips on it. I was amazed watching the machine run. It is incredible! The resulting plot was a little screwed up; some of the pads are not drawn correctly, but I am sure this has to do with the code and not my machine. I also drew a few circles so I could calibrate the machine. DAKCNC's algorithm for circular interpolation is amazing. Watching the steppers ramp up and down as the pen approaches an inflection point is very cool.

Verdict

This is by far the most fun I have ever had building anything. The pcb mill combines the best of working with your hands and your mind in multiple technologies and is very rewarding. The cost is attractive as well. I spent probably around $100 for everything I have so far. I haven't purchased the spindle motor and bits or the controller case/cables. Most of this money went to Home Depot for the hardware to complete the assembly. The most expensive items are the stepper motors, at $6 each. I highly recommend this to anyone involved in any kind of circuit board building or hobby part making. What a blast!!

 

Problems I encountered

Difficulty rating

***** Many hours of tinkering and thinking, lots of frustration, and numerous trial and errors

**** Several hours of fiddling, some frustration, usually could make the part the first time

*** One hour of work, usual difficulties in dealing with electro-mechanical items

** Nominal amount of work, sometimes the hole didn't quite line up, but I could ream it out.

* Simple construction

  1. Locating the raw materials ***
  2. Lead screw machining and nut alignment *****
  3. Slide mounting and accuracy ****
  4. End cap/bearing plate construction (I hope to replace these with high accuracy ones that I will machine on this mill) ***
  5. X to Y mounting ***
  6. Z axis arm **
  7. Controller circuit assembly *****
  8. DAKCNC setup (motor timing and step rate) *****
  9. Alignment/calibration and leveling (dial indicator required) **
  10. Overall difficulty of the project **** (intermediate to advanced project with lots of electrical/mechanical experience required)

 

Resources

http://www.dakeng.com/turbo.htm

http://halted.com/

http://www.geocities.com/crankorgan/index.html

http://vv.carleton.ca/~neil/robotics/stepper/ttl.html