Surplus Store Router Build/Conversion

[juzwuz]

Hi! Not sure if people have heard but the Boeing Surplus Store in Washington state has closed it doors for good at the end of December. There were lots of cool stuff there but my luckiest find was a 3-axis machine which in it's previous life was a "laser profilimeter" -- whatever that is. No laser equipment came with it but it did include a nice big Daedal x-y table and a Daedal actuator for the z-axis and some old motor drivers/controls. I kinda hesitate to call this a build log since I didn't really have to do too much to get it working. I sold off the old controls, attached some new Keling 280 oz-in stepper motors, and wired the motors to my stepper motor driver setup from my CNC Taig mill. I cobbled together a spindle holder with some scrap Baltic birch plywood and the machine was basically working.

The current travels are x 14", y 21", z 3" but I think I can eventually increase the x to 21" by making a better spindle holder. The leadscrews on the x and y axes are 10 tpi and it's 5 tpi on the z axis.

Overall, the machine is kinda unusual looking but it is essentially a moving gantry machine with the linear bearings spaced very far apart on the x and y axes. The weakest part is the z-axis support which is 0.5 inch aluminum plate and is not very stiff. The spindle holder also needs to be beefed up.

[juzwuz]

I'm running Ubuntu Linux on an old surplus store pc (Pentium III, 1GHz, 256Mb RAM) and EMC2. I've looked at various spindles but I've settled on the Proxxon IB/E rotary tool. Unfortunately the largest bit is 1/8" diameter which is kinda small when you want to mill out a lot of material. I wanted to keep the spindle very light so I could get good speed when doing 3D cuts. One of my interests right now is to make lithophanes using the Vectric PhotoVCarve software. I also want to eventually mill out a sailplane wing from balsa.

Anyway, everybody likes videos so here's my first try at uploading one on Youtube.



I'm attempting to cut an approximately 8" by 10" lithophane picture. I'm first cutting a 1/8" deep pocket into 1/4" Corian. Then I'm cutting the lithophane picture. I've set the feedrate at 50ipm but I haven't tuned the acceleration parameters yet so the 3D cuts are not as fast as I would like. There are about 2000 lines in the picture at about 30 secs a line! A couple problems occured during this lithophane project including a machine crash at around 8 hours into the project. More on that later.....

[juzwuz]

I rushed the lithophane project a bit and basically was testing the machine at the same time. I needed to get the project done because I was supposed to give the finished picture to someone retiring at the end of December. Anyway, the first problem I encountered was that my pocket corners were quite rounded. Much larger than my 1/8" endmill diameter. I mentioned this on the Vectric forum since I was using VCarve Pro to generate the pocket toolpath. GER21 mentioned that he thought it could be a problem with using the constant velocity mode in Mach 3. Of course I never heard of this since I am using EMC2 but a quick google search turned up the fact that the gcode post from VCarve I was using is generic and didn't have any trajectory planning codes for EMC2. Basically EMC2 defaulted to a constant velocity mode. I still have to play with adding in the proper EMC2 gcode to tell it to follow the corners as close as possible to a certain tolerance. It wouldn't hurt to actually figure out the proper acceleration parameters either.

Another problem on the pocket cutting was that cutting a 1/8" deep pocket with a 1/8" endmill was probably pushing the limits of the spindle and spindle holder. On the straight runs, at 50ipm everything was cutting OK but when the machine turned the corner, it seemed like the bit dug a little bit deeper. I wish I got a good picture of it. When the pocket was finally cut out, I had what looked like an x shaped depression. It wasn't deep enough to affect the lithophane though. I'm guessing that I need to beef up the spindle holder and z-axis in general. Maybe with some aluminum extrusion or angle.

Anyway, if anybody is interested, I'm uploading the dxf file for my spindle holder. I just drew it up in VCarve. I then used VCarve to generate the toolpath and I used my CNC Taig mill to cut it out of 1/2" Baltic birch.

[juzwuz]

Now the more serious problem of the machine crash. The lithophane was cutting OK, albeit slowly for about 8 hours since there were a lot of little x and z axis moves and the machine never could accelerate to the full 50ipm. We were supposed to go to a neighbor's house for dinner and I thought sure...why not?....the machine was fine for 8 hours straight! Well, when we got back from dinner I could still hear the machine working away but it sounded a little bit funny so I went to check on it. To my horror, the gantry was all the way to the right side of the machine with the z-axis jammed up next to the side. I quickly hit the software estop and shut down the spindle and surveyed the damage. The damage wasn't too bad luckily. The little c-clamp I was using to hold the spindle holder closed had hit the side first and the spindle holder gave way and sheared off. The x-axis helical coupler was also in pieces and at the time, I thought this was what saved the machine from more serious damage. Basically it looked like the coupler sheared itself apart when it hit.....

Now the problem was trying to figure out what caused the crash. I first thought it was a signal error from the stepper motor drivers or a weird bug in EMC2 or a electrical surge. I had heard that servo systems could possibly fail in a runaway condition but I couldn't think of a way that a stepper motor system could fail like that. I was definitely kicking myself for not hooking up the existing limit switches.

Anybody want to make a guess what caused the failure before I describe what I did next?

[ViperTX]

Power line glitches.....need more info.

[juzwuz]

I managed to make another spindle holder on my CNC Taig and I dug out another helical coupler and installed it. When I turned everything back on and tested the axis by jogging back and forth, it worked fine. Basically I couldn't make it reproduce the error.

I decided to try and make a smaller lithophane picture (4" X 6") since I had no time left for the bigger one. I started the machine going again and I just watched for at least 20 minutes but it behaved itself so I went back in and worked on other stuff. About 7 or 8 hours later I just happened to be sitting in the den surfing the net and I could hear the router working when suddenly I heard a funny squealing noise. I ran out to the garage and sure enough, the axis was heading in the wrong direction again. I managed to hit the software estop with about 3 inches to spare.

To my surprise, when I looked at the helical coupler that I had replaced, I noticed it was broken and the machine hadn't crashed yet. That was kinda weird because I had initially thought my first helical coupler broke on impact. Then the final straw was when I manually turned the axis....in one direction, the axis didn't move but when I turned it the other way, the broken helical coupler managed to still engage and move the axis! Basically the aluminum helical coupler was too weak for the application and broke but it was able to engage and move the axis when traveling in one direction but not the other.

Lesson learned: I need to use a better motor coupler

Anyway, I'd never heard of this happening before so I thought I'd share it with everybody. Basically, to fix the problem I bought some big steel helical couplers that I think should work better.

I'll work on replacing the couplers and then beefing up the z-axis next.

[High Seas]

Its good you have found the root cause of the problem - but maybe one of these:
http://cgi.ebay.com/2-4GHz-Wireless-...QQcmdZViewItem
and a long pendant with an E-Stop would be helpful too.

Got one for my wife (similar set up) and she uses it while firing her kilns. She toggles between net - tv - and shop surveillance.
Jim

[juzwuz]

Thanks for the tip! I was thinking for an estop I could also somehow get a network cable or use wireless to connect the machine computer to my home network and use some sort of remote terminal.

I think I found the root cause of the initial coupler failure too. While I was replacing the couplers, I found that there was a missing bolt where the stepper motor mount attaches to the table. There was only one bolt attached which basically could cause the whole motor mount to pivot. I think this caused too large of a misalignment and fatigued the coupler quicker. That axis always did have a different sound to it...it was a little more squeaky.

[juzwuz]

All the couplers are replaced and the motor mount is fixed. After looking at the specs for the Daedal table, I found that the manufacturer recommends lubricating the leadscrews with a light machine oil with teflon. Anybody have recommendations for an oil with teflon? I've been using ATF for my Taig mill and I was hoping that would be good enough for the Daedal table. The Daedal table does have some sort of brownish plastic nut (you can see it in the first post, 4th picture). I couldn't find any reference to what kind of plastic it is in the manual....any educated guesses? I'm wondering if it can withstand a high rate of travel and acceleration before wearing out.

[BobF]

I would guess it is delrin or something similar. Delrin is material of choice for this type application. I believe delrin is an acetyl plastic.

[diyengineer]

Stop by the local Tacoma screw, they have them everywhere in WA state. They have a lubricant with teflon in it. If you want to do it right, give Mcguire bearing company a call in Tacoma, Washington and ask them which lubricant to use.

[RomanLini]

I've seen a couple of those brown nuts in 1970's equipment. They are usually glass fibre reinforced and some kind of hard brittle plastic similar to bakelite. Not delrin, at least not in the machines that era I have seen. They will wear badly with high loads, in the machines i saw that kind of nut was only used in printers and measuring equipment.

I'd probably use a silicone oil lube.