Sindrian Arts

>Apologies for lapse in posts, but between wrapping things up in Boston, moving back to Cincinnati, starting classes back at DAAP, and being really sick twice, I’ve forgotten to make time to post here.  However, I’ve recently decided to get my act together, and part of that is going to be posting here at least once a week.
Here’s what’s happened since my last post:
That standing computer desk I was working on turned out beautifully:

Leaving out a bit of trouble-shooting, it only took 45 minutes to mill all the parts, and that’s with multiple passes.  With a more ridged machine, I could easily half that.  Design-wise, it was a little wobbly, but I think I know a neat way to fix that with a few eye bolts and some rope or cable.  Also, I discovered that it should probably be a bit shorter, since one girl I showed it to said she had a hard time seeing the screen.
After that, I made a few modifications to the Blackfoot pattern to integrate the additional bearing holders for the X axis drive shafts and to eliminate the parts I wouldn’t need before making the g-code for cutting it.  By the time I was actually ready to run the code I was running short on time, so I ran it a little faster that I probably should have.  The pieces came out with a few defects here and there, but nothing that would prevent them from fitting together correctly.  The next day I packed up all my stuff, and left the morning after that.  Thus ended my wonderful travel quarter.  I was sad that I didn’t get a chance to say goodbye to a lot of the folks at the Asylum, but I’ll be back.
Ok, I’ve got to go grab some lunch.  I’ll post the rest either later today, or in a day or two.

>Yesterday was a comedy of errors.  The day started well enough with some test cuts of small pieces to test slot clearances, and all went well.  Then I decided to mill out something big: the standing computer desk:

Now, while for the most part it fits together without hardware, the adjustable keyboard rack does require the drilling of the holes.  After figuring out how to get poor ‘ol CamBam (free version) to drill holes (and testing this), I started the program.
It started out ok, but then by the third hole it started generating a good deal of smoke.  It took lots of experimenting with feed rates, pecking, and other such stuff to figure out what was going on:  I was trying to drill holes with a downcut bit.  
For the non-machinists out there, allow me to explain: your standard drill bit is what’s called an upcut bit, that is the spiral flutes of the bit will bring chips up to the top of the bit.  This is good for high-speed cutting, where you have to worry about keeping the bit clear.  However, the drawback to using upcut bits in a CNC router is that it will cause the face of your material to chip and break near the cut, especially if you’re working on a laminate material like plywood.  For this reason I bought downcut bits, where the flutes of the bit spiral the other way.  This eliminates chipping on the face, but has other drawbacks, such as not being good at drilling holes.  What was happening to me is that the MDF dust created by the bit was being kept down in the hole where it was compressed and rubbed against itself a high speed, causing it to burn.  To fix this, I changed the program so that I’d only drill shallow holes, the plan being to take the pieces over to the drill press and finish them after they were cut out.
Just before I was going to start cutting the actual pieces out, suddenly everything shut off.  Turns out someone in the wood shop next to me blew a fuse.  To avoid this in the future I used an extension cord to plug into a different outlet.  I thought nothing of this at the time, but it would haunt me the rest of the night.  After that I could not get the machine to do what I wanted.  It wouldn’t always move when it was told, and nothing I did or changed seemed to help.  At 11pm I decided to call it a night and went home.
I came in this morning and after implementing a clamping system I thought up to keep the workpiece from moving (we thought that perhaps this was causing some of the error), I got back to work.
As one would imagine, the same problem was still there, but this time I noticed that it seemed to be limited to the Y and Z motors.  I checked the connections, moved cords around to avoid possible interference (even though they were shielded), but no dice.  The most annoying thing was that it would work fine until I actually went to run a program.  A stroke of luck finally showed me what was going on:  I turned the router on and had the program start and then when I saw it was moving I turned on the shop vac.  I look back to see that the Y motor had stopped moving right when I had hit the vac’s switch.  Remember when I changed outlets?  Turns out the new one didn’t supply the same amount of current as the old one, so the motors weren’t getting enough juice when the router and vac were on as well.
They’re now on a different circuit.  Off to cutting!

>It would seem that after some teasing, Murphy decided to give me a break yesterday.
First, the motors arrived right when UPS said they would.  Installing one on the x axis went smoothly, but I ran into some problems once I started testing and tuning it.  First, when I tried to run it it kept stopping at the same place on the table and clutching out.  Turns out the set screw that was keeping one of the drive sprockets in place had loosened somewhat and was catching on the chain.  For some reason it now wanted to be tighter than it had been before, which meant grinding a new flat into it’s head to prevent further catching.  Once that problem was solved I started the process of finding the right amount of current to drive the motor.  Surprisingly, even with the current turned all the way up to 4.3A (the motor should only get 3A), it was still not able to move at the speeds I thought it should.  In looking for a cause of this, I had the first of the day’s breakthroughs. 
It turns out the dip switches are “on” when they are in the down position.  This meant two things: that the auto half current function (which helps keep motors from overheating) was off when I thought it was on, and that the boards were set to microstep at 25 subdivisions instead of just 5.  While microstepping does give you much more precision (which I don’t really need), it also severely reduces the amount of torque (which I do).  I turn the auto half current function, set the driver to 2 subdivisions just to be on the safe side, modify the necessary settings in the software to reflect this, and resume testing.
The torque of the motor improves significantly, but it’s still not performing as well as I’d expected.  I could get it up to about 9 inches per second, but higher than that and it would choke, regardless of current.  Now, when I was running my previous jobs, I had the machine running at feed rates of 50 and 100, so I was understandably concerned.  I had no idea what was wrong, and was starting to get very frustrated.
Then it hit me: the units are different.
See, those feed rates of 50 and 100 were inches per minute, not second, so what I thought was a slow 9 was in fact 540 IPM, way faster than I’d ever need to go.
It’s amazing how one small realization can improves one outlook on life.
After that, it was all cake.  I re-tuned all three motors and cut the next prototype of the card dispenser.  The y axis motor got a little on the warm side, so I had to let it rest half-way through the job, but I can turn the current down on that one with torque to spare.
There were some issues with the card dispenser job itself, but that was because of how I had it cut, not any hardware problems.  I assembled it today, and it works quite well:

You can find other pics of it, along with screenshots of other things I’ve modeled in Solidworks that are destined to be milled out on a CNC here.

>I had to order two new motors today.  One to replace the z axis motor that got too hot the other day, and one to replace the x axis motor that got too hot today.  Not sure exactly why, since I thought the current was turned down on that one, but perhaps not.  The CNC is again out of commission until the motors arrive, which with any luck should be before the weekend.
*sigh*
Ah well, time to do some more CAD modeling.  I’ll have plenty to mill once it’s up and running again.

>A quick update today:
On Friday I milled out parts for a prototype business card dispenser.  Check out the flickr set here.

>Yeah, damn clichés.
Yesterday, as you all saw, the machine was up and running.  However, by the time I left it was once again non-functional.  I’m pretty sure what happend was that the current was turned up to high on the z axis motor, causing it to overheat and eventually burn out.  Usually I keep an eye on motor temp using an infrared thermometer, but it’s batteries died.  Thankfully my good friend Calvin had a small stepper I could borrow until I make sure the other motor is dead, or it gets fixed.  I picked it up from him this morning, and that was one problem licked. 
Another issue I had yesterday was that while the machine did work, it was painfully slow; that one part took around 45 minutes to mill out.  This was because the router I was using could only spin at 500 RPMs, which meant I could only make the machine move at six inches per minute.  I can’t afford to spend that much time waiting, so I decided to bite the bullet and buy a bigger router.  A wood working supply store near where I live happened to cary exactly what I needed: a replacement motor for a Porter Cable 3 1/4 hp router.  Of course, about an hour after I bought it I get an email from my friend Jeff telling me that he has a mailer from that same store that’s advertising a replacement router motor for $30 less than what I paid, plus a $25 gift certificate.  I call the place back, and they told me that if I come back on Saturday (when that sale starts) with just the receipt,  I can get the discount and the gift certificate.  Two problems licked.
I get to the Asylum, rebuild the z gantry using the parts designed to hold the larger router, slap it back on the machine, wire in the new motor, install that, and start testing.
The new motor won’t drive the gantry.  Instead it just clutches out.  No worries, the current is probably turned down too low, all I need to do is turn it back up until it has enough power.  After much headache, I realize that the potentiometer on the driver board that is used to adjust the current is ever so slightly broken, so that the board doesn’t work unless a slight pressure is applied to the dial.  A small screw and some electrical tape later, and I’ve MacGuyvered it back to working condition.  Third problem licked.  
Now that it’s all working again, I’m calling it a day.  We’ll see what happens tomorrow.

>Look!

My CNC made a thing!
It’s a replacement for part of the vacuum attachment that was originally milled upside-down.  As you can see here, the lobed part that hangs beneath the router has a channel that connects the two openings.  That channel is supposed to face downwards, but if you flip the part the bolt holes don’t line up.  To fix this, I simply milled out another copy of the part and used the current piece as a guide for drilling the holes.

Much better!

>Finally!
My CNC gantry router is finally ready to start making stuff!  I’m sure I’ll still do a lot of tweaking as I go along, but tomorrow I’ll actually be able to start making things.
First order of business will be to go back to Atlantic Carbide to talk to their Onsrud rep, who apparently is only there between 6 am and 2 pm.  I need to find out what the best bit for cutting masonite hardboard, which I plan to use in my first project: a business card dispenser.  I’ve already done the CAD modeling and part layout; now all I need to do is to write the g-code.  I should have a prototype to show off tomorrow!

>So what happens when you go and say that something will be finished on a certain day?  Murphy shows his ugly head.
Damn you, Murphy.
See, everything was ready.  I had bought an extension hose for the shop vac, and onsrud cutters for the router. I had even done testing!  Apparently, I hadn’t done enough.  Here’s what happened:
I got the new motor wired up to one of my cables, poped it on the y axis (it had a longer drive shaft than my other motors, so it fit better there), connected everything, and loaded up the config wizard to make sure my settings were right.  X and y axises still worked fine (so the new motor was good), but when I tried to test the z axis, nothing happened.  I called someone over to listen to the motor while I tried to drive it, and apparently it would make a faint humming noise, but that was it.  I disconnected it from the z axis lead screw, and it still didn’t move.  Out came the multi-meter.  The cable was good.  The port on the electronics box was good.  I tested the motor on a different stepper driver board and it worked just fine, so it wasn’t the motor.  This leaves (as far as I know) only one thing that could be failing to work: the z axis stepper driver board.
Some good news:  The guy who built the electronics is going to be here soon, he should be able to accurately diagnose the problem.  With any luck, it’s something silly I didn’t know about, and he’ll fix it.  Also, I have an extra driver board that only needs very minor repairs to be fully functional, so even if the current z axis board is busted, we’ve got a replacement; no waiting for shipping.
Wish me luck!

>The replacement motor has just arrived!  After I finish this post I’m off to buy some Onsrud cutters for the router and an extension hose for the shop vac, I’ll slap the new motor on the z axis, and I’ll be OFF AND MILLING!!!
It’s about damn time.
Over the weekend I did some testing and tuning using only the x and y axises and a sharpie mounted in the router.

As you can see, with the first run (at the bottom), it started off pretty good, but as it progressed the x axis kept slipping.  The problem was that since the set screw in the drive sprocket was so small (#8, no head), I couldn’t really torque it down tight enough without stripping the socket and/or allen wrench.  Thankfully my friend Calvin was around, and offered me a couple socket-headed screws that took a much larger allen wrench, allowing me to apply plenty of torque.  While the larger heads did rub the chain, a little quick grinding with my dremel took care of that.  As you can see, the final test showed no slippage at all.
Another mod I implemented at the suggestion of others was the addition of two brackets to help support the x axis drive shafts:

We had noticed that when the drive chains were tightened, it would cause the 1/4″ drive shafts to flex.  To counteract this, Jeff designed and milled out several pieces of MDF to hold an additional bearing for each drive shaft.  This way, the shaft would be supported on either side of the drive sprocket, taking most of the lateral tension off the sprocket.  They work great.
Off to finish the machine!