| | |
BMC-20 and the 4th axis I started working to add a 4th axis to my BMC-20 some time ago which begins in this thread. I've finally made some progress! (All pictures are clickable thumbs.)
The original plan: Get another dual axis board, add a drive & servo motor, attach servo motor to a 6" or 8" rotary table, and do my 4th axis work.
Following the plan, a dual axis card was sourced from ebay, and plugged into the rack.
As advertised, the control found the "extra" two axes. The 'a' axis showed up on the manual screen but wasn't available anywhere else. This is when I realized I've been using the machine wrong for the last 2 years. When you start a new program, you're supposed to clear everything out of the control. This gets you a set of config screens where you pick Inch/Metric, and 3/4/5 axis. I spent quite a bit of time before I figured this out, and I was really worried that my new dual axis card was no good.
The next step was to find a drive and motor. I had pretty much decided to use another Fanuc analog AC drive & motor. I planned to use my spare axis motor that I got while I was first troubleshooting the machine, and get a drive (which Fanuc calls a Velocity Control Unit) from ebay. Some info about how the control expects the axis to work was that the control expects 180000 counts for one rev of the rotary table, which is what you get with a 2000 PPR encoder motor directly driving a 90:1 rotary table. Since the axis motors have 2500 PPR encoders I would need a small reduction which I planned to do with a timing belt.
While I was pondering this I found a really good deal on a Model 5 motor with 2000 PPR encoder. This is the next size up from the Model 0 motors used on the axis.
While I was waiting for the new motor to arrive I ran across something locally that should make everything easier: It's a Nikken CNC200 rotary 4th axis. It's already got a Model 0 2000 PPR Fanuc motor, an air operated brake, and all the needed limit switches. This seems to be a really nicely made piece of equipment. The worm is in an oil filled compartment, as is the drive from the motor. Backlash is quoted as 5-15 microns measured on the rim of the table, with adjustment needed with 50 microns (!).
I rang out the wiring so that I could get it set up for my machine, and cleaned out 20 years of swarf. That's why it's apart.
The 4th was filled with tiny, razor sharp swarf needles, some of which I'm still pulling out of my fingers. Here's the back cover taken off, as the brake was stuck with gummed up oil and the home switch was set up NC where Hurco needs NO.
Other than a broken oil sight glass which leaked gear oil all over the inside of my car it seems to be in fine condition.
I also got the drive that goes with the 4th. Here it is, mounted in the drive cabinet of the machine. It's probably not the best idea to add another drive in this cabinet because of the heat. I will add another circulating fan. If I need to I will add another heat exchanger, too.
In this picture I have wired everything but not yet finished the connectors, in case I need to make some changes. For the control & encoder signals I used the same specification twisted, shielded wire as used on the other axis wiring harnesses. You can see the extra relay for the brake solenoid valve behind the drive.
The 4th needs air at slightly lower pressure than the machine for its brake. I installed a second air regulator with a quick connect fitting.
Now we're ready to go. I decided to take small steps. I hooked up only the encoder from my new spare Model 5 motor so I could make sure that the drive & the control could see the encoder pulses.
Trouble! The drive immediately throws a TG alarm, meaning it can't see the encoder. This was easy to fix, the plastic disk with the pins inside the military connector wasn't installed all the way. This cleared the TG alarm on the drive. More trouble! The control won't count the encoder pulses. Some detailed checking of the dual axis card showed that the suggested jumpering would not work. Changing the jumpers a little fixed this, so now the control counts! We're making progress!
Now its time to run the motor under power. I put the encoder back on the spare Model 5 motor as I didn't want to risk burning up the motor on the 4th. I hooked up the spare motor to the new drive, double checked all the wiring, and powered it up.
Now the real trouble began. Press manual, power, start gets the usual klunk from the contactors, then something different: unbelievably violent vibrations from the servo motor! I mean, grab that motor before it vibrates off the bench and hits the floor violent. I ended up strapping the motor to the bench to keep it from wandering all over.
Time to start troubleshooting again. These Fanuc drives have 2 connectors that go to the computer side of the machine, one for control signals and one for encoder signals. I disconnected the 'a' axis signals from the 'a' drive and substituted the Z axis (leaving the actual Z axis drive not connected to anything). It powered up and was quiet though I could feel the motor "trembling" slightly. Jogging the Z axis ran the 'a' motor smoothly.
Next, I put the Z axis plugs back on the Z axis, and I unsoldered the voltage command wires from the 'a' axis plug substituting them with a jumper to hold the voltage command to the drive at 0. Plugged in the 'a' connectors and powered up, the motor was completely noiseless. I adjusted the drive's gain and offset with this 0 voltage command.
I then resoldered the 'a' voltage command wires. It's now back to the correct wiring. Powered up again, when stopped the motor is still vibrating violently. When I jog the axis the motor runs completely smoothly as long as its turning - as soon as it stops it goes back to violent oscillations. During the oscillations/vibration the 'a' position readout on the control does change back and forth a very small amount, around 0.050 degrees, but it doesn't constantly increase or decrease, it always stays around the same value.
I wondered if the encoder for the spare Model 5 motor was not working correctly, or I had installed it incorrectly (I did follow the procedure in the Fanuc manual for aligning the encoder). I hooked up the motor on the 4th axis instead of my spare motor and it vibrated exactly the same way.
I did notice that each revolution of the spare motor shows a 2 degree change on the control, instead of the expected 4 degrees. I'm not yet sure this is important.
At this point I'm kinda at a loss. I believe the drive and motor are working correctly, as I showed by running it on the 'Z' axis of the control. I believe the encoder A & B signals are reaching the dual axis card, as the control counts when I turn the encoder by hand.
I don't want to hook up the 'a' axis of the control to one of the X/Y/Z axis because the vibration is so violent that I'm afraid that it will break something. I suppose I could take the belt off the X or Y axis motor pretty easily, which I will probably do tomorrow. This could prove that the 'a' axis dual axis board & wiring are OK or giving trouble.
This oscilloscope picture was taken during the vibration:
The yellow trace is the torque command from the drive - it's what the motor is trying to do. The green trace is the voltage command from the control (the blue and pink traces aren't hooked up to anything). Note the "notchiness" in the green trace, which is the voltage command from the control. When the control is actually command the motor to turn, such as when the axis is jogging, the voltage command is a nice straight line as expected.
The second oscilloscope picture is with the green probe hooked directly to the voltage command output on the dual axis card (the other 3 traces aren't hooked up). I'm certain that the control is commanding these voltage outputs - noise wouldn't have that notchy, stepped output. Exactly why the control is commanding such large motions (about 1/2 volt) when the motor should be stopped I can't quite figure out. I also suppose I could also scope the encoder A & B signals to see how far the control is moving the motor. I will do all this tomorrow, if I get the chance.
Until then any suggestions are welcome!
Ugh. I have swapped cables & tested at great lengths and I have determined that the 'a' axis dual axis card is the cause of the servo vibration. I also might have damaged the new servo drive. I hope I didn't break anything else.
Oh, the control reads the rotab's rotation as 1/2 the actual rotation as I expected. When the table turns 180 degress the control shows 360 degrees.
I guess we'll see what's next tomorrow.
Aaron, the relay panel in your machine has a set of switches on it (SW4) that if you change number 4, it'll change the rotary resolution between 2000/4000 count per degree. Sorry to hear you are having problems with setting this up.
Thanks Bloke! That will make things easier! I can't thank you enough for all your support!Originally Posted by bloke
I am sure that I'll get this sorted. I really hope I didn't harm the servo drive, that's why I was upset. I do have an extra control card for the single axis servo so even damaging that can't be too difficult.
This is my next step: I'm going to try the 2nd axis of the dual axis card which would connect to the 'b' axis. I may need to make a jumper plug to show the control that 'a' is still present. I hope I can then run the rotary 4th as if it were the 'b' axis. This will show if there's trouble in the 'a' channel of the dual axis card.
I'm almost certain that the trouble is in the 'a' channel of the dual axis card. At the card cage, I plugged the 'Z' DB25 wiring into 'a' and the terrible vibration moved to the 'Z' servo. When I plugged 'a' wiring into Z's DB25 the 'a' axis ran smoothly.
Last edited by fasto; 05-24-2011 at 08:49 PM. Reason: Spelling
Oh no, my CRP (415-0224-001 Rev C) doesn't have an SW4. I have SW1, SW2, and SW3 which are all 4 position dip switches but no SW4, not even a place for it to be installed.
The good news is that I didn't harm the drive as I was worried.
The bad news is that when I plugged the 25-pin connector into the 'b' axis output of the dual-axis card the control showed a position for the 'b' axis, but the terrible vibration from the servo was the same.
My bad! The doc I was looking at had it down as SW4 number 4 but should have said SW3 number 4.
SW4 is the little toggle switch for enabling/disabling the toolchanger.
Thank you again. I do know of the little switch for the toolchanger. I didn't think to look at its reference designator.
Considering everything that I have found and worked on, I believe that the 4th's drive & motor are working fine and the wiring is fine up to the DB25 connector. Since both 'a' and 'b' act exactly the same, I think the dual axis card is OK electrically.
This leaves the position control loop as the source of the trouble. I've done a lot of work with control systems over the years and the oscillations around the stopped position that this system is doing are indicative of too large Integral gain in the PID loop (if that's what is used), or of too high overall loop gain, or of too high an acceleration factor. These settings might not be adjustable in this system.
I've been following your thread since we're working on similar projects. I'm sure cliff will chime in and tell EXACTLY what's what, but anyway.
On my dad's machine the gain and acceleration factors are found in one of the secret screens, 101, 102, 103, 104, I can't recall which one, but it's there!
In the Diagnostics page press for example Enter 103 Enter.
I remeber seeing something on one of the secret screens, and I'll have a look tonight. I may also try changing the 2nd set of dip switches, the one that doesn't set the backlash comp. Do I recall that the servo power needs to be ON to get into the secret screens?
I've also been following your work on your 4th, Gregor. Sounds like things will be somewhat simpler for you.
Well, the axis acceleration & gain factors are on page 103 on my machine - that is, X Y & Z, no sign of 'a' or 'b'.
Just a notion, on the a/b dual axis card is there a switch block for backlash like the other dual axis cards? Just in case the control is making it oscillate by overshooting for backlash that isn't there. Scoping the command voltage at the board with the encoder feedback disconnected might indicate if this is the case.
Also, if it works nicely off the z axis I think by now I'd probably be copying the switch block positions off that board by way of experiment. As long as you can put them back again it's not going to make anything any worse right?!
You've probably tried all that already...
It does have two switch blocks for each channel. One block sets backlash just like X/Y/Z, I don't know what the other does. On my Z, the "other" block is set to 0x88. My X and Y have the 2nd block set of all OFF. On the 'a' the "other" block is set to all OFF and I have also tried (in hexadecimal) 0x01 0x02 0xAA 0xF0 0x0F 0xFF with no changes. Backlash on 'a' I tried only all OFF and all ON, now that I think about it switch#8 inverts the backlash comp so I should try 0x7f too.
I thought I had checked all combinations that were like Z thouugh I see that I didn't actually do this, I will tomorrow. I checked the card output with the output disconnected from the servo and it looked fine, which is not quite the check you suggest and I will do your suggestion tomorrow.Also, if it works nicely off the z axis I think by now I'd probably be copying the switch block positions off that board by way of experiment. As long as you can put them back again it's not going to make anything any worse right?!
You've probably tried all that already...
Interestingly enough the 'a' marker doesn't seem to do anything, at least it doesn't seem to add an extra '.' on the control as the markers on X, Y, and Z do. For example X at the marker shows as X. while 'a' at the marker doesn't add that extra '.'.
I tried all manner of combinations of the 3 sets of switches on the CRB. I found out how to get a 24 or 30 tool magazine, how to get single vs. dual solenoid valves, how to get a 2 speed transmission with a variety of different ratios including 2:1, 1.5:1, and 1.333:1, and how to get a BMC30/40/50.SW3 number 4
I didn't find anything that changed the encoder counts on the rotary, unfortunately. SW3 position #3 and #4, as well as SW2 position #3 and #4 don't do anything at all. I tried all 4 combinations. All the other switches do something. Incidentally SW2 #3 and #4, as well as SW3 #3 and #4 are all OFF.
I wonder if this points to some kind of trouble on the personality board. When I got the machine the toolchanger was switched OFF, and in fact mechanically locked out. It also had a crazy set of limit switches that would e-stop the machine if the toolchanger moved away from HOME with Z not at toolchange height. I changed everything back to how it's supposed to be and I haven't had any trouble with the toolchanger aside a burnt relay for the magazine rotate motor. All the circuit boards were very dirty which I thought explained any trouble the previous owners may have had.
OK, a little further investigation. All the DIP switches that do something are connected to 4-input optos on the personality board. SW3 #4 is connected to U16.7, and the other 3 dip switches that don't seem to do anything don't seem to be connected to anything. I did not pull out the CRB to check which I suppose I will have to do, I checked with a meter. This means to me that SW3 #4 should do something. I will check tomorrow to see if the opto is OK and the 74244 that buffers the signal onto the bus is working properly. The personality card is a 415-0177-002 rev E. It's got 3 banks of dip switches and all are OFF except S3 #8.
Posting Permissions
| | |
LinkBack URL
About LinkBacks
