[StickShifter]

I have a BP VMC 800/22 with a DX-32 control and Siemens Simodrive611 drives. The problem is that during a tool change the spindle seems to be searching for the proper orient position but never finds it. The spindle will go both CW and CCW while searching for the proper position. No alarms are set. Sometimes the spindle keeps searching even during the tool change. (tested without any tool loaded) In a possibly related observation, the spindle seems to search for which direction to spin at start-up, changing direction once or twice before going to full speed in the proper rotation.

I don't use the machine everyday but it had been working fine. Then the PC would not boot which turned out to be a bad motherboard. I changed that and got the machine back into service. Now this issue.

What I have done so far:
1) Removed the spindle encoder and tested it on the bench. All signals are fine and work according to the spec, at least the best I could tell.
2) Ohm-ed out the harness from the spindle encoder connector to the DB-50 connector on the BMDC(main embedded controller) card and all lines appeared to be fine.
3) Swapped the BMDC with one from my other mill. Tested by commanding a spindle speed and then a tool change, the operation of the spindle is exactly the same.
4) I ran the spindle at 1500rpm and the speed feedback was right on as displayed on the user screen.
5) I attempted a tool change with the spindle encoder disconnected which resulted it the motor going to full speed, which I stopped asap.
6) I put a scope on the analog speed control lines to the Siemens controller and this what I get during a tool change:


7) I tacked wires on the BMDC connector to look at the encoder signals while running and everything looks normal. I confirmed all six signals are present and accounted for.


So, what am I missing? What else could cause the BMDC to control the spindle in such a manner? It looks like the control loop is "broken" and it just keeps over shooting the target. It is seeing the encoder or the speed feedback wouldn't be accurate. And it must be seeing it or it wouldn't be driving the spindle that way during the tool change. It has the ability to stop the spindle since I can turn off the spindle from 1500 rpm and it stops almost instantly.

There are only two controls from the BMDC to the Siemens controller, the differential analog speed control and a control relay that disables the spindle and applies the brake. It appears that the BMDC just can't figure out where to stop the spindle. I did go into the maintenance program and found the parameter for the orient position which looked fine, the same as it was when I got the machine.

Any ideas?

Thanks for any help provided.

[machintek]

Have you looked at the marker pulse? It occurs once per revolution. It starts the counting process for orient.
Next, what is the drift? Command a S0M3. The drift should be less than 4 RPM.

[StickShifter]

Thanks for the response.

I did check the marker pulses, both the positive and inverted pulses are present when moving the spindle by hand.

I have no idea what the drift is. Never heard of it before in this context. So set the speed to zero and see what it does? I assume the spindle should stay stationary. And the drift is the speed greater than zero?

[jmullett]

Do I understand that the command signal sweep correctly +50% speed to - 50% speed over 500ms this seems way slow to me if its trying to hold the spindle in a stationary position then 1.5 seconds at +50% speed?

Is this TTL logic? Encoder signals are hovering around 3 volts, is it possible the voltage level is tapering off below threshold. I know some older gates are finicky about under voltages. Is your 5 volt supply all right?

[StickShifter]

Thanks for the reply.
During that input signal the spindle was going back and forth as if the controller was trying to find the proper position.

I also noticed the output voltage of the encoder signals. All signals are at 3.3V and that made me wonder. I checked the 5V supply but that was with the encoder disconnected. I thought the signals were a full 5V when I checked it on the bench but not so sure now. I will double check the supply voltage with the encoder connected later today.
Thanks.

[jmullett]

I guess what I'm trying to understand is why the command is so slow to change and why the magnitude is so high assuming its a +/-10V signal. I would think that a correct waveform should look like very small voltage and swinging very rapidly to hold the spindle in the tool change orient position in a proper running machine. Like maybe it is not getting all of the A B channel pulses or quadrature is messed up for the add/subtract to the counter. The fact that the speed display is reading correctly wouldn't necessarily tell us anything about the AB channels as it probably comes from timing the Z/index channel pulses in my estimation.

I'm trying to think is there a way to look at the spindle counter register in the DX32 and watch if it counts the pulses consistently when you turn it by hand. It probably is a ring counter and resets on the index pulse just as a guess.


The voltage is just a shot in the dark. technically the input should be determinate above 2 volts, but typically I would expect to see much closer to 5 volts from a TTL encoder driver. Sorry about your frustrations. At least it sounds like you have a second machine to get you through jobs though.

[StickShifter]

I messed with it some more and I am running out of ideas. I even changed the PC motherboard out of desperation. No difference.

I verified the 5V supply to the encoder and it is fine. I believe the 3.3V signal level at the BMDC is due to clamping the signals for the 3.3V logic levels on the BMDC.

So here are what the Z+/- pulses look like when turning the spindle by hand:



Here is a video of a tool change so you see what I am dealing with:


And here is a video after issuing a S0M3 command:
VMC800_zero_speed_zpsceahlkqg.mp4 Video by stickshifter75 | Photobucket

My next best guess is to change the encoder as I really don't know what else to do. I really don't like blindly throwing parts at it but I don't know of a better plan at this point.

[Jonp]

This is a wild guess but it appears the encoder is working and the motor is trying to orient it, is it that the brake isn't being applied at the proper time to hold it in position? Good Luck Jon P.

[StickShifter]

OK, I'm willing to say that the encoder is working.

But now what?

Does anyone actually know what the inputs and outputs are for a tool change?
My guess/assumption was that the BMDC would issue a command(analog voltage) to make the spindle spin until the spindle approached the proper position. At that time the BMDC would disable/brake the spindle controller in order to lock the spindle into position. But that is just a guess based off the wiring diagram and basic system knowledge. That also means that the number of inputs into the control loop are limited. Unless there is something else that I don't know about.

So, does anyone have real information as to what the software is trying to do and what inputs are being looked at?

Any other suggestions or other possibilities?

[andrewl]

3 phase power.. are all 3 legs working. All the way to the machine..

Sent from my HTC331ZLVW using Tapatalk

[StickShifter]

Yes, the AC supply is fine. As I said, everything else works fine on the machine. I can still use it if I don't do a tool change but that limits it big time.

[Jonp]

Is the mechanical connection between the encoder, spindle and motor firm ? no loose key ways, belts etc. My guess is it may be something simple that has been overlooked although I am not aware of how deep you have diagnosed the problem. Did you acquire the machine with this problem ? or did it just happen while using it? Jon P.

[StickShifter]

I have had the machine for 2 years and running it off and on for the past year with no issues. The PC mother board died and replaced it. This issue came up shortly after that. But I have no idea if changing the MB is related and one reason I swapped the MB again and ohm-ed out the encoder wiring.

The belt to the encoder was a bit loose when the problem first appeared so I thought for sure that tightening it would fix it. Nope.

I agree, it is probably something simple, but what? I plan on spending some more time on it today and will post anything I find.

Thanks.

[Jonp]

I can't tell you how many times with the old BOSS machines I had the card extender out with scope and data logger hooked up following signals and 9 out of ten times it was bad connection, micro switch, prox. switch, frayed wire etc. My last episode was about six months ago, picked up a late 90's TC3G hadn't ran for 6 or 7 years and intermitant before that, Y axis following error, previous owner had replaced Y servo motor, cables, servo drive, motherboard, and i'm not sure what else, along stupid me comes looking for a challenge, well I had a real challenge, after 2 days looking deep into the electronics I came to my senses, re-seat All boards check All connections and check All ribbon cables, Yup sure enough in a ribbon cable socket on the BDMC something looked funny, grabbed a O ring pick and pulled out a very small peice of that back anti-static foam like what new chips sometimes come in, the machine has ran flawless ever since. I know what your going thru, I don't know the details of your electronics but the later DX32 systems have a logic board just for the tool changer with many I/O and a link to the BDMC. Thats enough of my rambling, good luck Jon P.

[StickShifter]

Thanks again for the comments and ideas.

I have spent some more time reading the application guides for the SIMO drives and looking at the wiring diagrams. I have to admit that this would be a lot easier if I had access to the source code or at least a flow diagram of what exactly is supposed to be happening during a tool change.


Here is the wiring diagram for the spindle. As you can see, not many wires going to the controller. The main analog speed input is on terminals 56 &14. The only other control wires are to terminals 663, 65, & 81. Those terminals have the functionality as seen below.





It is my assumption that the controller provides an analog signal to get the spindle to start spinning. When it sees the correct position reached by using the spindle encoder feedback, it deactivates relay CR2. That in turn activates the brake according to the controller spec.

So I guess that breaks it down to a couple possibilities based on all the debug work I have already done.
1) The encoder does have an issue.
2) The brake system is not working.

Looks like I will be finding another encoder and double checking the enable/braking wiring and functionality. Since the spindle stops almost instantly from 1500rpm, I have to believe that the braking functionality is working. There just doesn't seem to be much left other than the encoder.

[Jonp]

So just a little checking this morning, on the Torq -cut the spindle has two encoders, one in the motor to keep track of speed and direction it goes to the spindle controller, the other belt driven is for coordinated Z axis cycles such as rigid tapping and tool change it goes to the axis I/O card, witch in turn is controlled by the BMDC. Maybe just for drill you could check the encoder in the spindle motor, it seems like some how or some where they must both have duty to do at the same time..... Jon P.

[StickShifter]

Jon,
The setup on the TorqCut is the same as the VMC800. And your description of their use is the same as my understanding. However, I see no way for the BMDC to receive any motor position information from the motor encoder. The motor encoder is plugged directly into the spindle controller and there are no connections back to the BMDC for that information. I have run the spindle at low speed without issue so I believe the closed loop speed control is working and has no bearing on the tool change. At this point I am going to get another encoder for the spindle and see what it does. I will either confirm the issue or eliminate the encoder as the problem.

[jmullett]

How did you make out on your orient issue? Did the new encoder fix it?

[StickShifter]

Still working on it. New encoder and cable made no difference. Used some test code and can see the encoder counts just fine. I now am waiting on a new old stock motherboard that is an exact match for the original. If that does not make a difference I don't know what I will try next.

[DarkStarMedia]

I have an issue on my machine where when I try to rigid tap the spindle rotates slowly and then the machine error's out. It orients fine for tool change though. I will have to look for that second encoder.

Given that the spindle keeps moving even during tool change I think is significant. For my tool changer (spun continuously) it was one of the electronic relays that I have to replace because it shorted out (I think). I wonder if there is a comparable component for the spindle that is used during a tool change.

Another thought... Given that my machine orients fine but cannot rigid tap, makes me think that the tool change orientation is handled differently than other orientation jobs (ie rigid tapping). I wonder if when the computer orders the tool change orient, the spindle drive handles it rather than the computer (or the other way around).

Just some thoughts that might or might not be helpful.

Oh, and changing the motherboard should have no effect either way.

Us Bridgeport CNC folks have to stick together

Richard