What's new
What's new

Help Needed: AMC Drive Tuning with Heidenhain TNC 355

BugRobotics

Stainless
Joined
Jun 22, 2015
Location
Denver, CO
I've been attempting to replace a faulty drive with an AMC drive (DPRAINIR-015S400). It's a tad undersized for the application but I'll be upgrading to the 030S variant if I can get this working.

Some specs:
Machine: Hermle UWF851
Control: TNC 355
Servo: Siemens 1HU 3076 with tacho

Currently I have the drive enabling fine. Tuned per the AMC instructions. Feedback is setup as an analog input and is scaled via a voltage divider so as not to exceed the maximum +/-10V input range. Command is setup as velocity and is coming in another analog input.

Velocity input scale: 222.22 RPM/V Current Offset that works best (-0.017V)

Tach input scale: 202.51 RPM/V

I can move the drive around with the manual controls most of the time until the control throws an asterisk (*) letting me know it's waiting to reach position. I can usually do a quick rapid (in the same direction) and the asterisk will go away and normal operation resumes. Same thing occurs when running a simple program (waits to reach position with asterisk displayed sometimes a few moves in but always at the end of a move)

I thought I had it dialed but then tried to run a probe cycle. After staring the cycle, I touched the probe to simulate hitting an edge. The control threw up the asterisk and the control froze, necessitating an emergency stop and restart of the drives. Following that restart I got a "Processor Check Error 1". I've never seen that before.

Messed around with almost everything I can think of and I'm out of ideas. Missing something though. Any help is greatly appreciated.
 
I had similar symptoms with my TNC355, it turned out to be signal quality issue from TNC to the Bosch servo amps, hadn't seen that error message though.

What I suspect is happening is that at the end of the move, the "missing" distance is so little, that the max voltage increase the TNC can put out is too little to get the amp to make the physical move, and TNC cannot further increase the voltage because otherwise it will not be able to stop the axis in time before it overshoots. When I was diagnosing this problem, I was adjusting a pot on the amp to match zero on the amps +/-10V input with zero from the TNC, in my case it would help temporarily, but still would "hang up" sometimes, and it didn't matter from which side it would approach the stop point

There is a section (C3~4) in this manual related to tuning axis control - https://product.heidenhain.de/JPBC/image/FILEBASE_PUBLIC/208667_00_a_02.pdf
 
Bug:
Got your private message, sorry for not responding until now.
Have had some exposure with the AMC drives, but in no way am i comfortable with all the ins and outs of them.
They do have support and real humans that you can talk to right here in the US.where English is their native tongue.
Sorry , i really can't help on your Heidenhain setup. No experience there. The AMC setup i worked with was coupled to a Dialog 4 control.
Have you talked with Heidenhain directly?
Cheers Ross
 
Thanks for the replies gents. Much appreciated.

I believe you are correct jz. I've gotten the lag down to 0.0002" but it's this last little bit of motion that is screwing me. Were you ever able to fix?

I'm gonna turn down the accuracy a bit to see if that helps. Currently set to 0.005mm.


I've spoken to AMC, Heidenhain and Hermle. Heidenhain wasn't interested in helping and told me what I already know (analog output). Hermle was helpful as always but they don't have anyone stateside who knows anything about this older control so everything is going to Germany and they just sent over the commissioning doc. AMC has been excellent but the tech I've spoken with doesn't really know about Heidenhain integrations.
 
I did fix the issue on my Mikron, some previous owner had some sort of customization done to the machine, there were non-OEM contactors installed in the electrical cabinet, they switched the signal coming into the Bosch amps from TNC to some external source (wires were cut inside the control cabinet) that didn't come with the machine, those contactors were bad, as soon as I eliminated them and rerouted the wiring as in diagrams from Mikron, the whole problem went away and it has been running flawlessly ever since.

But I don't think that is the case with your machine, I suspect you'll have to adjust some parameters responsible for acceleration/deceleration to match what is going on physically on the machine - I think the control doesn't know that the motor now stops the axis sooner than expected, and the signal it is issuing to travel that last tiny bit is just too small for the amp to react to, or it is trying to, but the undersized motor just can't overcome the friction - in this case the control (and amp) will not continuously try to increase signal voltage to the amp (and amp to the motor) because of the danger of overshooting the stop mark.
 
I did fix the issue on my Mikron, some previous owner had some sort of customization done to the machine, there were non-OEM contactors installed in the electrical cabinet, they switched the signal coming into the Bosch amps from TNC to some external source (wires were cut inside the control cabinet) that didn't come with the machine, those contactors were bad, as soon as I eliminated them and rerouted the wiring as in diagrams from Mikron, the whole problem went away and it has been running flawlessly ever since.

But I don't think that is the case with your machine, I suspect you'll have to adjust some parameters responsible for acceleration/deceleration to match what is going on physically on the machine - I think the control doesn't know that the motor now stops the axis sooner than expected, and the signal it is issuing to travel that last tiny bit is just too small for the amp to react to, or it is trying to, but the undersized motor just can't overcome the friction - in this case the control (and amp) will not continuously try to increase signal voltage to the amp (and amp to the motor) because of the danger of overshooting the stop mark.

Yea, I agree, not the same. My machine is stock except for the new drive. You mentioned an undersized motor (mine is stock). I'm going to tune the velocity loop at a much slower RPM to try and dial in the behavior at the slower speeds. Any other ideas are welcomed! Thanks again.
 
I may have fixed the issue by tuning at the lower speed. Originally I tuned the drive at around 300RPM and then again at 500RPM. This time I very carefully tuned the drive at 50RPM and the issue with the control waiting to achieve the correct position seems to have been resolved. Fingers crossed I'll do some test cuts shortly and then some bosses to ensure everything is tracking correctly.

Thanks again for the responses,
Skyler
 
Look at:

offset, if the bias is such it will not be able to overcome it due to:

accel/decel curve
make it a bit sharper

there might also be parameters for deciding when it has reached position

you might also be too fussy about trailing error

you should adjust for minimum attainable lag at rapid after it has reached speed
 
Look at:

offset, if the bias is such it will not be able to overcome it due to:

accel/decel curve
make it a bit sharper

there might also be parameters for deciding when it has reached position

you might also be too fussy about trailing error

you should adjust for minimum attainable lag at rapid after it has reached speed

Appreciate the reply gus. Offset is pretty simple with this drive and I adjusted roughly -17mV to get everything balanced right out of the gate. I was going to try manipulating the accuracy window from 0.005mm to 0.007mm but the tuning at 50RPM has seemingly fixed the issue I was having. I'm assuming my tune at 500RPM was too coarse and causing my issues. Just ran a probing cycle and all was well there. Cutting some parts and then will update the thread.
 
Happy (relieved) to report the machine made it's first chips with the new drive last night. Drove it with some HSM tool-paths and it chewed through a some 304 SS parts just like always.
 
I'm attaching my AMC setup file for the Y-axis in case anyone else happens to need a starting point or reference.

Important notes for this installation:

  • You must disable the axis of interest (parameter 72) in the control to allow tuning (IMPORTANT: CONTROL WILL NOT MONITOR AXIS MOTION!)
  • Be vigilant during the first tuning cycle. Before supplying servo power to the drive, center the axis and be sure to have your hand on the E-stop in case there is a runaway situation due to switched motor leads or tacho input. A runaway situation is not guaranteed to occur on first movement and the control is not monitoring the axis so you are in total control. VERY IMPORTANT: YOU AND/OR MACHINE COULD BE DAMAGED!
  • The Heidenhain's internal offsets must be cleared anytime you make a change to the drive with respect to any scale/offset/tune/etc. (Enter code "75368" and then "NOENT")
  • Set the machine in "LAG" position mode and set parameter 252 (auto offset adjust) to "0" while tuning. This will allow the control to quickly adjust to offset changes (see first note)
  • Tuning at a slow speed, in this case at 50 RPM, was critical to eliminate the issue with the controller waiting to achieve proper positioning. Tuning at 300RPM and 500RPM resulted in strange errors messages ("Processor Check Error 1") and control hangs when moving at slow speeds that were not entirely repeatable.
  • The tacho feedback is 20V/1000RPM and requires a scaled input. I did this with a 4:1 resistor divider. (PCB files attached for making a DIY PCB). After creating your resistor divider, measure your actual ratio (no resistor is perfect) and input that into AMC drive.
  • No machine/control parameters were modified to incorporate the AMC drive with the system.

AMC Configuration File for Hermle UWF851 Y-Axis w/ TNC 355
Notes:
  • Requires Driveware 7 to view

4-Axis Tacho Scaling Circuit
Notes:
  • This file is a mirrored view used to create a DIY PCB. Gerbers can be attached if there is any interest, just send me an email.
  • Parts needed: 1935226 (8 pin Pheonix Contact with 5mm pin spacing) and 0805 resistors (10K and 3.3K)
  • This PCB has signals entering in on a connector and leaving on directly soldered wires. Not ideal but I only had one of the 8-pins on hand.
 








 
Back
Top