JMC Servo in PP mode

Hello,
I'm currently trying to run a JMC servo IHSV57/60-EC in contour position mode (PP). I get it to the point where power is enabled to the motor. At least I can feel that it's getting holding torque. But it doesn't move. Here are the objects and the values I set:

• 0x6060 = 1 (profile position mode PP)
• 0x607a = 1000 (target position)
• 0x607b.1 = 0x80000000 (position limit minimum)
• 0x607b.2 = 0x7fffffff (position limit maximum)
• 0x6081 = 1000 (contour velocity)
• 0x6083 = 5000 (contour acceleration)
• 0x6084 = 5000 (contour deceleration
  
  All other values are at their defaults (for example, electronic gear = 1/1, ramp type 0x6086 = 0 (linear ramp). When I bring the state machine to "operation enabled" (0x6040 = 0x0f) I get a positive acknowledge (0x6041 = 0x0637) and the motor has torque. When I set a new target position by writing 0x607a and setting bits 4 and 5 of 0x6040 (new setpoint, latch immediately) the drive responds with 0x6041 = 0x1237. This means that bit 12 (setpoint acknowledge) is set and bit 10 (target pos reached) is cleared. However, the motor doesn't move. Actual position (0x6064) stays the same as before (somewhere close to rero).
  
  I'm not sure if the velocity and acceleration values are meant in steps/s or RPM units. The manual is not clear about that. But 1000 should work in both cases, I think.
  
  I have mapped target position and control word as RxPDOs and status word, actual position and actual torque as TxPDOs. All other parameters are set during the Pre-Op to Safe-Op initialisation. SyncMode is set to free run (0x1c32.1 = 0 and 0x1c33 = 0).
  
  Has anybody successfully run an EtherCat servo in PP mode? Could you please share your initialisation sequence or tell me what I have missed?

That looks like a CiA 402 servo. Does it support CSP mode as well?

You might want to look at the `cia402` HAL plugin, it automates quite a bit of this, but currently only supports CSP and CSV modes, not PP and PV.

Hello benezan,

i tested a JMC iHSV57-30-18-36-EC yesterday and got it working in LinuxCNC without any problems or needing to configuring the drive via its RS232 interface.
Mine started up in OPMODE 8 (CSP) and was usable right out of the box. Unfortunately not all inputs are working as expected, just the CW and CCW are working as expected.

You can find my config and machine files here:
github.com/PedPEx/plasma-cutter-LinuxCNC.../main/JMC-Servo_Test

Hello PedPEx,

I'm currently setting up my own EtherCAT based controller as a mock up before I replace my current control board. Your single axis configuration really did help getting my JMC iHSV57-30-18-36-EC moving.

One question though: You are setting "setp cia402.0.pos-scale 3600". Do you have any reference where this value comes from? Can you give me a hint how it gets calculated?
My final setup has ball screws with 5mm/rev thread pitch.

Second question: Did you ever try to use the "basic_cia402" device type in the EtherCAT XML file and got it working?

Thanks,

Uwe

Servus Uwe,

at the moment i'm dedicated to learning for my exams next month, unfortunately i don't have the time right now, to revisit that topic. Eventually i will test that in august.

Personaly i would guess, the encoder has a resolution of 0.1 degree, so the pos-scale variable is just the maximum value, the encoder can reach.

For the second question, i'm not using the standard CiA402 XML-mapping, but i use the exact same syntax for all variables. I'm not entirely sure, why i went with that solution back in the day (i suspect, because i wanted to map the inputs/outputs of the drive too, which is not possible with the basic_402 defintion) and i do also not know, whether the basic_402 part of the XML file worked. As i said, eventually i will revisit that topic in late july or august.

Best Regards
Patrick

Hi Patrick,

so, regarding the encoder resolution I have come one step further. It actually is configurable and the default value is 4000 counts/rev.
I suspect that the drive doesn't have a "real" hardware encoder but uses hall sensors or other electronic tricks to know/estimate the position. Thus the resolution can be set kind of arbitrary and is dependent on the accuracy of the sensors.

I was able to set the resolution to 10_000/rev without problem. But the question remains how accurate that actually is and if such a high resolution makes any sense. I think I'll settle with just 1_000/rev which will still give me an accuracy of 10µm in combination with my 5mm/rev ball screw.

If you're interested in more details, let me know and I will share what I found out.

Regarding the XML configuration I'd be interested to see what you came up with.

Good luck with your exams,

Uwe