Acontis and Kollmorgen MKD

[tecnico]

Hi all,

I have a very weird issue with Ethercat. We moved from a bunch of AKD drives with shared DC bus to the multi-axis MKD from Kollmorgen.

On the cabinet side it was a great improvement, minimizing all the power connections.

 

What I did not know is that the MKD behaves differently on the network and seems to shut down the Acontis master (on NYB2C-31PK1, but with ARM CPU is the same).

 

Upon a ecat reset command I can see the slaves (8 to 16 in my configuration) being detected while the communication leds are still flickering, but after that Masterstate has value zero.

 

Below the terminal reply with commencts between << >>

 

>>>>>>>>>>>>> all slaves connected >>>>>>>>>>>><

 

ecat slaves

ecat reset

Starting ACONTIS communication

Restarting ACONTIS stack

 

Licenses granted by /opt/etherlab/etc/sysconfig/.ethercat: 64

 

Ecat Type: 1; ECAT License for 256 axis

 

>>> done while leds flickering >>>>

 

ecat slaves

0 VID=$00000002 PC=$044C2C52 0:0 PREOP + ?s?

1 VID=$00000002 PC=$03F03052 0:1 PREOP + ?s?

2 VID=$00000002 PC=$03F03052 0:2 PREOP + ?s?

3 VID=$00000002 PC=$03F03052 0:3 PREOP + ?s?

4 VID=$00000002 PC=$03F03052 0:4 PREOP + ?s?

5 VID=$00000002 PC=$07D83052 0:5 PREOP + ?s?

6 VID=$00000002 PC=$07D83052 0:6 PREOP + ?s?

7 VID=$01000056 PC=$00002301 0:7 PREOP + ?s?

8 VID=$0000006A PC=$4D4B4331 0:8 PREOP + ?s?

9 VID=$0000006A PC=$4D4B444E 0:9 PREOP + ?s?

10 VID=$0000006A PC=$4D4B444E 0:10 PREOP + ?s?

11 VID=$0000006A PC=$4D4B444E 0:11 PREOP + ?s?

12 VID=$0000006A PC=$4D4B444E 0:12 PREOP + ?s?

13 VID=$0000006A PC=$4D4B444E 0:13 PREOP + ?s?

14 VID=$0000006A PC=$4D4B444E 0:14 PREOP + ?s?

15 VID=$0000006A PC=$4D4B4332 0:15 PREOP + ?s?

16 VID=$0000006A PC=$4D4B4333 0:16 PREOP + ?s?

ecat slaves

0 VID=$00000002 PC=$044C2C52 0:0 PREOP + Slave_001

1 VID=$00000002 PC=$03F03052 0:1 PREOP + Slave_002

2 VID=$00000002 PC=$03F03052 0:2 PREOP + Slave_003

3 VID=$00000002 PC=$03F03052 0:3 PREOP + Slave_004

4 VID=$00000002 PC=$03F03052 0:4 PREOP + Slave_005

5 VID=$00000002 PC=$07D83052 0:5 PREOP + Slave_006

6 VID=$00000002 PC=$07D83052 0:6 PREOP + Slave_007

7 VID=$01000056 PC=$00002301 0:7 PREOP + Slave_008

8 VID=$0000006A PC=$4D4B4331 0:8 PREOP + Slave_009

9 VID=$0000006A PC=$4D4B444E 0:9 PREOP + Slave_010

10 VID=$0000006A PC=$4D4B444E 0:10 PREOP + Slave_011

11 VID=$0000006A PC=$4D4B444E 0:11 PREOP + Slave_012

12 VID=$0000006A PC=$4D4B444E 0:12 PREOP + Slave_013

13 VID=$0000006A PC=$4D4B444E 0:13 PREOP + Slave_014

14 VID=$0000006A PC=$4D4B444E 0:14 PREOP + Slave_015

15 VID=$0000006A PC=$4D4B4332 0:15 PREOP + Slave_016

16 VID=$0000006A PC=$4D4B4333 0:16 PREOP + Slave_017

 

<<<< leds on the ECAT slaves stop flickering and stay still - MasterState = 0 >>>>>>>>

ecat slaves

 

<<<< no more replies to ecat slaves >>>>>>>

 

Did anyone use MKD with Acontis succesfully?

[Omron Forums Support]

Normally the only state change I see happen on it's own after ecat reset is INIT->PREOP. Then I don't see a state of 0 without trying to enable (ECAT[0].Enable=1). Do you know if a PLC may be trying to enable ethercat? (Not kollmorgan specific)

 

If you right click the logs folder in the IDE solution explorer, you will get an upload option. Do any of these logs have information that looks relevant?

[tecnico]

According to the tests performed together with Brad @ODT CH there could be some EEPROM issues causing the master to shut down at the very first scan after a ecat reset (or first power on).

 

With a workaround (i.e. connecting the MKD after succesfully enabling the network with a reduced set of lasves) it is possible to scan the network and even to enable it, but at the next power on the original problem remains.

[Omron Forums Support]

If you right click the Logs folder in the IDE and chose to upload log files, do you see any ethercat errors?

[tecnico]

After some further investigation with the help of DTCH we narrowed the problem - that is not solved yet -

 

Here is what I get with "system dmesg" from IDE terminal

 

[ 1093.097868] tEcJobTask[1757]: segfault at 4 ip 0807e024 sp aa5230f0 error 4 in EcMasterDemo[8048000+118000]

 

So there is something coming from MKD causing a segmentation fault in the master module

[tecnico]

After more than one month the issue has been figured out.

The Acontis stack had to be modified to tolerate a missing configuration in the ESC (register 0x410 was not implemented)

This, along with a modified EEPROM from Kollmorgen to fix some wrong mailbox settings, allowed to have a working ethercat network