TMS320F28P650DK: E-Cat cannot be connected stably

Hi Joe,

I have some questions:

1. Are there any reported errors from the ECAT host side?
2. P65x ESC registers can be checked for errors as well, there is a long list:

   1. ESC Register

      0x100 ESC DL Control word 1

      0x102 ESC DL Control word 2

      0x110 ESC DL Status

      0x120 AL Control

      0x130 AL Status

      0x134 AL Status Code

      0x300 RX Error Counter P0

      0x302 RX Error Counter P1

      0x308 Forwarded RX Error Counter P0 & P1

      0x30C ECAT Processing Unit Error Counter

      0x310 Lost Link Counter P0 & P1

      0x440 Watchdog Status Process Data

      0x510 MII Management Control/Status

      0x518 Register PHY P0 Status

      0x519 Register PHY P1 Status

3. Is customer doing power cycle testing? How often does this issue occur?
4. After re-power cycling the board is the P65x ESC able to connect to the ECAT host again?

Best,

Kevin

Hi Kevin

The situation in customer side:

Step1: TwinCAT connect with LAUNCHXL-F28P65X properly and TwinCAT in OP mode.

Step2: Connect and Disconnect the Cable / press the reset button on the LP / Power cycle the board.

the LP cannot connected to the host properly, the LED in the RJ45 is turned off, TwinCAT is in the init mode.

Kevin Allen18 said:

Are there any reported errors from the ECAT host side?

ECAT host is TwinCAT, no error is reported, TwinCAT state is Init.

Kevin Allen18 said:

P65x ESC registers can be checked for errors as well, there is a long list

ESC registers are attached:

F28P65 debug.docx

Kevin Allen18 said:

Is customer doing power cycle testing? How often does this issue occur?

No the issue happens during the development, frequency 1/10

Kevin Allen18 said:

After re-power cycling the board is the P65x ESC able to connect to the ECAT host again?

Yes, power cycle the board/ reset MCU/ reconnect the cable can reconnect again.

Thanks

Joe

Hi Joe,

How many P65x devices are being tested in the network? More than 1?

Are they running TwinCAT on a Windows PC or a Beckhoff IPC? Standard Windows PCs can have issues sometimes. I use https://www.beckhoff.com/en-us/products/ipc/embedded-pcs/cx5100-intel-atom-r/cx5130.html in the lab here.

Joe Ji said:

Step2: Connect and Disconnect the Cable / press the reset button on the LP / Power cycle the board.

the LP cannot connected to the host properly, the LED in the RJ45 is turned off, TwinCAT is in the init mode.

Can you elaborate on these steps? 

• Which cable is disconnected and then reconnected? Power cable or EtherCAT cable to RJ45 port?
• The behavior they are expecting is for all the P65x devices on the network to go back to OP mode after being re-connected to the network, correct?
  • In my quick testing with the LPs + TwinCAT here it seemed to work OK. Did not do extensive testing yet however.
• Both LEDs on the RJ45 are OFF (Yellow and Green)?
• Can TwinCAT still identify the P65x devices or do they need to re-scan? If P65x devices are identified then can they be put back in OP mode?
  • Some images of TwinCAT view may help too

Joe Ji said:

ESC registers are attached:

OK, I will try and review but putting in a table for comparison is often easier to identify differences. Or exporting from CCS memory browser even.

Best,

Kevin

Hi Kevin

Please see my reply below:

Kevin Allen18 said:

How many P65x devices are being tested in the network? More than 1?

There are 5 pcs PCBA in customer side and 1 Launch-pad, all the devices can reproduce the issue.

Kevin Allen18 said:

Are they running TwinCAT on a Windows PC or a Beckhoff IPC? Standard Windows PCs can have issues sometimes.

They use the windows PC integrates the internet port.

Kevin Allen18 said:

Which cable is disconnected and then reconnected? Power cable or EtherCAT cable to RJ45 port?

RJ45 port

Kevin Allen18 said:

The behavior they are expecting is for all the P65x devices on the network to go back to OP mode after being re-connected to the network, correct?

Yes

Kevin Allen18 said:

Both LEDs on the RJ45 are OFF (Yellow and Green)?

Yes

Kevin Allen18 said:

Can TwinCAT still identify the P65x devices or do they need to re-scan? If P65x devices are identified then can they be put back in OP mode?

Do not need to re-scan, once the computer recognize, the TwinCAT can identify P65 and back in OP mode.

Kevin Allen18 said:

OK, I will try and review but putting in a table for comparison is often easier to identify differences.

Sorry for the inconvenience, if there is further data export requirement, I will let customer do that in register format.

Thanks

Joe

Hi Joe,

Thanks for answering my many questions.

Joe Ji said:

They use the windows PC integrates the internet port.

I think this could be the cause of the issue. Using a Beckhoff IPC is usually better since it doesn't have other software running on it unlike standard windows PC.

I can do some more testing to try and reproduce the issue with set-up we have here with Beckhoff IPC. If I cannot reproduce it then I think it is eCAT Master / PC related.

Best,

Kevin

Hi Kevin

Is it the same issue that you met before? Or you think it should be the issue caused by customer set up?

Thanks

joe

Hi Joe,

Sorry for the delay. Do you have any update on this case over the past week?

I have not personally seen or re-produced this issue with the set-up we have in our lab.

I think it is possibly caused by customer set up, since they are using a normal Windows PC and not an Industrial PC (IPC) like we use in our lab.

Best,

Kevin

Hi Kevin

To avoid the Windows PC issue, customer uses the servo motion controllers to connect the F28P65_LP, the issue appearance is different from using Windows PC, the RJ45 LED is correct, but the OP LED in launchpad cannot be lit up for some time, you can see the video attached, you can ignore the conversation in this video:

Support needs: is there any onsite debug actions I can take? this issue blocks the project going into mass production.

Thanks

Joe

Hi Joe,

Ok, I see. Run LED behavior is explained below, so P65x ESC is stuck in Init state in fail case.

Joe Ji said:

Support needs: is there any onsite debug actions I can take? this issue blocks the project going into mass production

I would suggest reading out the ESC and PHY registers in a PASS / FAIL case for value comparison to debug the issue. ETG also has diagnostic material that can be referenced.

When I have some time I'll see if I can re-produce what you've shown with set-up here.

Best,

Kevin

Hi Kevin

Could you please share the .out file to me through email? I will let customer download the .out file to see whether the same issue can be reproduced.

I want to exclude the customer setup issue with the .out file.

Thanks

Joe

Hi Joe,

Sent over email.

Best,

Kevin