AM335X DC with 250us

Eddie Hsieh

Other Parts Discussed in Thread: SYSBIOS

Dear Sir,

We are developing the EtherCAT Slave by using the AM335X ICE. We use the latest slave code from ETG and the Accontis master. We do the test under RTX8.2.1 in WinXP.

In the DC 500us, the slave works well.

In the DC 250us, the master will report the slave "Out of Sync".

Does the AM335X has the limitation of DC timing? How fast can it achieve?

Eddie

over 12 years ago

0 PratheeshGangadhar over 12 years ago

TI__Mastermind 47541 points

Hi,

We have tested with Beckhoff PLC CX9020 upto 100us cycle time in DC mode. As well TwinCAT2.11 running on a Windows XP PC upto 250us (http://processors.wiki.ti.com/index.php/Running_AM335x_EtherCAT_Application_in_DC_Mode)

How many slaves are present in network? Do we have more details from master side? Any wireshark logs of the session will be of help.

0 Frank Walzer over 12 years ago

TI__Mastermind 44306 points

Eddie,

can you try another slave (e.g. Beckhoff) at 250us DC to see if there is no issue on the master side?

Otherwise you might need to work with your master vendor. We do not have that configuration at TI.

Regards.

0 Eddie Hsieh over 12 years ago in reply to Frank Walzer

Intellectual 290 points

Sir,

I only use one TI AX3335X ICE module connected the master controller. The master controller runs the RTX8.1.2 + Acontis EtherCAT master.

I also use this master controller to communicate with Yaskawa Servo Drive working at 250us DC. It is OK. The master should be OK.

As I know, the DC is controller by the master and TI's PRU code. So, I will think the problem maybe caused by the slave protocol stack or ESI file. Is it correct?

The TI AM335X ICE module has tested at 250us DC in XP OS and Beckhoff controller. Has it?

If the answer is YES, what version of slave code do you use? I use the am335x_sysbios_ind_sdk_1.0.0.81/sdk/ecat_app.

Thanks!

Eddie

0 Frank Walzer over 12 years ago in reply to Eddie Hsieh

TI__Mastermind 44306 points

Eddie,

we can only test fast DC modes against Beckhoff PLCs (e.g. CX5020) as Windows XP runtime of Twincat is not reliable enough. Obviously we test all our releases and 1.0.0.8 is the latest.

Would it be possible to provide us a wireshark log of the test with ICE and the Acontis master?

Regards.

0 Eddie Hsieh over 12 years ago in reply to Frank Walzer

Intellectual 290 points

Sir,

I have Beckhoff CX1020 running in WinCE. I will do the same test on this platform.

I also need to prepare the switch Hub for doing the log. I will connect EtherCAT master controller, slave device TI ICE, and log computer switch hub. Then, I run Wireshark on the log computer to do the log. Is this topology correct?

I review the following link. It seems only change the master controller task timing. It doesn't show how to modify the DC timing in DC Tab of slave device.

(http://processors.wiki.ti.com/index.php/Running_AM335x_EtherCAT_Application_in_DC_Mode)

Eddie

0 Eddie Hsieh over 12 years ago in reply to Eddie Hsieh

Intellectual 290 points

5468.EtherCAT_250us.rar

6404.EtherCAT_500us.rar

Sir,

I log two DC timing for TI ICE board. One is 250us DC, it has the "out of sync" issus. I set the monitor window is 2us.

The 500 us DC works well.

We are verifying our designe EtherCAT slave board, most of function work OK. Now the only issue is the DC issue. Please help us to classify the issue.

Eddie

Processors

Processors forum

AM335X DC with 250us