[Etherlab-users] Timeout while setting state INIT.

Sun Mar 15 21:28:21 CET 2026

Hi Oguz,

I haven’t looked at all closely to the issue (and a couple of emails got blocked by our spam/virus filter), but if you suspect that the slaves SII information is incorrect then you can try supplying the SII information yourself, rather than the master downloading if from the slave.  If that then works, you can check for the differences between the SII on the slave and the data provided by the esi file.  If confirmed you can request the vendor provides an update for the slave with the correct SII information (or continue using the manual SII file).

We use quite an old EtherLab master with patches to allow loading your own SII file instead of the master downloading it from the slave (from gavinl’s patchset “0001-load-sii-from-file.patch”).  I don’t know what’s available in your master configuration, but in ours we copy the SII file to “/lib/firmware/ethercat”

To create an SII file:

  *   Start a new TwinCAT project (TwinCAT XAE Project (XML Format))
  *   Under I/O, Devices add an EtherCAT Master device
  *   If you don’t have the correct esi (.xml) file loaded:
     *   copy the slaves esi (.xml) file into: C:\TwinCAT\3.1\Config\Io\EtherCAT
     *   Select the TwinCAT, EtherCAT Devices, Reload Device Descriptions menu option
  *   Add you EtherCAT slave to the EtherCAT master device
  *   Open your slave
  *   Select the EtherCAT tab
  *   Select Advanced Settings...
  *   Select the ESC Access, E2PROM, Hex Editor item
  *   Confirm the SII contains the expected values (for Vendor ID , Product Code etc)
  *   Select the Write to File... button
  *   Save with a filename in the format “ec_{vendor_id_hex}_{product_code_hex}[_{revision_hex}].bin”
     *   The revision part is optional.
     *   e.g.: ec_00000539_02200001_00010000.bin

Regards,
Graeme.

From: Etherlab-users <etherlab-users-bounces at etherlab.org> On Behalf Of Bilko AS, Oguz Dilmac
Sent: Saturday, 7 March 2026 02:30
To: Richard Hacker <ha at igh.de>; etherlab-users at etherlab.org
Subject: Re: [Etherlab-users] Timeout while setting state INIT.

Hi,

We are still working on the problematic slave which doesn't go to INIT. I'm trying to understand the difference by examining wireshark logs.

I noticed these following messages are sent at the beginning. And the device changed to INIT itself. It seems as if there is something wrong with SM configuration?

FPRD: 0x80d Return value: 0x30 (SM error / sm enable acknowledged)

FPWR: 0x800 16 byte data: 0    (Clear SM0 and 1?)
FPWR: 0x800 8 byte data: 0010000426000100  (Re write SM according to XML file?)
FPWR: 0x808 8 byte data: 0014000422000100
FPWR: 0x810 16 byte data: 0     (Clear SM2 and 3?)

FPWR: 0x620 16 byte data: 00000009010000000d08000101000000 (define FMMU?)
FPWR: 0x500 1 byte data: 01

FPWR: 0x120 2 byte data: 0012 (PreOP + Error ACK)
FPRD: 0x130 Return value: 0x0002 (PREOP)

FPRD: 0x80D Return value: 0x00 (The error is gone!)

FPRD: 0x1400 Return value: 0x00 (but WC is zero. I guess the slave didn't processed it)

FPRD: 0x130 Return value: 0x0001 (INIT) (Slave went to INIT. But I didn't see a write to 0x120 to go to INIT)

FPWR: 0x800 16 byte data: 0
FPRD: 0x1400 Return value: 0x5208
FPWR: 0x620 16 byte data: 0
FPRD: 0x1400 Return value: 1024 byte data...0x80
...

Could this be the reason why the slave don't go to INIT when IgH master ask?

By the way I tried to write the right values to 0x800 via ecrt_reg_request_write. I tried to call this function right after ecrt_master_activate. But it didn't work.

You can find the log file at the attachments. I hope someone can give me a suggestion.

Best regards,

Oguz.

On 25-Feb-26 3:42 PM, Bilko AS, Oguz Dilmac wrote:

Hi Again,

I just sent some dmesg outputs and a test program using a new ethercat library. Unfortunately it hit a max 200KB attachment limit. Sorry I didn't notice this limit before. And could not cancel my message either.

Anyway, I uploaded the files to wetransfer. I also uploaded a wireshark log of twincat while the problem device goes to OP. Here is the link:

https://we.tl/t-Yl7iRMSadB

I also copy my previous message here:

Hello,

We went to customer side and used a test PC with ethercat library 1.6.7. We also used slave timeout function with 10 seconds.

We build a simple setup with only one device and one test PC. Still the old device is going OP, and the new device don't.

At the attachments you can find some dmesg outputs. I also attached the simple test program I used.

Some explanations for dmesg outputs:

* dmesg.OK.txt: Output for the working device. I just put it here as a reference

* dmesg.Problem.txt: Output of the problematic device.

* dmesg.Problem.freerun.txt: I tried to activate freerun with ecrt_slave_config_dc(sc, 0, 0, 0, 0, 0); I put it right before ecrt_master_activate().

* dmesg.Problem.ManualStateChange: I tried to change state via command line tool.

We also checked and see that both devices are going to OP with twincat. If you suggest a way to get some log from twincat, we can try.

We are stuck 🙁 If you have any idea It would be great.

Best regards,

Oguz.

On 20-Feb-26 1:30 PM, Richard Hacker wrote:

On Fri, 2026-02-20 at 12:49 +0300, Bilko AS, Oguz Dilmac wrote:

Hi,

We will try to install the newest ethercat library version.

By the way, we tried one version older of this device. It goes to OP

with no problem.

Well, this shows that your configuration is probably not the issue and

is at the problem is at the slave's side

Also our customer has a bechoff controller as well, and they say that

with twincat both devices are going to OP without a problem.

TwinCAT is not the standard, neither is our EtherCAT master. However,

both should implement the standard, but they do it differently

(naturally), especially regarding timing. The standard however ensures

that EtherCAT masters and slaves are immune to timing issues, but there

are always exceptions!

We will try to go to the customer and check with the newest ethercat

master version. We are also considering to connect the both devices

to

the twincat and see if the twincat behaves different. Do you think

checking the transmitted data via wireshark worth trying?

You can try, but I do not see that you'll get much insight using

wireshark. The master commands the slave to go to INIT and there is

simply no reply from the slave acknowledging the state change. That is

why the master times out and gives up trying to bring the slave online.

Going to INIT is such a primitive state change for a slave that there

should be absolutely _no_ reason it takes any amount of time. So I am

not even convinced that ecrt_slave_config_state_timeout() would work.

I only know that some (complex) slaves may take time between PREOP-

SAFEOP and SAFEOP->OP transition when certain hardware needs to be

configured based on its configuration. But your slave is not there yet.

Before you use wireshark you can try switching debug to 1

`ethercat debug 1`

--

Oguz Dilmac

Bilko AS, R&D Manager

====================================

Perpa Ticaret Merkezi B Blok Kat 13 Nr. 2536

TR-34384 Okmeydani Istanbul Turkey

Tel : +90 212 563 00 00

e-mail : odilmac at bilko-automation.com<mailto:odilmac at bilko-automation.com>

web site : http://www.bilko-automation.com

https://www.youtube.com/@LyncaCNC

--

Oguz Dilmac

Bilko AS, R&D Manager

====================================

Perpa Ticaret Merkezi B Blok Kat 13 Nr. 2536

TR-34384 Okmeydani Istanbul Turkey

Tel : +90 212 563 00 00

e-mail : odilmac at bilko-automation.com<mailto:odilmac at bilko-automation.com>

web site : http://www.bilko-automation.com

https://www.youtube.com/@LyncaCNC
-------------- next part --------------
An HTML attachment was scrubbed...
URL: <http://lists.etherlab.org/pipermail/etherlab-users/attachments/20260315/d4a83d20/attachment-0001.htm>