Random ticking sound using elmo gold drives.

MankaranSingh

This issue has been randomly popping up now. When we tried motorcortex very first time with elmo with 2 joints, ticking sound was there. Then one day it randomly went away. It came back today again. We have double checked our grounding and electrical connections and the issue seems to be very random. We tried setting the distributed clock which resulted in even more ticking sounds. We are using latest elmo firmware.

Philippe

Hi Mankaran,
would you share your configuration (master.xml)? Please post the one where you configured DC clock. We will look if we can see something obvious.

We did have a lot of EMC issues with Elmo drives in the past, so as you guessed, this may also be EMC related. You can also check if EtherCAT packets get lost now and then by monitoring dmesg.

MankaranSingh

Here is the config. https://filebin.net/rfg4wlidjkjumwf4 . The DC clock setting was same as the one you linked earlier. The drives show no sound when running with elmo application studio or SOEM code.

I am not sure if this helps: but for 2 joints it was working fine, on connecting third joint, the ticking sound started happening, then we removed the third joint and now the ticking sound is still there. The cpu usage is around 30%

Also, how do I monitor if ethercat packets get lost ?

Philippe

MankaranSingh In you config there are no DClocks enabled

Philippe

Is it only occurring when multiple slaves are configured?

If you log in to the controller with ssh and then type "dmesg -w" you can watch the output. You will see "Dataframes Lost" or something similar when there communiction issues. Maybe you can correlate this to the sounds?

MankaranSingh

Yes, configuring dc clock causes more infrequent ticking sounds. its been working for a month without dc clock now.
This sometimes start happening with 2 slaves, but for 3 or more, its always happening.

When the sound is not there, enabling DC clock brings back the sound and its very infrequent.

I configure DC clocks as shown here https://support.motorcortex.io/assets/files/2023-02-03/1675418770-91875-screenshot-from-2023-02-03-11-03-26.png

Roel

Hi Mankaran,

To get a better understanding and investigate the cause, we'd like to build a similar setup with an Elmo drive and Motorcortex and replicate the issue. I will perform tests on the setup next week and keep you updated on this.

To start the investigation; can you provide some information and elaborate on the ticking noise:

Can you give me the specific firmware version number of the Elmo drive?
Can you send me the elmo's configuration file?
Can you send me a picture of your setup?
What is the version of Motorcortex image: 2022.09?
On what Hardware are you running Motorcortex? (type of CPU, memory)
At what sample time are you communicating? (1 kHz = default)
Does it changes in magnitude (strong/soft/away), does it fade in/out?
Infrequent in time and/or motor position?
Using Distributed Clocks it is always there (as you mentioned)
(btw, there is no offset in the DC settings in the figure, try set this at ₀.2 ms)
you checked on EMC?
are you using position mode (use Setpoint generation on the drive) or cyclic position with SOEM? (mode of operation)

Philippe

To double-check that it is not software related can you plot the setpoint that is going to the drive in our DESK tool:

root/Ethercat/Robot/Axis1/In/Target position

Make sure you set the Frequency Divider to 1, so you see all samples
In DESK, click on Settings and then set the Frequency Divider to 1::

The signal should be smooth, also when you hear the ticking sounds.

Please post your findings here.

MankaranSingh

Thanks for guidelines, will try to gather data.