suntravel wrote: ↑Wed Oct 05, 2022 2:23 pm
You can run stresstest.cnc if you have updatet to 4.80
This will show you if the connection ist ok or not.
If your PC have an Intel Network chip, is needs some changes in the advanced setting to get rid of resends.
Uwe
Given that I have been experiencing same issues with two drastically different wiring configurations with two different setups, I am willing to bet that Intel NIC is actually the culprit. Do you by any chance know which settings should be modified? My NIC is
Intel® Ethernet Connection I219-V
Today I purchased USB Ethernet interface from TP-Link which I will test. I also plan on completely plugging out spindle VFD just to make sure that nothing outside of cabinet makes noise (except Clearpath servos which I obviously can't turn off).
Changing the NIC settings wont help if you have a noise issue. Leave the spindle off and test the system to see if you still have resends. Run several air cuts without the spindle and then look at your error log again. Please do one change at a time so you know what the actual issue was. USB to ethernet adaptors are also not recommended by Centroid. It'll just add another layer of trouble and they are not supported.
Did either of your "drastically different wiring configurations" use shielded wire direct to the motors?
ShawnM wrote: ↑Wed Oct 05, 2022 9:32 pm
Changing the NIC settings wont help if you have a noise issue. Leave the spindle off and test the system to see if you still have resends. Run several air cuts without the spindle and then look at your error log again. Please do one change at a time so you know what the actual issue was. USB to ethernet adaptors are also not recommended by Centroid. It'll just add another layer of trouble and they are not supported.
Did either of your "drastically different wiring configurations" use shielded wire direct to the motors?
I agree, idea is to shutdown everything not needed and retest. I will do that today.
Both setups used shielded wiring, even magnetic ferules on cable ends. One of the reasons I switched to Clearpath was getting their OEM wiring which seems to be good and moving drivers away from electronics (since they are part of servos and not housed within cabinet).
As with anything complex, doing one thing at a time is a must
Also on my PC with Intel chip I have to turn on Acorn first then the PC, or for some reason I will get resends.
Uwe
Thanks for this, I will do the Intel NUC maintenance today (OS upgrade, drivers refresh). My setup already has the same turn on order since as soon as I hit the master switch on the cabinet door Acorn turns on. I need to open door and manually turn on PC which is mounted on the other side of door.
bbatarelo wrote: ↑Thu Oct 06, 2022 7:05 am
I agree, idea is to shutdown everything not needed and retest. I will do that today.
Both setups used shielded wiring, even magnetic ferules on cable ends. One of the reasons I switched to Clearpath was getting their OEM wiring which seems to be good and moving drivers away from electronics (since they are part of servos and not housed within cabinet).
As with anything complex, doing one thing at a time is a must
Bruno
The Clearpath OEM wire is not shielded as you know, nor is it even a twisted pair. This is why I don't use their overpriced cables. I love the motors but not their cables. Do they work? I guess. I make all my own shielded motor cables for Clearpath motors. Shielded (and grounded) signal wires from the drive all the way to Acorn is good practice.
I also notice what appears to be two cables coiled up next the the IPC5. Is that the AC line voltage for the IPC5 and the DC voltage going to the power hub? If so try and separate them and cut the a/c cable to length, coiling a/c wires can lead to more noise.
It appears from your report that this noise issue has been there from the start so hopefully we can help you find the solution.
ShawnM wrote: ↑Thu Oct 06, 2022 10:20 am
The Clearpath OEM wire is not shielded as you know, nor is it even a twisted pair. This is why I don't use their overpriced cables. I love the motors but not their cables. Do they work? I guess. I make all my own shielded motor cables for Clearpath motors. Shielded (and grounded) signal wires from the drive all the way to Acorn is good practice.
I couldnt agree more. For simple signaling cable the CP/mini-junior cables suck. Why you need high current current cable for signaling is a head scratcher. I made some RJ45 to mini junior adaptor boards and it worked ok, but not great. You cant get mini-jr cable to pcb headers so I had to cobble in a cable connecter glued to the pcb. It works but isnt great. Running shielded CAT cable it so much nicer and you can color code the axis.
bbatarelo wrote: ↑Thu Oct 06, 2022 7:05 am
I agree, idea is to shutdown everything not needed and retest. I will do that today.
Both setups used shielded wiring, even magnetic ferules on cable ends. One of the reasons I switched to Clearpath was getting their OEM wiring which seems to be good and moving drivers away from electronics (since they are part of servos and not housed within cabinet).
As with anything complex, doing one thing at a time is a must
Bruno
The Clearpath OEM wire is not shielded as you know, nor is it even a twisted pair. This is why I don't use their overpriced cables. I love the motors but not their cables. Do they work? I guess. I make all my own shielded motor cables for Clearpath motors. Shielded (and grounded) signal wires from the drive all the way to Acorn is good practice.
I also notice what appears to be two cables coiled up next the the IPC5. Is that the AC line voltage for the IPC5 and the DC voltage going to the power hub? If so try and separate them and cut the a/c cable to length, coiling a/c wires can lead to more noise.
It appears from your report that this noise issue has been there from the start so hopefully we can help you find the solution.
I can cut ac line to length. DC will have to remain as is since I don’t have spare terminals to crimp.
As far as shielding is concerned, axis cables are unshielded from acorn to chassis. Later from XLR connector to drives it’s a shielded connection. I know it’s not ideal, but again, I don’t have spare terminals to crimp.
So, interesting news here - the only thing I needed to do was to upgrade to latest version of centroid. After that things magically started working without resends, stress test passed without issues. I took computer, brought it home, made OS upgrade which also updated all Intel drivers and retested everything today. Also all good - no resends no matter how many probing cycles I did. I will consider this thing fixed now. I will periodically check for errors, but things look promising
I'm attaching latest report just for the curious.
