Parport to 7i92

Channel 1 of the oscilloscope on 485+
Channel 2 of the oscilloscope on 485-
and Ground of the oscilloscope on Gnd.
Gives IMG_20220611_121629.jpg

Is that what you mean ?

If this is the device you have, it appears to have a termination resistor built in. so you might only need it at the drive end.
See R13.
Try that and see if the overshoot on A is reduced.

OK I will.

One question about this module : I connected Gnd and Vcc to an Usb port of the PC.
But with my 1.5kw it's working without the Usb plugged. Looking at the schematics, I don't understant how this is possible.

Today I added a 120 ohm resistor only on the VFD side. The Rs485 converter is powered with 5V.
With SpindleTalker2 on a windows laptop, I can talk to the VFD .
But with my LinuxCnc, it is still a failure.
I unpluged again the 2.2Kw and connect the old 1.5Kw, and LinuxCnc comm is OK.
I guess I have to get my oscilloscope to compare SpindleTalker2 and LinuxCnc  signals ! Any idea on what should I look at ?

1.5kw and 2.2kw, are you refering to spindle or VFD.

To the VFD.
I have spindle for both. At the moment only the 2.2Kw spindle is connected to  the 2.2Kw VFD.
The 1.5Kw spindle is stored

May be serial Com port is faulty. (even if I could talk to the 1.5Kw VFD )
So I took an other desktop computer, fresh install LinuxCnc 2.8.2
With the same Rs485 adaptor, same wiring, 2.2Kw VFD, I have the same result as with dedicated linuxCnc box : no comms

Maybe it is worth trying the HY_GT_VFD version of HY_VFD?

I tried thé hy_gt_vfd but also comms error.
Do you know à mobus test application for linuxcnc ?

You could try mb2hal
linuxcnc.org/docs/stable/html/drivers/mb2hal.html
But my understanding is that HY VFDs don't talk real ModBus, but rather their own modbus-like protocol.