When using a MC403-Z P821, we are facing with an issue related to the "direction" signal.
Please take a look at the following picture:
Scope screenWe have connected the oscilloscope on the step signal (channel 1) and direction signal (channel 2).
From what we see, it seems that the direction signal has a strange behavior... multiple commutations...
Please could you explain the behaviour?
=====================================================================
We have reproduced the same effect at Trio.
The theory we have is that the divide by 16 in the FPGA is not at a count which will generate a pulse. As the MOVECIRC crosses 0 position, the internal count from DPOS jitters slightly so it sends a count up by 1, then down by 1 and repeats this for a few milliseconds. The count is not enough to generate a pulse at the output, but it is enough to cause the direction output to change between forward and reverse. The change in the direction output does not matter to the drive because there is no pulse on the step output.
The reason the output changes immediately there is a direction change internally is because some drives need a minimum time between direction change and the pulse. The maths for the MOVECIRC create some very small errors as it crosses 0 and they are generating "noise" in the demanded position which then changes the direction output.