Seth, one of the few people with a Replicape (Rev A3A) on the way has asked a lot of questions about the inner workings of the Redeem/PRU software. Today I recorded some G1-motions. Here are the results. A Step happens on rising flank.
what speed? What are the two signal pins you are looking at (x / y step pins?) One thing we wanted to see is how the steps are generated on fast moves. The arduino stuff is doing double and quad stepping in the interrupt and does not produce smooth stepping. I don’t have a image handy but it would be a bunch of quick back to back pulses then a gap and repeats.
Yes, x and y steps.
Did you ever get around to sampling some higher speed steps? We are looking to compare to a know problem with the arduino / Atmel based products.
No, I’m right in the middle of merging with the work of Methieu: https://bitbucket.org/zittix/redeem/overview
He has a higher level in PRU implementation of the step procedure, so the duty cycles are not symmetrical like in these pictures. I’ll post something once I have it working!
Ok. That is what we are looking for. Most firmware isn’t symmetrical once it gets past a certain speed because they do 4 steps in a single interrupt because it is too fast for each step to be clocked on its own timer.