I recently upgraded from 3.16 Release to Master 4.21d, primarily to get the spiral search feature. At one point I believed I got it to work, at least intermitently but a bigger issue arose with the primary functioning of the controller (running STM32 Blue pill with TMC5160 drivers with DS3231 with Wifi and SHC on a kit circuit board). Before upgrading, I had been running 3.16 robustly for many months. Description of the 4.21d problem:
When the controller powers up, the RA axis starts rotating at a rate that is more like a fast quide rate as opposed to a tracking rate....maybe 60 times the correct tracking rate. This behavior happens consistently when 4.21d is loaded, seemingly independent of config.h setting changes. During this behavior, the DEC axis remains stopped at exactly 90 00 00. The RA axis rotation rate (observed both visually on the axis itself as well as on the SHC display) is about one angular minute per second of time (roughly). This behavior appears to continue indefinitely (I have let it run for many minutes and it just continues to go round). All the while, the controller's LED is flashing as if were tracking. What I have tried so far to resolve this:
* Using the EEPROM wipe tool to wipe the NV memory and then reload 4.21d
* Checked the config.h file many, many times to look for issues such as the incorrect steps per or microstepping
* One by one changed what would seem to be the potentially relevant config.h settings within reasonable limits to see if I could perturb this behavior...to no avail.
* Reverted the system back to 3.16 release. It works perfectly suggesting that this is not a HW issue. I have done this multiple times...back and forth from 3.16 to 4.21d...3.16 works, 4.21d does not.
* Pulled a fresh download of 4.21d from github and loaded. Same issue
* When powering up 4.21d (start tracking on boot set to on), the LED turn on for a few seconds (normal), then goes to flashing (to indicate tracking). The RA axis does not turn while the LED is steady ON (though I do feel a couple of "clicks" during the steady on period...but no real rotation. When the LED starts flashing, the RA axis starts rotating.
* If I ignore the high RA rotation rate and play with syncs and gotos, the controller acts strangely, sometimes to the point where it screws up the SHC (display goes off/acts wierd). This seems to suggest that whatever is going on, it may be screwing up the SW operation in general.
* I found and loaded 4.16f (just took a stab and an older release to see if it was different). It appears to be behaving normally; no high speed "tracking" behavior. I have limited test time on 4.16f but so far, it does not show the high speed rotation problem.