- Posts: 3
- Thank you received: 0
Solved PHD guiding problems with M-UNO
21 Jan 2021 23:22 #470
by ken.self
-- Ken
Avalon M-Uno; GSO RC8; ASI1600; Optec focuser; Aaeon UP/Ubuntu/INDI
Replied by ken.self on topic PHD guiding problems with M-UNO
As one of the folks at Cloudy Nights trying to help out Peter we went through his guide logs to see if we could understand what was happening.
In the log he attached, run #3 is a 53 minute unguided run using the guiding assistant. That looks totally normal for an Avalon although the RA drift looks larger than usual. This could be due to an imperfect calibration.
Run #4 is a calibration and is where things begin to look strange. Peter appears to have set guide rates in Stargo of 0.35x in RA and 0.4x in Declination.
The attached graph shows the raw calibration movements in pixels.
The west leg shows erratic guide pulse movements. Short ones in between longer ones. However it is straight.
The east leg has evenly spaced guide movements and returns along the same path but stops short of the start position. The pulses on this leg are longer due to the fast return option being used. 2500ms pulses vs 1150ms on the west leg.
There is a bit of backlash correction but that appears to be an artefact of the dec movement.
The north leg shows the same erratic pulse lengths as the west leg. I think PHD2 interprets the short movement as backlash. The north leg is orthogonal to the east/west leg
The south leg is similar to the east leg and has evenly spaced pulses - also longer than the north pulses: 2500ms vs 1150
After calibration the guide rates based on the calibration results come out at 6.9 as/s RA (0.46x) and 8.1 as/s in DEC (0.54x). This can be due to an incorrect focal elngth specified but it it also affected by the unevenly spaced pulses west and north.
Overall the calibration looks reasonably good.
The issues start when guiding commences e.g run #5. RA guiding looks good. But dec starts out with a small 5s oscillation that finally settles to a large (10 arcse peak to peak) oscillation ith 10s period. This is using the LowPass2 algorithm with 90% aggression.
In an earlier session, Peter was using the Z-filter algorithm and got a similar Dec oscillation with a period of 20s. I think this is because the Z-filter was less aggressive with its corrections. But that indicates that it is not a guiding algorithm issue. And in any case there are many Avalon users using both LowPass2 and Z-filter algorithms successfully.
Looking in the Debug log, whenever a guide pulse is issued, PHD2 asks for a IsGuiding status and most of the time a 0 (not guiding) is returned even before the pulse should have finished. It does occasionally return IsGuiding = 1 (guiding in progress) but that is often when the pulse should have finished.
This is definitely a strange problem. As best as we can figure out, it may be due to the guide commands being delayed along the communications chain and being applied out of phase. But it only seems to affect the Dec axis. It could also be electromechanical.
In the log he attached, run #3 is a 53 minute unguided run using the guiding assistant. That looks totally normal for an Avalon although the RA drift looks larger than usual. This could be due to an imperfect calibration.
Run #4 is a calibration and is where things begin to look strange. Peter appears to have set guide rates in Stargo of 0.35x in RA and 0.4x in Declination.
The attached graph shows the raw calibration movements in pixels.
The west leg shows erratic guide pulse movements. Short ones in between longer ones. However it is straight.
The east leg has evenly spaced guide movements and returns along the same path but stops short of the start position. The pulses on this leg are longer due to the fast return option being used. 2500ms pulses vs 1150ms on the west leg.
There is a bit of backlash correction but that appears to be an artefact of the dec movement.
The north leg shows the same erratic pulse lengths as the west leg. I think PHD2 interprets the short movement as backlash. The north leg is orthogonal to the east/west leg
The south leg is similar to the east leg and has evenly spaced pulses - also longer than the north pulses: 2500ms vs 1150
After calibration the guide rates based on the calibration results come out at 6.9 as/s RA (0.46x) and 8.1 as/s in DEC (0.54x). This can be due to an incorrect focal elngth specified but it it also affected by the unevenly spaced pulses west and north.
Overall the calibration looks reasonably good.
The issues start when guiding commences e.g run #5. RA guiding looks good. But dec starts out with a small 5s oscillation that finally settles to a large (10 arcse peak to peak) oscillation ith 10s period. This is using the LowPass2 algorithm with 90% aggression.
In an earlier session, Peter was using the Z-filter algorithm and got a similar Dec oscillation with a period of 20s. I think this is because the Z-filter was less aggressive with its corrections. But that indicates that it is not a guiding algorithm issue. And in any case there are many Avalon users using both LowPass2 and Z-filter algorithms successfully.
Looking in the Debug log, whenever a guide pulse is issued, PHD2 asks for a IsGuiding status and most of the time a 0 (not guiding) is returned even before the pulse should have finished. It does occasionally return IsGuiding = 1 (guiding in progress) but that is often when the pulse should have finished.
This is definitely a strange problem. As best as we can figure out, it may be due to the guide commands being delayed along the communications chain and being applied out of phase. But it only seems to affect the Dec axis. It could also be electromechanical.
-- Ken
Avalon M-Uno; GSO RC8; ASI1600; Optec focuser; Aaeon UP/Ubuntu/INDI
Attachments:
Please Log in or Create an account to join the conversation.