Help with another DEC pattern
Magnus Larsson
Here is anonther DEC pattern that is new to me, as of tonight. See the attached image and PHD log file. I'm using a guide scope with 500 mm focal length, and a Losmandy G-11 mount. The weather was nice, almost no wind. I was indoors asleep for most of the time, so I cannot say there were no wind gusts. But they would not cause these drifts, would they?
Best,
Magnus
Andy Galasso
Magnus Larsson
Thanks, Andy!
Yeah, I know. This is precisely the nice curve I am aimin for. I've been fiddling a lot with aggresiveness, backlash and min motion. This time, it was all set high - as an attempt to come to terms with even more movements. I guess I'm in a sense hunting seeing, but interestingly, I get far better results with 1 sec exposure and somewhat aggressive settings, than with say 2 sec exposures. But there are many factors in that.
For the moment, it seems that my Dec has become worse, and this odd saw.-like pattern is totally new to me. I am wondering if there is anything mechanically wrong (introduced recently). I've tried tightening all screws I can find (guide scope, az and elevation on mount, guide cam in guide scope etc). I'm wondering if the backlash might grow on me and play tricks - on the G11 this is a matter of adjusting the "blocks" where the worm gear runs (not sure about the english terms here). Or is this a matter of bad PA or strange balance (for some reason)?
What I try to do is focus on one of the two axes at the time and experiment to improve it. This time it is Dec.
Interestingly enough - I imaged tonight as well, and this saw-like pattern is not there at all. Better values with aggressiveness set at 40 and min motion at 0,15. But still far from what I'd like (the slow pattern).
So any ideas wecome - I'll keep experimenting!
Magnus
peter wolsley
Magnus Larsson
Here's a follow up on this (I hope I post it correctly - there are two different "reply" buttons...)
I've been checking all I can and cannot find anything that is loose or that could cause backlash. That is not to say it is not there. I also used the guiding assistant yesterday, output below.
So how bad is this backlash?
I've not seen the "saw" pattern since I posted it. What I've done is extensive polar alignment, and mabe that has helped.
However, yesterday, I had another experience that I cannot understand, concerning the DEC. Early in the evening, I imaged M33 and got the guiding results that I expect - what I used to get until about a week ago. At this point in time I had the mount to the west pointing east. See logfile called "first_log" and here's a screenhot.
I think this looks quite good, right, Andy?:) More similar to the one you posted, and actually what I used to get. Here I have the DEC aggressivness turned far down, and little backlash compesation.
Because of bad transparency, I stopped imaging and did other things. A few hours later, I resumed - this time with mount east pointing west (capable of running the whole night with out meridian flip). And M33 was close to the meridian. Now I get a very different pattern in DEC, and far worse guiding results. This is the situation where I up until about a week ago consistently had 0.7-0-8 RMS total. There was not much wind, still and nice, but quite damp outside, so seeing was not that good. Guide log attached, called "second_log". I paused and ran the guiding assistant - and got this pattern. THis surprised me. First: DEC does not drift consistently, but oscillates in a way I'd expect from RA but not from Dec. Secondly, DEC follows RA very very closely. What is going on here....?
I also noted that when turning guiding output totally off from DEC, I get the same oscillating pattern, but very little in terms of drift. I ended up finally turning aggressiveness somewhat up, guding set to auto, and little backlash compensation, and shorter exposures. Then there came clouds and I feel asleep...
Any ideas on this is warmly welcome. Obviously, my guiding has deteriorated vastly over the last few weeks. I have not changed anything but of course moved stuff in and out (do not have an observatory) although the mount itself is left where it is (sheltered of course).
Magnus
Magnus Larsson
Magnus
Andy Galasso
Magnus Larsson
Hi!
Thanks a lot for looking at it!
Well, I would not rule anything out. However, I changed the
position of the guidecam slightly on the afternoon yesterday, then
fastened it again (although I apparently forgot to secure the
focusing, that that cannot account for rotation). I just checked
it. It is tightly fastened. And no rotator. I didn't touch it
between the sessions.
However, given the MF between these, I wonder if there is
something with this. Cables becoming tight? Although east pointing
west is where cables have MORE slack. I like the idea of
invalidating calibration. How else could that be accomplished?
Could it be that PHD2 was not aware of my MF, thereby messing
things up? (I use Indi/Ekos for scope control and imaging).
Magnus
--
You received this message because you are subscribed to a topic in the Google Groups "Open PHD Guiding" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/open-phd-guiding/TeiM9J9S29I/unsubscribe.
To unsubscribe from this group and all its topics, send an email to open-phd-guidi...@googlegroups.com.
For more options, visit https://groups.google.com/d/optout.
Andy Galasso
However, given the MF between these, I wonder if there is something with this. Cables becoming tight? Although east pointing west is where cables have MORE slack. I like the idea of invalidating calibration. How else could that be accomplished? Could it be that PHD2 was not aware of my MF, thereby messing things up?
magnus
Andy Galasso
Magnus Larsson
Here are some updates from tonights session. I cannot repeat the oscillation in Dec that follows RA. However, there seems to be somthing going on when I do a meridian flip.
First, I have the scope west pointing east, pointing south, a little bit over the equator. This is the calibration data, the GA, and a bit of guiding graph. The graph is as I would expect, between 0.7 and 0.9 arc sec RMS total
Then I flipped the mount. Now east pointing west. No change in anything else. Here is the GA, GA graph, and a bit of guide data:
Now there is quite some difference. But stil no DEC oscillation following the RA. the polar aligment suddenly is far worse. And tracking in RA is far worse as well. So I did a new calibration, the data is here:
And after this, voila, a far better guide result:
I think I got it right. Log is attached, including some quite failed imaging in the end, due to wind, I guess. So my question for tonight (a new question every night, it seems...:)) - how can the polar alignment shift so much when doing a meridian flip? What is going on here?
Magnus
Andy Galasso
my question for tonight (a new question every night, it seems...:)) - how can the polar alignment shift so much when doing a meridian flip? What is going on here?
Magnus Larsson
Here's a kind of follow up - somewhat loosely coupled to the original topic, but I choose to post it here for consistency in regards to my situation.
I have now disassembled, cleaned and regreased the mount, and tried to adjust the worm gears with as much precision as I can. I've also worked on the polar aligment, so it is at this moment in time about 2 arc min off at most. I see no significant differences in GA output when doing the meridian flip anymore (pointing at the meridian), and I have not seen any more oscillation in DEC following the RA pattern. I've trained ("manually", that is not through PemPRO) and turned on PEC.
What happens now is that I get some DEC spikes. THey all seem to
1. go in the same direction. Direction shifts after meridian flip. However, I'm not sure how that is represented in the graph - is "guiding North" thus the opposite direction for the mount after meridian flip? If so, the "side" of pulses are consistent in terms of the mount.
2. be triggered by a pulse.
I take this to be the effect of backlash compensation. How do I understand this? At times backlash compensation works nicely, and then suddenly, there is way too much....? In the end, as seen in the log, I turned backlash comp off, and then changed to guide only north. And then no more spikes. But no other backlash problems either.
So, if it works nicely without backlash comp, how come PHD2 insists on a backlash compensation of up to more than 1000, when turned on?
Practically, I guess this indicates that I should turn it off. But I'd like to understand it as well :)
Interestingly, as seen by the two guiding sessions, guiding RMS was some 15 % worse after meridian flip (the chaotic end of first session was caused by reaching the limit, and I executed a meridian flip, repostioned and restarted guiding.
Any ideas warmly welcome!
Magnus
bw_msgboard
Hi Magnus, see below.
From:
open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Magnus Larsson
Sent: Monday, December 04, 2017
11:40 PM
To: Open PHD Guiding
Subject: Re: [open-phd-guiding]
Re: Help with another DEC pattern
Hi!
Here's a kind of follow up - somewhat loosely coupled to the original topic,
but I choose to post it here for consistency in regards to my situation.
I have now disassembled, cleaned and regreased the mount, and tried to adjust
the worm gears with as much precision as I can. I've also worked on the polar
aligment, so it is at this moment in time about 2 arc min off at most. I see no
significant differences in GA output when doing the meridian flip anymore
(pointing at the meridian), and I have not seen any more oscillation in DEC
following the RA pattern. I've trained ("manually", that is not
through PemPRO) and turned on PEC.
What happens now is that I get some DEC spikes. THey all seem to
1. go in the same direction. Direction shifts after meridian flip. However, I'm
not sure how that is represented in the graph - is "guiding North"
thus the opposite direction for the mount after meridian flip? If so, the
"side" of pulses are consistent in terms of
the mount.
‘Guide North’ and ‘Guide East’ describe the direction of the guide pulses that were sent to the mount. They don’t have any fixed relationship to sky locations or directions relative to the ground. Most mounts don’t change the direction the Dec motor will turn because of a meridian flip. So a ‘north’ command will often mean ‘rotate the Dec axis counter-clockwise as viewed from the top of the mount’ regardless of side-of-pier. I think the Losmandy behaves that way but some other mounts do not.
2. be triggered by a pulse.
I take this to be the effect of backlash compensation. How do I understand
this? At times backlash compensation works nicely, and then suddenly, there is
way too much....? In the end, as seen in the log, I turned backlash comp off,
and then changed to guide only north. And then no more spikes. But no other
backlash problems either.
So, if it works nicely without backlash comp, how come PHD2 insists on a
backlash compensation of up to more than 1000, when turned on?
Practically, I guess this indicates that I should turn it off. But I'd like to
understand it as well :)
First off, your overall guiding seems quite good – you should have had nicely round stars except for the few Dec excursions. Second, there’s pretty good evidence that the seeing degraded a bit after you did the meridian flip. The Dec spikes I see in the second session look like they were triggered by a too-small setting of min-move. You were getting more motion in Dec just because of seeing effects and some of these relatively small moves triggered guide pulses and some of those triggered backlash comp pulses. Those big corrections are going to be wrong because it was a random seeing event that caused the apparent move, not an error in how the mount was tracking. If you had increased the min-move value up higher, for example to 0.25, you would have largely avoided these issues. During the first session, the random movements due to seeing were smaller and didn’t trigger the apparent direction reversals. If you’re going to use an aggressive approach like backlash compensation, you’ll have to keep an eye on things and be sure you aren’t chasing seeing in Dec.
Personally, I think less is better with guiding. If you get consistently good results without backlash compensation, then leave it off. Unless the RA and Dec guiding numbers stay fairly close together, you’re not gaining much by improving one axis without also improving the other.
Hope this helps,
Bruce
Interestingly, as seen by the two guiding sessions, guiding RMS was some 15 %
worse after meridian flip (the chaotic end of first session was caused by
reaching the limit, and I executed a meridian flip,
repostioned and restarted guiding.
Any ideas warmly welcome!
Magnus
--
You received this message because you are subscribed to the Google Groups
"Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to open-phd-guidi...@googlegroups.com.
Magnus Larsson
Thanks, yes, it is very helpful. But ...:) if you have the time and energy, I still have a few questions - more aiming towards me understanding than improving guding, for the moment.
Guide North and Guide South: so, when the spikes change direction after a meridian flip - it means that they also have changed direction "on the mount"? That is, if the spikes earlier was in the clockwise direction, they are now in the counter-clockwise? I'm thinking: if they are in the direction, say, away from the side that is slightly heavier (if not perfectly balanced) that direction would change on the meridian flip, right? So that they STILL are in the same direction, seen from the perspective of balance?
Which would indicate that spikes depend on how the tooths are "aligned"...right? If not balanced, we would expect backlash only in one direction, right? Or am I messing it up?
Secondly, on the pulse that trigger the spikes, I follow you in that the seeing maybe deteriorated (although it coincided with my meridian flip...hrm). But: evenif I have a too low min motion (I get that), most pulses do not trigger a spike, but a few does. Should I understand that as spikes occuring when there for some reason is no backlash at that moment (for instance becuase of improper balance) - or what is happening here?
And perhaps most centrally:
If it now works nicely without backlash compensation - how come PHD2 insists on setting a rather high value, when I have it turned on? Should it not instead gradually become lower and lower....?
I'm thinking that next clear night, I'll use the Guiding Assistant many times, to get a precise handle of the measured backlash at different times. Typically it is reported by GA to be some 40-70 px. Anything else you would advice me to do to understand this better?
Magnus
Brian Valente
I have found it automatically adjusts the amount more and less depending on what it sees. The amount of backlash in your mount does not go down, so PHD does not gradually remove it, instead it adjusts to what is required.
Thanks
Brian
Brian Valente
Brianvalentephotography.com
bw_msgboard
See below.
From:
open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Magnus Larsson
Sent: Tuesday, December 05, 2017
10:20 AM
To: Open PHD Guiding
Subject: Re: [open-phd-guiding]
Re: Help with another DEC pattern
Hi Bruce,
Thanks, yes, it is very helpful. But ...:) if you have the time and energy, I
still have a few questions - more aiming towards me understanding than
improving guding, for the moment.
Guide North and Guide South: so, when the spikes change direction after a
meridian flip - it means that they also have changed direction "on the
mount"? That is, if the spikes earlier was in the clockwise direction,
they are now in the counter-clockwise? I'm thinking: if they are in the
direction, say, away from the side that is slightly heavier (if not perfectly
balanced) that direction would change on the meridian flip, right? So that they
STILL are in the same direction, seen from the perspective of balance?
I think we’re having trouble communicating here – it’s hard to describe in words. I think it would be easier for you to just play with the telescope/mount and see for yourself how it works. Set up on the west side of the pier and push the ‘North’ button on the hand-controller – see how the Dec axis moves. Let’s say it moves counter-clockwise when you’re looking down the Dec shaft toward the counterweights. Now move the scope to the east side of the pier and again push the ‘North’ button. Does the axis continue to move counter-clockwise? With the Losmandy mount, I think it will stay the same regardless of side-of-pier. So when PHD2 issues a ‘north’ guide pulse, that’s the direction the Dec axis will move. Depending on where in the sky you’re pointing will determine whether the nose of the scope moves closer to the ground or further away.
Which would indicate that spikes depend on how the tooths are
"aligned"...right? If not balanced, we would expect backlash only in
one direction, right? Or am I messing it up?
Simple backlash is caused by looseness in the mesh of the drive gears. Some amount of looseness is required so the gears don’t bind and the axis is allowed to reverse direction at all. That means there will be a very small delay in getting the gears re-meshed in the opposite direction when the direction of movement switches. If the delay is on the order of a few hundred milliseconds it in unlikely to have a noticeable effect on guiding. Once it gets up above 1 second it can become more of an issue. Some mounts can take up to 10-15 seconds. If simple, small backlash is the only imperfection, PHD2 backlash compensation can work pretty well. But there can be other complications. For example, the amount of backlash might be different when going from north to south vs south to north – perhaps because of a serious imbalance. In that case, the “correct” amount of compensation will keep changing depending on the direction of reversal. Other times, there can be some form of stiction or resistance in the system that may depend on the gravitational load applied to the gears. That is much more difficult because it can create over-shoots and dependence on the scope’s pointing position. These kinds of problems probably won’t be handled well by backlash compensation or at least not always handled well.
Secondly, on the pulse that trigger the spikes, I follow you in that the seeing
maybe deteriorated (although it coincided with my meridian
flip...hrm). But: evenif I have a too low min motion (I get that), most pulses
do not trigger a spike, but a few does. Should I understand that as spikes
occuring when there for some reason is no backlash at that moment (for instance
becuase of improper balance) - or what is happening here?
The simple premise behind PHD2’s backlash compensation is to issue pulses that will keep the Dec gear system fully meshed in one direction or the other. But guiding is a statistical process, hampered by all sorts of random behaviors – both mechanical and seeing-induced. If the gear train gets into a state where it is “floating” – not engaged in either direction – the next backlash compensation pulse will probably create an over-shoot.
And perhaps most centrally:
If it now works nicely without backlash compensation - how come PHD2 insists on
setting a rather high value, when I have it turned on? Should it not instead
gradually become lower and lower....?
The backlash compensation is automatically adjusted upward or downward depending on the recent history. Maybe your mount is showing some kind of asymmetry in how the backlash behaves so that there isn’t convergence on a single best estimate for compensation. Since you’re interested in all the details, here’s what I suggest. Open the debug log file associated with your guiding sessions and do a text-search on BLC. Each entry will show you exactly what was done with backlash compensation and how the amount was adjusted (if it needed adjustment).
Bruce
Magnus Larsson
Hi!
Brilliant!! Thanks a lot!!
Magnus