|
I'm seeing about a 4" per minute (varies) linear drift in RA on my G11 that I'd like to address if possible. ?I've scoured all the info I could from here and the web first, but would love some suggestions on where to go.
Gear: - G11 Gemini 2 (latest firmware) - brand new brass worm ? - ABEC7 bearings in 2-block configuration - Belleville washer - new-ish clutch pads - axes balanced (done with worm removed during replacement) - recently cleaned and re-greased whole mount (SuperLube)
Data acquisition: - ST80 guide scope (WO fixed mounting rings on a Losmandy dovetail) piggybacked on a C8 - ASI 120mm? - PHD2 latest dev - acquiring at meridian on equator for ~ 8 minutes - polar alignment <1'
Things I've tested so far and ruled out:
- correct RA divisor (sidereal) - swapping cables - dew controller off - east bias (wouldn't expect this to do anything for a drift)
Probably the most interesting observation is that when I acquire with the scope on the east side, the mount appears to track fast (star drifts east in PHD), but when I flip to the scope being on the west side of the mount, the RA drift in PHD is in the opposite direction. ?So is this clutch slippage, flexure, balance? ?I could test the clutches slipping I suppose by "tracking" a fixed object in the day and looking for any drift. ?Other things I could think of to test are swapping motors, worm mesh (but this shouldn't matter here), but would like to hear your thoughts before moving on.
Thanks, Keith
|
Hi Keith ? Is it possible to see the guide logs from PHD2 when it¡¯s unguided? ? You may also try posting the guidelogs and questions to PHD2 user groups for tips as well ? Thanks ? Brian ? ? Brian Valente Brianvalentephotography.com ?
toggle quoted message
Show quoted text
From: Losmandy_users@... [mailto:Losmandy_users@...]
Sent: Monday, August 14, 2017 9:44 AM
To: Losmandy_users@...
Subject: [Losmandy_users] RA rate drift (revisited)? ? I'm seeing about a 4" per minute (varies) linear drift in RA on my G11 that I'd like to address if possible. ?I've scoured all the info I could from here and the web first, but would love some suggestions on where to go. ??????????? - G11 Gemini 2 (latest firmware) ??????????? - brand new brass worm ??????????? - ABEC7 bearings in 2-block configuration ??????????? - Belleville washer ??????????? - new-ish clutch pads ??????????? - axes balanced (done with worm removed during replacement) ??????????? - recently cleaned and re-greased whole mount (SuperLube) ??????????? - ST80 guide scope (WO fixed mounting rings on a Losmandy dovetail) piggybacked on a C8 ??????????? - PHD2 latest dev ??????????? - acquiring at meridian on equator for ~ 8 minutes ??????????? - polar alignment <1' Things I've tested so far and ruled out: ??????????? - correct RA divisor (sidereal) ??????????? - swapping cables ??????????? - dew controller off ??????????? - east bias (wouldn't expect this to do anything for a drift) Probably the most interesting observation is that when I acquire with the scope on the east side, the mount appears to track fast (star drifts east in PHD), but when I flip to the scope being on the west side of the mount, the RA drift in PHD is in the opposite direction. ?So is this clutch slippage, flexure, balance? ?I could test the clutches slipping I suppose by "tracking" a fixed object in the day and looking for any drift. ?Other things I could think of to test are swapping motors, worm mesh (but this shouldn't matter here), but would like to hear your thoughts before moving on. ?
|
Hi Keith,
Very nice summary of the system...?
My guesses:
1. Not clutch slippage, not anything wrong with the drive system. ?
2. My guesses are: ? ? a. slight error in polar alignment. ?(most likely cause).? ? ? ? ? (Did you do a long drift test for your polar alignment... as long as your ?sub exposure time? ? I use the PHD2 Drift Alignment tool and this helps very much. )
Reason: ?I am not sure that your polar alignment which you wrote as " < 1 ' " is good enough, for the 8 inch SCT (f/10?)?has a high angular mag, and the 80mm guide scope (f/6?) is probably wide view with low angular mag.? So a single pixel shift on the guide scope, to cause a PHD error correction, already would result in many pixels shift on the main scope image. ? ?Some folks put a Barlow on the guide scope to improve its error detection. ?
Other possible effects:? ? ? b. you may have a difference between the center star in your guide scope and the center star of your main telescope image frame. ? This will cause the autoguider to perfectly track the guide star, but the main image will seem to rotate. ?
and/or
? ? c. you have some differential flexure between your Guide Scope to SCT coupling.? These are awfully hard to eliminate due to changes in the gravity pull on the guide scope as the mount tracks. ?
and/or
? ? [d. you may have some mirror flop in your SCT. ?(SCTs are prone to this due to the way the primary mirror is "floating" by a gap around its moveable shaft, even if the shaft is "locked" in place.) ?A flop would give a quantum jump in image though...not a gradual one. ]
My experience was this:
I never was able to get perfect tracking/autoguiding with an external piggyback 80mm guide scope on my 10 inch f/10 SCT, even when assuring they were pointing to the exact same center image.? The reason (for me at the time) was that my polar alignment was imperfect, and the guide scope is guiding the mount to rotate about the guide scope guide star.? Any slight polar alignment error will then make the Main scope image look like it is drifting.
Recommendation: I gave up on piggyback guiding, and went to an OAG.? For me, the large prism and solid construction of the Lumicon giant easy guider was superior to other ones that have many adjustment screws and a small prism.? The OAG method this eliminated the problems of autoguiding... at least for me. ? But the polar alignment drift is a necessary step... when I try to skip that step I may the price... it is easy to knock the mount slightly, and mess the alignment up. ? ?
Hope these comments help you get to the finish line!
Michael
On Mon, Aug 14, 2017 at 9:43 AM, keithnakonechny@... [Losmandy_users] <Losmandy_users@...> wrote:
?
I'm seeing about a 4" per minute (varies) linear drift in RA on my G11 that I'd like to address if possible.? I've scoured all the info I could from here and the web first, but would love some suggestions on where to go.
Gear: - G11 Gemini 2 (latest firmware) - brand new brass worm ? - ABEC7 bearings in 2-block configuration - Belleville washer - new-ish clutch pads - axes balanced (done with worm removed during replacement) - recently cleaned and re-greased whole mount (SuperLube)
Data acquisition: - ST80 guide scope (WO fixed mounting rings on a Losmandy dovetail) piggybacked on a C8 - ASI 120mm? - PHD2 latest dev - acquiring at meridian on equator for ~ 8 minutes - polar alignment <1'
Things I've tested so far and ruled out:
- correct RA divisor (sidereal) - swapping cables - dew controller off - east bias (wouldn't expect this to do anything for a drift)
Probably the most interesting observation is that when I acquire with the scope on the east side, the mount appears to track fast (star drifts east in PHD), but when I flip to the scope being on the west side of the mount, the RA drift in PHD is in the opposite direction.? So is this clutch slippage, flexure, balance?? I could test the clutches slipping I suppose by "tracking" a fixed object in the day and looking for any drift.? Other things I could think of to test are swapping motors, worm mesh (but this shouldn't matter here), but would like to hear your thoughts before moving on.
Thanks, Keith
|
Hi Brian. ?I've posted a couple screenshots and the actual guide log in the Files section:
|
Hi Michael. ?I started the session with a PHD2 drift align. ?According the guide log and pics I have now posted (with tracking off), the DEC drift (meridian, equator) was approximately 1' over 8 minutes. ?The ST80 is f/5 (400mm focal length).
Right now the C8 is being completely factored out and is really just supporting the ST80 acquiring the data. ?I know there are differential flexure issues when actually imaging, but I really only do photometry where the subs are short enough that I'm not too worried. ?I worry more about finding guide stars,accurate flat fields, etc.
Really since my subs are short, the linear drift is not a huge concern since guiding will largely take care of it. ?But fundamentally a mount should do only one thing well, and that is track at the rate (sidereal) accurately. ?Of course PE etc etc comes in as well, but I'd like to nail down the root cause of this drift if possible.
I could look at fine tuning the PA but not sure how good I can get it. ?PA is a whole other issue (for example, why did I have very little DEC drift/PA error with the scope on the east side, then swing it around to the west side to track the same spot in the sky, then now have DEC drift indicating PA error?). ?Seems to me the PA routine in PHD is sensitive to the calibration, but testing this is on the to-do list (e.g. calibrate, measure PA, calibrate, measure PA,...).
Keith
|
Keith,
?
I think there are so many factors affecting the possibility of having *perfect* tracking from a mount, that you should cease striving for perfection and get on with doing photometry. As some
have said, ¡°perfect¡± is the enemy of ¡°really good¡±.
?
Even if you could achieve perfect polar alignment, (and I think your assessment of being within 1 arc minute of polar alignment is probably too good to be true), that would not be the best alignment
for tracking. You would have to take into account refraction and the correct polar alignment for particular needs is not perfect polar alignment in any case. Latitude changes the departure from perfect alignment for good tracking.
?
Copernicus used to think that celestial objects should orbit in perfect circles at uniform speed. It was a conviction he drew from ancient Greek philosophy. Johannes Kepler put him right on this
one, saying that physics applied in the celestial sphere as well. Mathematical perfection exists as only an idea in the human mind. Reality is a bit more messy. By the time you figure in error margins for every measurement, there is no point seeking improvements
in one or the other. Let¡¯s imagine that you could spend years perfecting your mount, only to have mirror shift in your C8 make it pointless.
?
?
?
?
?
?
toggle quoted message
Show quoted text
From: Losmandy_users@... [mailto:Losmandy_users@...]
Sent: Tuesday, 15 August 2017 3:35 AM
To: Losmandy_users@...
Subject: [Losmandy_users] Re: RA rate drift (revisited)
?
?
Hi Michael. ?I started the session with a PHD2 drift align. ?According the guide log and pics I have now posted (with tracking off), the DEC drift (meridian, equator) was approximately 1' over 8 minutes. ?The ST80 is f/5 (400mm focal length).
Right now the C8 is being completely factored out and is really just supporting the ST80 acquiring the data. ?I know there are differential flexure issues when actually imaging, but I really only do photometry where the subs are short enough
that I'm not too worried. ?I worry more about finding guide stars,accurate flat fields, etc.
Really since my subs are short, the linear drift is not a huge concern since guiding will largely take care of it. ?But fundamentally a mount should do only one thing well, and that is track at the rate (sidereal) accurately. ?Of course
PE etc etc comes in as well, but I'd like to nail down the root cause of this drift if possible.
I could look at fine tuning the PA but not sure how good I can get it. ?PA is a whole other issue (for example, why did I have very little DEC drift/PA error with the scope on the east side, then swing it around to the west side to track
the same spot in the sky, then now have DEC drift indicating PA error?). ?Seems to me the PA routine in PHD is sensitive to the calibration, but testing this is on the to-do list (e.g. calibrate, measure PA, calibrate, measure PA,...).
|
Thanks Greg. ?All very good points, and I have no intention of killing myself on this one. ?I guess its a bit of a statistical process control thing, i.e. for other G11/Gemini owners repeating the same experiment, how many would the same behaviour (10%, 50%, 90%)? ?That's what I have no idea about, and if I'm the outlier it would be nice to know why (or at least what I'm doing wrong in running the experiment).
Keith
|
The 4¡±/minute drift rate is kind of close to the difference between sidereal and solar rates. ?That could explain being slow, but wouldn¡¯t explain being fast on the other side.?
toggle quoted message
Show quoted text
Thanks Greg. ?All very good points, and I have no intention of killing myself on this one. ?I guess its a bit of a statistical process control thing, i.e. for other G11/Gemini owners repeating the same experiment, how many would the same behaviour (10%, 50%, 90%)? ?That's what I have no idea about, and if I'm the outlier it would be nice to know why (or at least what I'm doing wrong in running the experiment).
Keith
|
All great comments... a great group.
Another thought: ?Could the problem of a difference in tracking from East - West be just be uncorrected Cone Error?
See page 21 of this PDF manual (for the Orion Atlas but same issue for all GEMs):
I am wondering about that as you are saying that your 80mm guide scope is shifting image differently from East to (a meridian flip) to the West.? Maybe your scope (guide scope) is not "Cone Error" adjusted on the dovetail perfectly (what is perfect anyway)?? To eliminate that "Model parameter" down to zero, you have to get that right (and I never did that either... more work to do!).? So you might check on that. ?
[Also... did you build a complete sky model before this?? I think the Gemini1 (and 2) require a complete T pointing model that has the cone error as a parameter, and so should somewhat correct for East - West meridian flip position, though I can't imagine it also adjusts the tracking rate...? I'll stick with a cone error issue as my guess! ] ------------
It is the pursuit of perfection that gets people (not us of course!) Nobel prizes... finding out the reason for experiment not matching theory...?
There are many effects that affect our images, and our measurements.
I found out the hard way about atmospheric refraction... easy to see that if you do planetary imaging with a one-shot color camera and the R, G, and B are not overlapping exactly from the planet edge. ?
I also started to get into spectroscopy... trying to keep a star at the input of an optical fiber was quite difficult. ? So I made up a beam-splitter system (maybe like the newer form of "on-axis guider") .? I am not sure if you are doing the same... it is possible since you are not really going to split that starlight into colors, but instead just quantify the starlight intensity.? Anyway... I still use the beamsplitter side reflection to provide an image fo an autoguider camera.? Like I said... I never got the piggyback system to work. ?
(Also ... privately send me some info on the stellar measurements you are doing!? It must be far easier than the spectroscopy work!)
All the best, Michael
On Mon, Aug 14, 2017 at 11:56 AM, Les Niles les@... [Losmandy_users] <Losmandy_users@...> wrote:
?
The 4¡±/minute drift rate is kind of close to the difference between sidereal and solar rates.? That could explain being slow, but wouldn¡¯t explain being fast on the other side.?
Thanks Greg.? All very good points, and I have no intention of killing myself on this one.? I guess its a bit of a statistical process control thing, i.e. for other G11/Gemini owners repeating the same experiment, how many would the same behaviour (10%, 50%, 90%)?? That's what I have no idea about, and if I'm the outlier it would be nice to know why (or at least what I'm doing wrong in running the experiment).
Keith
|
Hi All
Just so you know, Ren¨¦ the writer of the firmware in the Gemini-1
and Gemini-2 does not
normally monitor this group.? He says he just does not the time to
do so.
He does try and monitor the , the and the
group.? Now I am not trying to get anyone to leave this
group,
but if the subject is specifically about the Gemini-2 and if you
need his help, you might try dual posting it this group and one of
the Gemini groups.
Tom Hilton
|
Good suggestion Tom. To be honest, I'm sometimes not sure which is the appropriate group to post to since as a customer, Losmandy and Gemini are interchangeable.. I'd like to test a few more things, then will repost in Gemini II.
Keith
|
Keith, ? I would suggest, if you haven¡¯t already done it, that you mount the guide scope directly onto the G-11 rigidly, not with the C8 involved at all. Simplify the rig as? much as possible and then make sure the focuser on the ST80 is really torque down as well.? I found that having a fitting threaded on with T-threads provided a much tighter coupling for the extender that I have to use to mount and focus my guiders on my ST80. A simple 1.25 tube extender would twist. So I use T-extension tubes threaded right onto the back of the focuser and then either thread a Meade DSI directly onto the other end of the tubes, or use a Baader T-threaded nose piece threaded onto the T-tube extensions with my QHY5. And then really lock the focuser down. ? If after doing all this (simplify the rig and verify the rigidity of the focuser) ?the problem is still there then back to square 1: Motors, software, electronics, or drive train. ? 4 arc seconds in 1 minute translates to 4 parts in 900. That is a lot and it is unlikely that the electronics are at fault (the crystals are good to parts per million or so). If the problem still persists after you simplify the rig and make sure the ST80 focuser and camera couplings are tight, then the only thing I can think of off the top of my head is a problem with the encoder on the motor (assuming all the timer parameters in the software are correct). That would tend to make the motor run fast. I am not familiar with the encoders on the Gemini motors but 256 lines ( 4 parts in 1024, nearly the same as your error) are common for encoder wheels, so if one line is messed up the RA motor would run faster. That may be a long shot. If you¡¯ve swapped the motors completely, including encoders, and the problem is still there then you can probably ignore this suggestion. ? Hopefully, you¡¯ll find it isn¡¯t an intrinsic problem with the mount but rather with the configuration of the rig (total weight, flex, cable drag, you name it) or the focuser on the ST80. Then start working forward adding in elements of your final configuration until the sinner is exposed. ? Best wishes, ? Mark Christensen
|
Keith
I added a third dovetail and adjustable ring that holds the extension tube tightly. ?I could have put it on the focuser directly, but went with the extension tube to make the third ring as close to the end of the contraption as possible, i.e. limited wiggle.
See image in?
Although I rarely have to adjust focus, its easy to do by loosening just one of the three bolts on the third ring.
Bryan
|
Keith
None of the user groups are run by Losmandy.? Now Ren¨¦ is a co-owner
of the Gemini-2 users group and the Gemini users group.
Tom H
toggle quoted message
Show quoted text
?
Good suggestion Tom. To be honest, I'm sometimes not sure
which is the appropriate group to post to since as a
customer, Losmandy and Gemini are interchangeable.. I'd
like to test a few more things, then will repost in Gemini
II.
Keith
|
Thanks for the suggestions Mark. ?All my fittings on the ST80 to the camera are screw-on as well, but to piggyback I did get the Losmandy-style dovetail instead of Vixen specifically for that reason (to attach directly to the mount), so I will give that a try.
Regarding the encoders, that's a good point. ?During my most recent attempt at installing the McLennan gearboxes, then subsequently going back to put the original pinion back on after not-so-good results, I had to disassemble the motor including the encoder cover in order to support the back of the motor shaft to press fit the pinion. ?This exposed the encoder disk, and while I was very careful not to disturb the disk, I can't rule it out. ?I did this to both motors, so I'll start by swapping the motors and test, and if the problem is still there, then its probably not the encoder (though I suppose its possible I damaged both encoders!). ?
Keith
|
Hi Tom. ?Yes I know these groups have no affiliation with Losmandy, poor word choice of "customer" in my last post implied that.
Keith ---In Losmandy_users@..., <tomh@...> wrote :
Keith
None of the user groups are run by Losmandy.? Now Ren¨¦ is a co-owner
of the Gemini-2 users group and the Gemini users group.
Tom H
?Good suggestion Tom. To be honest, I'm sometimes not sure
which is the appropriate group to post to since as a
customer, Losmandy and Gemini are interchangeable.. I'd
like to test a few more things, then will repost in Gemini
II.
Keith
|
Gentlemen,
I was recently enlightened about polar alignment and drift in the two axis on the Astro-Physics group.? The polar alignment most of us are used to is designed to provide minimal declination drift, not minimum RA drift.? If we were on the moon there would be no difference.? But due to atmospheric refraction, the minimum declination drift polar alignment results in RA drift.? What Roland Christian recommends is to align the azimuth in the normal fashion minimizing declination drift at the celestial equator and the meridian.? But instead of going East or West to do the elevation adjustment, move? up the meridian to the zenith.? Then adjust the elevation to minimize the RA drift.? Good PEC helps here to identify the RA drift. ?
The difference is that the old method was designed when scopes typically did not have a declination adjustment that could be used in guiding, but the RA could be adjusted for guiding.? So the polar alignment routine was designed to minimize declination drift over the entire sky.? With scopes that can effectively guide in declination, the new polar alignment routine minimizes both declination and RA drift within 30 degrees of the meridian, where most imaging is done.? This can allow for unguided imaging in this range and is still readily guided outside of this range.
I have yet to try this, but the theory seems sound.? Rolando would not steer me wrong.
Greg
|
Hi!
THis is really interesting (as are most things related to this tricky issue of PA...)!!
Could you say a little more about how this adjustment at zenith would be done? For instance, I use PHD2 for drift alignment - I guess that is not useful in this situation. So more precisely how could I try this?
Best,
Magnus
|
This subject is a current topic on the Astro-Physics group (#59424) and the subject of using PHD2 was discussed.? PHD2 displays both the DEC and RA drift in the Drift Alignment tool.? So instead of letting PHD2 slew the scope to the East for the altitude adjustment, slew it manually to the zenith.? Turn on your mount's PEC to minimize RA motion.? Run the drift session for a full worm cycle to cancel out the effects of the remaining PE.? Adjust the altitude to minimize the RA drift.? There is no? point in doing the Eastern altitude adjustment first as this will always change for the minimum RA drift adjustment.
I found it interesting to think about this at the limit of doing polar alignment at the equator, where the intersection of the celestial equator and the meridian is also the zenith.? Using this method there you would not move the scope from the zenith for either adjustment
Greg
|
Hi Greg,
Thanks for the notes.??
When using the PHD2 drift tool, it is only the DEC drift you want to null out (not RA drift), as that tool says in its instructions.
[ repeat these 2 steps...
When you are pointing first near meridian and celestial equator, only adjust the AZ knobs.
When pointing next, as far East or West as possible, then again look only at DEC drift, and null out with the AZ knob.??
]
Repeat...[...] til you are satisfied.
You really don't need to bother about the PEC... you can ignore errors in that for the drift alignment (by PHD2 anyway).
Best, Michael?
On Nov 5, 2017 1:03 PM, " beneckerus@... [Losmandy_users]" < Losmandy_users@...> wrote:
?
This subject is a current topic on the Astro-Physics group (#59424) and the subject of using PHD2 was discussed.? PHD2 displays both the DEC and RA drift in the Drift Alignment tool.? So instead of letting PHD2 slew the scope to the East for the altitude adjustment, slew it manually to the zenith.? Turn on your mount's PEC to minimize RA motion.? Run the drift session for a full worm cycle to cancel out the effects of the remaining PE.? Adjust the altitude to minimize the RA drift.? There is no? point in doing the Eastern altitude adjustment first as this will always change for the minimum RA drift adjustment.
I found it interesting to think about this at the limit of doing polar alignment at the equator, where the intersection of the celestial equator and the meridian is also the zenith.? Using this method there you would not move the scope from the zenith for either adjustment
Greg
|
Michael Herman