Still chasing differential drift on my RASA
87 views
Skip to first unread message
Chris Lewicki
Mar 14, 2021, 1:56:14 AM3/14/21
to Open PHD Guiding
I have a RASA 11 + ZWO ASI294MC on a hypertuned Celestron CGX, concrete pier-mount. I guide with a William Optics GuideStar 61 and a ZWO ASI 290MM.
I did 2 minute exposures, and it averaged about 2.5s readout, for a spacing of 122.5 sec on average, for a little over 2 hours total.
I used actual timestamps from the FITS headers in the file to make sure that I had accurate image spacing in the log file, although given the peculiarities of the INDI camera simulator, I'm not sure that I got the guide log and Main scope log to exactly overlap. In the ~2 hour session, they're probably within 10 minutes of each other.
The drift looks almost perfectly linear (R=0.9984) in overall distance, at a rate of about 0.5 pixel/min in angular drift. As I usually do 2 minute exposures to get to sky background at my site, this means that I'm getting about 1 pixel of smear for regular OSC imaging, and if I go to longer exposures for narrowband, I can expect 5 pixels of smear if I'm doing 10 minute exposures.
During long guiding sessions I regularly see large amounts (5-15% of the frame over a few hours) of drift on the primary camera, while the guide scope shows good tracking during the session. I've shared this here before, and the assessment has been "differential drift" given the external guide scope.
I've done the usual work of making sure the guide scope is securely mounted, cable runs are careful to prevent any tension from the guide camera/scope, etc. It has reduced the drift a bit.
Given that my main scope is a RASA, I don't have the option to use an OAG, and I'm determined to find a solution to this, if I have to program an extension to do it!
Towards that end, I've set up INDI/Ekos/Kstars to feed PHD2 saved frames from my recorded main-camera imaging session, and post-processed the log files to make it "look" like a guiding session, although of course it wasn't able to make any corrections. These frames were acquired while PHD2 was guiding through the guide scope, so the resulting drift SHOULD only be the effects of differential drift.
Given that my main scope is a RASA, I don't have the option to use an OAG, and I'm determined to find a solution to this, if I have to program an extension to do it!
Towards that end, I've set up INDI/Ekos/Kstars to feed PHD2 saved frames from my recorded main-camera imaging session, and post-processed the log files to make it "look" like a guiding session, although of course it wasn't able to make any corrections. These frames were acquired while PHD2 was guiding through the guide scope, so the resulting drift SHOULD only be the effects of differential drift.
I did 2 minute exposures, and it averaged about 2.5s readout, for a spacing of 122.5 sec on average, for a little over 2 hours total.
I used actual timestamps from the FITS headers in the file to make sure that I had accurate image spacing in the log file, although given the peculiarities of the INDI camera simulator, I'm not sure that I got the guide log and Main scope log to exactly overlap. In the ~2 hour session, they're probably within 10 minutes of each other.
The drift looks almost perfectly linear (R=0.9984) in overall distance, at a rate of about 0.5 pixel/min in angular drift. As I usually do 2 minute exposures to get to sky background at my site, this means that I'm getting about 1 pixel of smear for regular OSC imaging, and if I go to longer exposures for narrowband, I can expect 5 pixels of smear if I'm doing 10 minute exposures.
You can take a look at the real guide session and "synthesized main camera" guide session in this file.
I use Kstars/Ekos/INDI for mount control, and I've posted a pair of feature ideas there which would allow Kstars to request "dithers" which would accommodate this drift. For context, you can see them in Issue #59 (plate solving) and Issue #60 (the dither part).
I use Kstars/Ekos/INDI for mount control, and I've posted a pair of feature ideas there which would allow Kstars to request "dithers" which would accommodate this drift. For context, you can see them in Issue #59 (plate solving) and Issue #60 (the dither part).
I had previously been thinking of this problem as a centering problem, but now I'm seeing it as a continuous drift problem. Rather than correcting for the drift with discrete (dither) jumps, it would be much better to detect the drift rate, and build it into the overall drift rate correction that PHD2 already accounts for for things like Predictive PEC. This would prevent the pixel smear entirely. To implement it would probably require another INDI camera driver which could be set up to watch a folder of images and send them over to PHD2 to evaluate the drift rates. Since it represents a different imaging system, with a different alignment of control, it could be implemented with a second instance of PHD2 configured for the main scope and parameters, but the measured drift rates could be fed to the primary guider as a "bias".
I realize I'm chasing just a few pixels here, but I'm happy to be obsessive about that and get my stars as small and round as the optics will produce!
I'd be interested in people's assessment of the log files, and thoughts on the above idea to try and correct the drift.
I realize I'm chasing just a few pixels here, but I'm happy to be obsessive about that and get my stars as small and round as the optics will produce!
I'd be interested in people's assessment of the log files, and thoughts on the above idea to try and correct the drift.
steve
Mar 14, 2021, 8:32:54 AM3/14/21
to open-phd...@googlegroups.com
On 14/3/21 7:56, Chris Lewicki wrote:
> INDI/Ekos/Kstars
Hi
Simply set the EKOS scheduler to recenter every, say, 6 frames.
Make one sequence of six frames and repeat it x times. Don't forget to
check 're-align before starting each job'.
Cheers and HTH.
Chris Lewicki
Mar 14, 2021, 5:06:40 PM3/14/21
to open-phd...@googlegroups.com
Hi Steve,
A few points of discussion:
1) I don't believe EKOS scheduler supports "recenter" at some frequency (as do camera sequences allow for focus and/or DITHER at some frequency). It DOES support re-alignment at the beginning of a new sequence, but this means dividing up the sequence into chunks, and also suffering the overhead of the time interruption of re-alignment/re-guiding at this periodicity, plus the introduction of sequencing risk from failing to align for some reason, or failure to re-start guiding. EKOS scheduling is already tedious enough, so I'd prefer to avoid this complication. If I'm just wrong about the feature, I'd appreciate a pointer on how to check the appropriate
Platforms like Voyager Astro probably DO allow you to implement this inter-sequence in a DragScript action, but I'd really prefer not to run windows for this solution. My proposal in the two Ekos issues I identified could deal with it in "dither" without interrupting sequences or pointing, but it requires the substantial changes described in the changes.
1) I don't believe EKOS scheduler supports "recenter" at some frequency (as do camera sequences allow for focus and/or DITHER at some frequency). It DOES support re-alignment at the beginning of a new sequence, but this means dividing up the sequence into chunks, and also suffering the overhead of the time interruption of re-alignment/re-guiding at this periodicity, plus the introduction of sequencing risk from failing to align for some reason, or failure to re-start guiding. EKOS scheduling is already tedious enough, so I'd prefer to avoid this complication. If I'm just wrong about the feature, I'd appreciate a pointer on how to check the appropriate
Platforms like Voyager Astro probably DO allow you to implement this inter-sequence in a DragScript action, but I'd really prefer not to run windows for this solution. My proposal in the two Ekos issues I identified could deal with it in "dither" without interrupting sequences or pointing, but it requires the substantial changes described in the changes.
2) While a sequence kludge in this area would solve the alignment problem I mentioned in an earlier post (and above), it doesn't address what I brought up in this new post - the pixel-smear of differential drift during longer integrations.
I'll admit I am definitely "optimizing" in looking for a solution to this, and it seems like something that's easily solved in an elegant way with the data that is already available to the system. The current "Inner" PHD2 loop running at high frequency (seconds) with the Guide Scope to deal with typical mount tracking, and a slower "Outer" loop (minutes) running on the Imaging Camera frames to deal with the differential flexure drift rates. It may be something that you react to with "Why bother? There are workarounds." But you could also use that argument against previous stages of development: "PEC modeling, why bother?" This is just another potential evolution in the refinement of more-perfect tracking capabilities.
I'll admit I am definitely "optimizing" in looking for a solution to this, and it seems like something that's easily solved in an elegant way with the data that is already available to the system. The current "Inner" PHD2 loop running at high frequency (seconds) with the Guide Scope to deal with typical mount tracking, and a slower "Outer" loop (minutes) running on the Imaging Camera frames to deal with the differential flexure drift rates. It may be something that you react to with "Why bother? There are workarounds." But you could also use that argument against previous stages of development: "PEC modeling, why bother?" This is just another potential evolution in the refinement of more-perfect tracking capabilities.
I produced another synthetic guide log (linked here) based on a ~5 hour imaging session I did with M101 last week, which includes 3-pixel spiral dither every 3 frames / 6 minutes. There's about 4 arc-minutes total drift, but in the middle, there's an ~hour period which has no drift at all. It could be that the changes in RA/DEC drift cancelled each other out for a time. With this longer track you can also see the long-term trend is not actually linear, but something that looks like a slow sine-wave of the scope-orientation/flex being affected by gravity as the projected angle gets steeper towards the meridian.
It could be that the easiest thing (no-code) is to simply modify my guide scope mounting to be side-by-side, tying the guider more directly to the mount, and bringing the differential flexure to negligible values. However, hanging the guide scope off the main scope, and the resulting differential flexure, is a common enough problem, I suspect that a software solution to it would be useful (and cheaper) for many.
In short, a solution to cancel out differential drift has the following potential benefits:
1) More useable frame for maximum FOV in integration
2) Improved FOV/overlap in mosaic imaging
3) Lower Eccentricity, FWHM in long integrations
4) More integration time on target by correcting for drift during guiding, rather than with re-alignment
5) Easier "Push-Here-Dummy" guiding without requiring tedious configuration of multi-platform sequencing software, care-abouts of particular camera drivers, etc. Just point PHD2 at your imaging folder and forget it.
-Chris
@interplanetary
@interplanetary
--
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/GcoJt2PzIFQ/unsubscribe. To unsubscribe from this group and all its topics, send an email to open-phd-guiding+unsubscribe@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/acb1a6a6-bd24-7951-fa63-1cc837ace342%40gmail.com.
Steve
Mar 14, 2021, 5:35:43 PM3/14/21
to OpenPHD Guiding
Hi
Yes, I should perhaps have used the term realign rather than recenter. It doesn't however involve dividing the sequence into chunks, adds little if anything to the time needed and (thinking you may have a clear night(s) coming up) takes care of the drift. JTOL that you wanted something now.
Look at:
Job completion conditions > Repeat for
Cheers
----
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/GcoJt2PzIFQ/unsubscribe. To unsubscribe from this group and all its topics, send an email to open-phd-guidi...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/acb1a6a6-bd24-7951-fa63-1cc837ace342%40gmail.com.
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/km9legd1.2b1a53d9-074d-4f16-9e32-2330721f6934%40we.are.superhuman.com.
Chris Lewicki
Mar 14, 2021, 5:56:59 PM3/14/21
to open-phd...@googlegroups.com
Look at:Job completion conditions > Repeat for
Ah - got it. You could divide the sequence duration by whatever re-alignment frequency you want, and check this option. I guess you could Track/Focus/Align/Guide on the first one, and then include a second instance with the "Repeat For" option, and just "Align/Guide" checked.
Might add this in with periodic re-focus I'm already doing.
Thanks for the clarification.
-Chris
@interplanetary
@interplanetary
On Sun, Mar 14, 2021 at 2:35 PM, Steve <stev...@gmail.com> wrote:
HiYes, I should perhaps have used the term realign rather than recenter. It doesn't however involve dividing the sequence into chunks, adds little if anything to the time needed and (thinking you may have a clear night(s) coming up) takes care of the drift. JTOL that you wanted something now.Look at:Job completion conditions > Repeat forCheers
--
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/GcoJt2PzIFQ/unsubscribe. To unsubscribe from this group and all its topics, send an email to open-phd-guiding+unsubscribe@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/acb1a6a6-bd24-7951-fa63-1cc837ace342%40gmail.com.
--
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-guiding+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/km9legd1.2b1a53d9-074d-4f16-9e32-2330721f6934%40we.are.superhuman.com.
--
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/GcoJt2PzIFQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to open-phd-guiding+unsubscribe@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/CAB7Jd2K%2BJk%3DCswYkM5HeLG2qGajLHStMReiHmb1TyPE0sVH5KQ%40mail.gmail.com.
Ron Kramer
May 24, 2021, 6:26:32 PM5/24/21
to Open PHD Guiding
I fought with my RASA 11 for quite some time. They are an awkward beast. I have a V1 and it guided okay at about .70
I replaced it with a RASA 11 V2 which for some reason was a lot heavier. I put more counter weight on. This was a nightmare.
I had moved the guide scope to the top (NOT GOOD) and fought with .90-1.60 for months.
Finally gave up reworked the entire mount, replaced guide scope I could fit a smaller one UNDER the scope taking the moment arm down.
Moved most of my stacked gear and lightened the load. Removed a 7lb counterweight... moved the other weights up...
I was happy with my make over. (got a picture of your RASA and stuff on the mount? to share how you have it stacked up?
I always had EGGS - no longer.
I replaced it with a RASA 11 V2 which for some reason was a lot heavier. I put more counter weight on. This was a nightmare.
I had moved the guide scope to the top (NOT GOOD) and fought with .90-1.60 for months.
Finally gave up reworked the entire mount, replaced guide scope I could fit a smaller one UNDER the scope taking the moment arm down.
Moved most of my stacked gear and lightened the load. Removed a 7lb counterweight... moved the other weights up...
I was happy with my make over. (got a picture of your RASA and stuff on the mount? to share how you have it stacked up?
I always had EGGS - no longer.
--
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/GcoJt2PzIFQ/unsubscribe. To unsubscribe from this group and all its topics, send an email to open-phd-guidi...@googlegroups.com. To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/acb1a6a6-bd24-7951-fa63-1cc837ace342%40gmail.com.
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/km9legd1.2b1a53d9-074d-4f16-9e32-2330721f6934%40we.are.superhuman.com.
--
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/GcoJt2PzIFQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to open-phd-guidi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/CAB7Jd2K%2BJk%3DCswYkM5HeLG2qGajLHStMReiHmb1TyPE0sVH5KQ%40mail.gmail.com.
--
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.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/km9p0bau.350eec33-5638-4fd6-84b5-36c7317831d7%40we.are.superhuman.com.
Ron Kramer
Reply all
Reply to author
Forward
0 new messages