¿ªÔÆÌåÓý

ctrl + shift + ? for shortcuts
© 2025 Groups.io
  1. ServoCAT
  2. Topics

Promising Autoguiding results in PHD2

 

Thanks to the moon i have been able to dedicate the last 3 clear nights on getting autoguding to work with my ServoCat driven 30" F3.4 Starstructure dob.
?
The first night i used a guidespeed of 25 arc sec/sec then 20 arc sec/sec in ServoCat Sky, no matter what i did with the settings in PHD2 the guide corrections would cause the scope to shake similar to wind gusts which showed up in the subs as double or triple stars.
?
On night 2 i used a guidespeed of 15 arc sec/ sec, this time the guiding looked pretty good but it kept loosing the guidestar after some time so i suspect that one of the motors was changing directions = backlash which caused the loss in tracking.
?
The above scenario is similar to what i have seen a few others run in to trying to autoguide with their dobs after searching the internet.
?
Then it struck me, that the guidespeed PHD2 was measuring was not the same as in the PHD2 calibration section.
?
The measured guidespeed was about 0.44 x Sidereal when using 15 arc sec/sec in ServoCat sky, but in the PHD2 section 1.01x Sidereal was set.
Which means that the calibration guide pulse length was calculated for twice the actual guidespeed = the calculated pulses for calibration and guiding would be way shorter than they should be.
?
So i changed the guidespeed in the PHD2 section to 0.44 x Sidereal, (does not change it in the mount, only for calculations). For the value to stay it is essential that you don't connect PHD2 with the mount, so one should connect to "ask for coordinates". If you connect to the mount, at least with the Meade Generic driver, it will automatically change to 1.01x in PHD2.
?
So i calibrated once again and this time the guiding looked like something i could work with by only changing the settings in PHD2.
?
On a sidenote to my great surprise PHD2 could always calibrate during all my attempts these 3 nights with almost no Orthogonality error or any warnings except the "DEC and RA rates varied by an unexpected amount" which went away after i entered the correct guidespeed in the calibration section. For calibration i trie both bright and dim stars and always at an altitude between 30 and 40 degrees which the Alt/az enthusiast Niels Haagh recommends in his alt/az PHD2 guide: /?
?
So with that out of the way the tweaking began, i ran the guide assistant which only recommended that i should lower the min move on both axis which did not make sense since that causes more aggressive guiding.
?
After a while the picture started to become clear, if you want to guide "a huge ship" you can't do that by bouncing it around like a small stick, the stick will instantly settle but the huge ship will start to oscillate and you will loose control. If the ship starts to drift, you want to cancel out the drift, then slowly bring it back on track.
?
Finally i found settings that where able to give me guiding around 1.3" arc seconds RMS. Tracking accuracy unguided was about 2.0-2.5 arc seconds. Alt and Az drift rates where at about 5 arc seconds per minute for each axis. In reality the improvement manifested as many 10s subs in the 2.6-3.0 FWHM range guided, if i remember correctly no unguided subs where below 3.0 FWHM.
?
Here are the settings i used:
Default MinMo on both axis is 0.20. I raised this to 0.40 so that the guider does not correct just because of slight deviations.
Default Hysteresis on both axis is 0.10. I set this to 0 since i don't want the guider to repeat previous corrections to avoid risk of oscillations.
Default RA and DEC Agression is 0.70 on both axis, i lowered this to 0.20-0.40 depending on conditions to avoid too agressive corrections.
Default Max Guidepulse duration on RA and DEC is 2500 ms, i set this as low as 50-400 ms depending on conditions in order for the guider to slowly bring back the scope to the guide star and not send a long pulse to quickly center the star.
?
Of course every rig is different, the difference may even be huge but summarized the key to autoguiding a large dob seems to be to first of all get the unguided tracking as good as possible with the 2-star and very accurate gear ratios and then use a slow guidespeed and short guidepulses since the opposite will cause too much movement which will not settle like on a small scope.
?
For guiding i used a 50mm guidescope, 120MM Mini guidecamera and ST4 cable. Using 180mm of focal length on the guidescope to guide the main scope at 2725mm is of course insanity but this is how i see it:
?
The guiding setup is good enough to verify if the guiding is about 0.8" arc seconds RMS, so it will be able to provide corrections that exceed my mounts unguided performance.
?
However, since the platescale of my setup is 0.297 arc seconds per pixel it is of course impossible for a 50 mm guidescope to make corrections within such a small tolerance, but this is a dob not a CDK 700 or ASA 800 and even if my scope would be able to stay witin 0.297 arc seconds it would not be seen since the seeing here is to bad, in the same way the seeing also hides tracking errors.
?
As for field rotation the guidescope and main scope are pointing at the same spot, or rather i know which spot on the Bullz eye in PHD2 my main scope is pointing. The chances of a guidestar being on that exact same spot are not big, but here is how i see it:
?
I like watching the subs roll in, so if the object starts to drift in the main scope i can re-center. If you want to sleep then NINAS re-center after drift will take care of that and PHD2 will automatically pick a guidestar and start guiding again after the re-center.?
?
I find it interesting that i have so far been able to image and get decent results despite the fact that the unguided performance seems to be about 2.0-2.5 arc seconds RMS.
?
I wonder though, how much in all this is the star moving because of seeing vs the scope itself moving, and when it comes to guiding it is of course strongly affected by seeing conditions.
?
It will be interesting to see how this evolves in the coming sessions!
?
Clear Skies
Gabriel
?
?
?
 

¿ªÔÆÌåÓý

Thank you Gabriel for the hard work and time you put in in this project so far.?
?I am hopeful to use this in some way with my rig, using SharpCap. We will see.

Thank you
Mickey


On Feb 12, 2025, at 6:59?AM, Gabriel Wiklund via groups.io <gabriel.wiklundholeshot@...> wrote:

?
Thanks to the moon i have been able to dedicate the last 3 clear nights on getting autoguding to work with my ServoCat driven 30" F3.4 Starstructure dob.
?
The first night i used a guidespeed of 25 arc sec/sec then 20 arc sec/sec in ServoCat Sky, no matter what i did with the settings in PHD2 the guide corrections would cause the scope to shake similar to wind gusts which showed up in the subs as double or triple stars.
?
On night 2 i used a guidespeed of 15 arc sec/ sec, this time the guiding looked pretty good but it kept loosing the guidestar after some time so i suspect that one of the motors was changing directions = backlash which caused the loss in tracking.
?
The above scenario is similar to what i have seen a few others run in to trying to autoguide with their dobs after searching the internet.
?
Then it struck me, that the guidespeed PHD2 was measuring was not the same as in the PHD2 calibration section.
?
The measured guidespeed was about 0.44 x Sidereal when using 15 arc sec/sec in ServoCat sky, but in the PHD2 section 1.01x Sidereal was set.
Which means that the calibration guide pulse length was calculated for twice the actual guidespeed = the calculated pulses for calibration and guiding would be way shorter than they should be.
?
So i changed the guidespeed in the PHD2 section to 0.44 x Sidereal, (does not change it in the mount, only for calculations). For the value to stay it is essential that you don't connect PHD2 with the mount, so one should connect to "ask for coordinates". If you connect to the mount, at least with the Meade Generic driver, it will automatically change to 1.01x in PHD2.
?
So i calibrated once again and this time the guiding looked like something i could work with by only changing the settings in PHD2.
?
On a sidenote to my great surprise PHD2 could always calibrate during all my attempts these 3 nights with almost no Orthogonality error or any warnings except the "DEC and RA rates varied by an unexpected amount" which went away after i entered the correct guidespeed in the calibration section. For calibration i trie both bright and dim stars and always at an altitude between 30 and 40 degrees which the Alt/az enthusiast Niels Haagh recommends in his alt/az PHD2 guide: /?
?
So with that out of the way the tweaking began, i ran the guide assistant which only recommended that i should lower the min move on both axis which did not make sense since that causes more aggressive guiding.
?
After a while the picture started to become clear, if you want to guide "a huge ship" you can't do that by bouncing it around like a small stick, the stick will instantly settle but the huge ship will start to oscillate and you will loose control. If the ship starts to drift, you want to cancel out the drift, then slowly bring it back on track.
?
Finally i found settings that where able to give me guiding around 1.3" arc seconds RMS. Tracking accuracy unguided was about 2.0-2.5 arc seconds. Alt and Az drift rates where at about 5 arc seconds per minute for each axis. In reality the improvement manifested as many 10s subs in the 2.6-3.0 FWHM range guided, if i remember correctly no unguided subs where below 3.0 FWHM.
?
Here are the settings i used:
Default MinMo on both axis is 0.20. I raised this to 0.40 so that the guider does not correct just because of slight deviations.
Default Hysteresis on both axis is 0.10. I set this to 0 since i don't want the guider to repeat previous corrections to avoid risk of oscillations.
Default RA and DEC Agression is 0.70 on both axis, i lowered this to 0.20-0.40 depending on conditions to avoid too agressive corrections.
Default Max Guidepulse duration on RA and DEC is 2500 ms, i set this as low as 50-400 ms depending on conditions in order for the guider to slowly bring back the scope to the guide star and not send a long pulse to quickly center the star.
?
Of course every rig is different, the difference may even be huge but summarized the key to autoguiding a large dob seems to be to first of all get the unguided tracking as good as possible with the 2-star and very accurate gear ratios and then use a slow guidespeed and short guidepulses since the opposite will cause too much movement which will not settle like on a small scope.
?
For guiding i used a 50mm guidescope, 120MM Mini guidecamera and ST4 cable. Using 180mm of focal length on the guidescope to guide the main scope at 2725mm is of course insanity but this is how i see it:
?
The guiding setup is good enough to verify if the guiding is about 0.8" arc seconds RMS, so it will be able to provide corrections that exceed my mounts unguided performance.
?
However, since the platescale of my setup is 0.297 arc seconds per pixel it is of course impossible for a 50 mm guidescope to make corrections within such a small tolerance, but this is a dob not a CDK 700 or ASA 800 and even if my scope would be able to stay witin 0.297 arc seconds it would not be seen since the seeing here is to bad, in the same way the seeing also hides tracking errors.
?
As for field rotation the guidescope and main scope are pointing at the same spot, or rather i know which spot on the Bullz eye in PHD2 my main scope is pointing. The chances of a guidestar being on that exact same spot are not big, but here is how i see it:
?
I like watching the subs roll in, so if the object starts to drift in the main scope i can re-center. If you want to sleep then NINAS re-center after drift will take care of that and PHD2 will automatically pick a guidestar and start guiding again after the re-center.?
?
I find it interesting that i have so far been able to image and get decent results despite the fact that the unguided performance seems to be about 2.0-2.5 arc seconds RMS.
?
I wonder though, how much in all this is the star moving because of seeing vs the scope itself moving, and when it comes to guiding it is of course strongly affected by seeing conditions.
?
It will be interesting to see how this evolves in the coming sessions!
?
Clear Skies
Gabriel
?
?
?
 

I appreciate your gratitude Mickey, luckily the hard work has also been very interesting and rewarding!
?
/Gabriel
 

Thanls Gabriel,
Really appreciate how you share your experience !
Try astrophotography using my 36" Fullum alt/az is a target for future and with your explainations it becomes clearer what to do.
And the ressource is really great !
I'm currently in the gear ratio tracking improvement phase, which you pointed out to be a must have before thinking of astrophoto.
When you say "the unguided performance seems to be about 2.0-2.5 arc seconds RMS" I believe this error of 2.0-2.5 arc seconds is for a amount of time, right ?
Per seconds ?
For example you measured that the unguided star was drifting in the field for 20' in 10 min and so 20*60=1200 arc second in 10min so? 1200/(10*60)=2 arc seconds / s, right ?
Just curious what accuracy you achieved to compare with my actual precision of tracking.
Regards
Rapha?l
 
Edited

Hello Raphael
Glad it is of interest!
?
2.0-2.5 arc seconds unguided for my sub exposure time which? is 10-13s with only a slice of time from 1 night as reference so would need more nights to verify and i don't know the tracking accuracy for the images i have aquired so far. The Fullum alt/az scopes look really cool, exiting times ahead if you try imaging!
?
When it comes to the mesuread drift rate it was generally around 5 arc seconds per minute per axis so that would be 0.08 arc seconds per axis per second = total of 0.16 arc seconds drift error per second. If you take 0.16 arc seconds x 13 seconds = 2.08 arc seconds so the numbers agree fairley well.
?
Best regards
Gabriel
?
?
 
Edited

I would also like to take the opportunity to correct myself. I have now read the PHD2 documentation (mostly the troubleshooting section) and according to what is written there, if you get large/abrupt guidestar deflections like i was getting, wildly changing the settings in PHD2 is not likley to solve the problem, at least not in the long run, it is almost always due to hardware issues which should always be adressed first.?
?
This is what the PHD2 documentation states about the symptoms i had:
?

Large/Abrupt Guide Star Deflections

Most users eventually encounter situations where the guide star appears to make a large, abrupt excursion away from the lock-point. ?The great majority of these problems?arise from neither the mount nor?PHD2's guide commands. ?Instead, they usually come from?unwanted mechanical movement in the gear that is riding on top of the mount, especially the guide camera/guide scope assembly. ?This is especially true if the large deflections occur in declination because the Dec motor is normally idle except for executing the very short, relatively infrequent guide commands it receives. ?The unwanted mechanical movement usually comes from several sources:
  1. Tiny movements of the various components in the guiding assembly as a result of the changing gravitational forces while the mount tracks the target object
  2. Dragging, binding, or snagging of cables, especially those that are connected to the guide camera
  3. Wind gusts or less commonly, effects from camera filter changes, auto-focusing, or mirror movement
  4. Use of mount features for backlash compensation - these should not be used with PHD2 guiding
Before rejecting these things as likely sources of? problems, think again about the tiny measurement scales and tolerances described in the previous section. ?With many guiding set-ups, a movement of only 5 microns can create an apparent tracking error (guide star deflection) of over 6 arc-sec, the equivalent of many star diameters. Every mechanical interface, every set-screw, every movable element has the potential to shift or move on its own by these tiny amounts. ?Even when cables have been routed in a purposeful way, they may bind or pull in certain sky positions or after a meridian flip. ?Cable ties or ribbed plastic cable guides hare small protrusions that can briefly catch on stationary parts of the mount. ?For large Dec deflections, it's easy to determine if these things are coming into play. ?Just use the PHDLogViewer tool to zoom in on the time of interest and see if the deflection was immediately preceded by a correspondingly large guide command in the direction of movement. ?In most cases, you will find this didn't happen. ?It can sometimes happen at the beginning of a guide session if you're using PHD2 Dec backlash compensation, but those events should disappear quickly. ?If the abrupt deflections occur in RA, the analysis is less straightforward because the RA motor runs continuously. ?But even then, unusually large, randomly space deflections are more likely to arise from the sorts of mechanical problems described here than from errors in the RA drive system.
?
?
So i took a closer look at how my guidescope is mounted and it could be a problem, i have tightened the screws and we will se if it makes a difference.
?
Guidespeed was probably also the problem in the beginning and just to see what will happen next time i will try 10 arc seconds /per second in ServoCat Sky, which is actually what the link with the Panther mount recommends and also the defualt value in the Meade Generic driver.
?
I also read in the documentation that you should not change the guidespeed in PHD2 "manually" you should reset and create a new profile where that is entered in the first steps, i noticed that if i enter 10 arc seconds per second in the Meade Driver and in the beginning in PHD2 when creating the profile, PHD2 does not change it to 1.01 arc seconds like i wrote earlier so now both the Meade driver, PHD2 and ServoCAT sky has 10 arc seconds /per second, i hope it clears up tonight so that i can give it a try! (i have also set backlash comp to 0 on both axis in ServoCat Sky this time)
?
In the first calibration, If the measured guidespeed in PHD2 turnes out to be say 5 arc seconds per second instead of 10, then i will re-do the profile and enter the new speed (0.33 x Sidereal in that case) so that the correct guide-pulse length is calculated for calibration and guiding.
?
Here is the PHD2 user guide: https://openphdguiding.org/PHD2_User_Guide.pdf
?
/Gabriel
?
?
 

Thanks Gabriel for clarification.
You indicate your unguided precision is "0.08 arc seconds per axis per second".
In the last session I did to improve my ratio, I measured initially a 0.05 arc seconds per second in az and 0.32 arc seconds per second in alt.?
I have then calculated at home the correction factor and the new ratio in alt and az have been put in servocat but bad weather then.
I will check next time I have good weather at observatory what are my new az and alt precision. But I will certainly be globally not that far from you on precision.
I then have some questions :
?
* I see wonderfull pictures taken with alt/az scopes here. Are these pictures taken using the autoguiding method ? Or with unguided scopes ? In the latter case what is your precision of these unguided scope in arc sec/ sec ??
Because I wonder if, as a newbie in astrophoto, autoguiding is necessary or not...
If I read correctly your initial message Gabriel, I understand that you have improved precision of guiding by a factor 2 when using the autoguiding method (from 2-2.5 arc" to 1.3 arc " for your exposure time of <13"). It this improvment necessary for a newbie ?
?
* I have a 10" (f=1150 mm) scope in // of the 36" (f=3150 mm). This could be used for autoguiding, no ?
The only potential problem is that the parallellism of the 2 scope may vary a little bit when I change location in the sky because the 10" is fixed by its back on a plate (with adjustment screw at back of the plate to parallelize the 10"), and I believe a little bit of flexion occurs certainly. Not a big deal in visual, but in astrophoto I don't know... Maybe on a given zone of the sky it will not change its flexion for a total 30 min posing? I don't know...
?
Thnaks for your feedback
Rapha?l
?
 

I have seen very few images taken with dobs that have been autoguided, i can only remember one and that was on an EQ platform from Tom O, so i have only read about 2 that have succeded in doing this in alt/az (Dan Price, and another Obsession owner).
?
So you don't need autoguiding, and you don't need to take even 10 second exposures, with a modern low read noise camera you can get away with 3 seconds even.
?
The fact that your guidescope has a long focal length is a lot better than the 50mm guidescope i am using which only has 180mm focal length, even if they are not perfectly aligned i would still give it a try and see the results, then you know for sure (if you want to try autoguiding).
?
The only reason i am trying to autoguide is because i am having great fun problemsolving and i am very fascinated bu the subject of tracking accuracy, it may even be difficult to tell the difference on my future images even if i get autoguiding to work well.
?
Even if the guidescope and mainscope are not pointing at the same spot in the sky, as you are indicating i suspect you can still keep the object reasonably centered probably for even more than 30 minutes, then you would just re-center and pick a new guidestar.
?
/Gabriel
?
?
 

Thanks Gabriel,
clear !!!

1 - 9 of 9