PHD2 graph - Comments please

[malc-c]

Guys,

 

It's been four years since I last used my 200P/HEQ5 combo, but with all this lockdown I needed something to stem the boredom, so over the past four / five days I've stripped down and overhauled the scope and took advantage of the clear(ish) evenings to set things up.  The optics were cleaned and re-collimated and the star test at both extremes of focus produced nice circular images with clear evenly spaced concentric rings.

The ST80 was removed and the 9 x 50 finder converted to a finder/guider using the QHY5.  Over the past couple of nights I've spent time checking focus, aligning the finder to the main scope and Celestron star finder pointer so that all three have the target central in the field of view.  The mount was polar aligned using SharpCap's tool to 10 arc seconds.  So that left two things to check out, the first being to gather some guide data, and to generate a PE curve for EQMod to use in its correction.

So having tweaked EQMod's pulse guiding settings and a few PHD2 settings, the calibration ran fine with about 18 steps in either direction E/W and 10 steps N/S - It didn't complain and PHD2 started guiding.  I was pointing at NGC2903 (darks are still running so it will be a while before I get to play with processing) so the weight bar was out fairly level, with a declination of 21.4 degrees.  I've attached the resulting log file trace, and would like some opinions if this is good enough for guiding.  The RMS values of 1.24 arc seconds for RA and 0.59 arc seconds for DEC sound OK, but the graph doesn't look as smooth as some I used to get when using the ST80.

One last question.  There is obviously an underlying sine wave on the RA trace, which could suggest the mount needs PE training.  In order to do so I guess I should fit the camera to the main scope and log the guiding trace through it in order to get better accuracy ?

[alacant]

10 hours ago, malc-c said:

underlying sine wave

Hi

If you guide RA using PHD2's PPEC with an initial preriod of,say 120s, it should flatten the sine after a few minutes.

Before the days of the PPEC algorithm, I wasted a lot of time trying to PEC train my eq6;)

HTH

[masjstovel]

51 minutes ago, alacant said:

Hi

If you guide RA using PHD2's PPEC with an initial preriod of,say 120s, it should flatten the sine after a few minutes.

Before the days of the PPEC algorithm, I wasted a lot of time trying to PEC train my eq6;)

HTH

Not wanting to hijack the thread, but how did you PEC train the EQ6? If you could explain it like i was 5 years old i would really appreciate it - its a mess in my head now between PHD2, PECprep, and EQMOD, where do i do what, when and how sort of. If @malc-c doesn't want it here on the thread, i would love it on PM

[jimjam11]

1 hour ago, alacant said:

Hi

If you guide RA using PHD2's PPEC with an initial preriod of,say 120s, it should flatten the sine after a few minutes.

Before the days of the PPEC algorithm, I wasted a lot of time trying to PEC train my eq6;)

HTH

Likewise, I havent found eqmod pec helpful vs straight guiding using the PPEC algorithm.

[jimjam11]

18 minutes ago, masjstovel said:

Not wanting to hijack the thread, but how did you PEC train the EQ6? If you could explain it like i was 5 years old i would really appreciate it - its a mess in my head now between PHD2, PECprep, and EQMOD, where do i do what, when and how sort of. If @malc-c doesn't want it here on the thread, i would love it on PM

Eqmod has an AutoPec feature. Get everything setup and start guiding on a target, start autopec and it will do everything for you.

 

[jimjam11]

12 hours ago, malc-c said:

Guys,

 

It's been four years since I last used my 200P/HEQ5 combo, but with all this lockdown I needed something to stem the boredom, so over the past four / five days I've stripped down and overhauled the scope and took advantage of the clear(ish) evenings to set things up.  The optics were cleaned and re-collimated and the star test at both extremes of focus produced nice circular images with clear evenly spaced concentric rings.

The ST80 was removed and the 9 x 50 finder converted to a finder/guider using the QHY5.  Over the past couple of nights I've spent time checking focus, aligning the finder to the main scope and Celestron star finder pointer so that all three have the target central in the field of view.  The mount was polar aligned using SharpCap's tool to 10 arc seconds.  So that left two things to check out, the first being to gather some guide data, and to generate a PE curve for EQMod to use in its correction.

So having tweaked EQMod's pulse guiding settings and a few PHD2 settings, the calibration ran fine with about 18 steps in either direction E/W and 10 steps N/S - It didn't complain and PHD2 started guiding.  I was pointing at NGC2903 (darks are still running so it will be a while before I get to play with processing) so the weight bar was out fairly level, with a declination of 21.4 degrees.  I've attached the resulting log file trace, and would like some opinions if this is good enough for guiding.  The RMS values of 1.24 arc seconds for RA and 0.59 arc seconds for DEC sound OK, but the graph doesn't look as smooth as some I used to get when using the ST80.

One last question.  There is obviously an underlying sine wave on the RA trace, which could suggest the mount needs PE training.  In order to do so I guess I should fit the camera to the main scope and log the guiding trace through it in order to get better accuracy ?

What imaging camera are you using?

Your RA RMS is approx 2x your dec so that could lead to non-round stars, your PA might actually be too good? 

[malc-c]

Thanks for the comments guys, and masjstovel  I have no objections to you posting the request here.

@jimjam11 Several of my previous attempts to calibrate resulted in messages saying the RA was X percent of the DEC, yet both were set to use x0.40 rates under pulse guide settings in EQMod.  The only setting I changed in PHD2 was to uncheck the option for DEC compensation which resulted in no warnings after calibration.  Looking at the images whilst guiding all the stars are nice and round

I've used PECPrep to generate a PEC file and will load that into EQMod and try a repeat of the guide session if the sky is clear tonight.

I wasn't intentionally out to do an imaging session last night, but took advantage of the fact and ran off 25x 180s subs at 400ISO + 40 darks and here's the resulting image stacked in DSS using default settings - I'm quite happy with my first attempt in nearly four years  

Edited April 11, 2020 by malc-c

[alacant]

1 hour ago, masjstovel said:

how did you PEC train the EQ6?

Hi. VSPEC is part of eqmod. But try phd2 before spending too much time on it

HTH

[masjstovel]

 

@malc-c Thank you! 

3 hours ago, jimjam11 said:

Eqmod has an AutoPec feature. Get everything setup and start guiding on a target, start autopec and it will do everything for you.

 

What do you mean now? didn't you say you did not find EQMOD PEC helpful vs straight guiding PPEC?
If i start AutoPEC while guiding wouldnt it be wrong when PHD2 is correcting every error? Shouldnt it be running unguided? What do i dont understand here?:)

 

@alacant thanx. That was the explanation for a 7 year old. But i truly need for a 5 year old:) How do i do it in PHD2? Yes i have googled it, but cant find the recipe.

[alacant]

3 minutes ago, masjstovel said:

How do i do it in PHD2?

Ok. In PHD2, click on the brain -> algorithms and in the first column, click hysteresis. In the menu which emerges, choose PPEC. That's it.

Come back for some further settings when you've got that going.

HTH.

[malc-c]

I'm no expert, but I think all that happens when you record the PE using EQMOD is that it's effectively logging the corrections PHD2 is making, and then after it has enough cycles it then works out the error and generates its own PEC file, rather than importing a PHD logfile, removing the non harmonics, and then creating the PEC file which is loaded into EQMOD.  

I'm running another guiding session as I type this, having used last nights log files to produce said PEC file.  PEC was off whilst calibrating, but running whilst guiding.  Its early into the session, but so far the traces in PHD seems a lot smoother, and when I left the RA/DEC errors were still below 1px, but that was only after 15 minutes or so.  I'll probably post the results up tomorrow to see how they compare

[malc-c]

9 minutes ago, alacant said:

Ok. In PHD2, click on the brain -> algorithms and in the first column, click hysteresis. In the menu which emerges, choose PPEC. That's it.

Come back for some further settings when you've got that going.

HTH.

 

If I get time tonight I might have a look at that and give that a go (disabling EQMODs PEC) and see how it compares....

[malc-c]

2 minutes ago, malc-c said:

 

If I get time tonight I might have a look at that and give that a go (disabling EQMODs PEC) and see how it compares....

 

Maybe not.... Just remote desktopped into the observatory PC and this was the result...... where did all that come from, it was crystal clear 1o minutes ago !

[michael8554]

The difference in Dec and RA guide rates is worrying, and the Dec and RA RMS difference may be why the stars in your image are elongated.

Could you post that Guidelog ?

Michael

[alacant]

8 hours ago, malc-c said:

uncheck the option for DEC

Make sure it's checked.

8 hours ago, malc-c said:

both were set to use x0.40 rates

Set them both to 0.9 and recalibrate.

[malc-c]

26 minutes ago, michael8554 said:

The difference in Dec and RA guide rates is worrying, and the Dec and RA RMS difference may be why the stars in your image are elongated.

Could you post that Guidelog ?

Michael

Michael, the log file is attached - 

When zooming in on the image the stars looked OK to me

PHD2_GuideLog_2020-04-10_203526.txt

[malc-c]

19 minutes ago, alacant said:

Make sure it's checked.

Set them both to 0.9 and recalibrate.

When checked PHD complained about RA being x% of DEC - 

I noticed that whilst calibrating, the East West would not count up and down correctly... it would say go from steps 1- 16 counting up a digit at a time, but when going the other way it would jump from sat 12 to 7, then 5, then 4, 3, 2.  With the compensator unchecked it would count the steps up and down equally in all directions... unless it was stepping correctly, but just not displaying it.  I also noticed that the N/S steps were always around half the E/W number of steps when calibrating.

 

[michael8554]

1 hour ago, malc-c said:

I noticed that whilst calibrating, the East West would not count up and down correctly... it would say go from steps 1- 16 counting up a digit at a time, but when going the other way it would jump from sat 12 to 7, then 5, then 4, 3, 2.

Normal behaviour, after the East steps it deliberately goes back quickly to the centre, to save time.

Your started your first Cal with a good Guide Rate of 13.5 arcsec/sec, and at Dec 5, but got actual guide rates of RA = 7.5, Dec = 13.8.

That's why Dec Cal'd in a reasonable 12 steps, but RA took 22 steps.

So something is wrong with the RA rate.

With guiding off, does it actually track a star reasonably well in RA, allowing for PE ? 

Is it belt drive with wrong ratios set in RA ?

After that the preset Guide Rate for subsequent Cals goes 7.5, 6, 12, 7.5, 1.5, 4.5, 6 - what's that about ? 13.5 was fine. And at different Dec too, not near Dec 0.

Not that it would have helped until you get the guide rate sorted, but no sign of a Guide Assistant run, which would have suggested the right Min Mo settings etc, but I did see loads of manual changes in the logs.

Star HFD started at 4.31, so your focus could be improved, but worsened as the night went on to 6.4, which is pretty poor - were these Auto Selected stars ?

Might be worth seeing tonight's Guidelog

Michael

 

 

[malc-c]

Thanks for the in depth analysis.  It's given me a fair amount to go through and check.

This was a complete fresh overhaul of the scope after it had been sitting unused in the observatory for almost 4 years.  Optics were removed, cleaned and reinstalled- focuser squared to the tube and the optics collimated with a Hotech laser in self centering adapters.  Star test proved good as shown above.  I did use APT's focus aid, but normally just adjust the Bahtinov mask  pattern by eye.  Focus on the finder guider was done on the moon, so could probably do with adjustment, but the only way to do so is screwing the finders objective in or out so its not as precise, but will see if I can sharpen the focus a little better.

The PC was a fresh install of windows 10 64bit.  I then downloaded the current versions of EQMod, ASCOM, Sharpcap, PHD2 and installed my old copy of APT.  I only use APT for controlling the 400D DSLR which works well, so no need to change.  The latest version of CdC was also installed.

The mount is belt driven (I was messing about with belt conversions in 2011, and effectively proved the concept before Dave from Rowans used his CNC to make the kits we see in the shops).  The ratio is 4:1 and this has been selected in the "driver" set up (but will confirm tomorrow).  PA was done using Sharpcap as shown above.  Once the first target was synced in CdC goto's have been spot on (from park to star two nights in a row).  If the belts were slipping (which they are not) or the ratio was out then Goto would be way out.

I will see how well the mount tracks a star unguided with no PEC, and then enable PEC to see if there is any issue there.

The events shown in the log were my attempts as seeing if there was any dramatic impact on the guide graph having hovered over each item and reading the help text.  Other than that everything was left at default other than entering the details of the focal length of the finder as 181mm based on info found on this forum.  On the last 1hr 23m guide I believe the application selected the guidestar.

The guide camera is an original QHY5 mono - under windows 10 I have to use the ASCOM driver in order to get an image, although the native driver works with QHVideo.

When I installed PHD I simply followed the wizard.  Nothing else was run other than the initial calibration routine.  GA was not run (new to me).  I do remember that in the original PHD under the brain icon there used to be a button that having inputted the focal length of the guide scope one the button was clicked it came up with the thresholds for various things including pulse length etc... that seems to be missing or is now called something else.

Initially both RA and DEV guide rates were set to x0.5 - I tried x0.9, but dropped it down.  I'll reset it for the next session.

I'll grab the log for the last session and upload that, although the cloud cover may have screwed the data ?? and upload it in the morning.  I'll also confirm the EQMOD settings (seems there are now presets rather than having to enter the values manually, but like I said above, if RA was set to the default ratio for the handset goto would be way off.

It may be a few days before I get any further log files given the forecast for the next few days....  

[michael8554]

Sounds like the ratio is correct, so Sidereal should be right too?

And PHD2 is correctly showing the Guide Rates set in EQASCOM.

But check just in case.

Then try a Cal at Dec 0 with RA Guide Rate set to 14 and Dec at 7, see if the Cal steps come out the same !!

Then a Guide Assistant run, accept settings, then do an initial guiding session of 10 mins without changing the settings. then you can have a play !

And post log.

Michael

[michael8554]

Daytime you could run the mount for 60 minutes, check the setting circle for an hour of movement. 

Do you have any Backlash Comp set in the mount or EQASCOM?

They will conflict with PHD2

Michael 

Edited April 12, 2020 by michael8554

[alacant]

8 hours ago, malc-c said:

Star test

Hi. Ah, so that's the main telescope. I wondered what it was. It looks like a (very) out of focus star though.

8 hours ago, malc-c said:

see if I can sharpen the focus

An easy way (don't look at the stars of the screen) is to twist focus whilst looking at the star HFD value. Leave it a few seconds to stabilise after each tweak. Get that value as low as you can (without going overboard).

Cheers and good luck next clear night. It's gotta be close now;)

Edited April 12, 2020 by alacant

[malc-c]

@alacant Yes it is a very out of focused star.  To check collimation running the focuser to the end of its travel in both directions and taking an image will show up any errors in the optical alignment.  If you get a nice round image, that has evenly spaced concentric diffraction rings (as shown) then it means the optics are collimated. Any errors in star shape is thus not down to a mis-aligned optical path.

@michael8554 Thanks for the additional things to look at.  I'll post up last nights log later this afternoon, and report back on the settings in the software.

[alacant]

14 minutes ago, malc-c said:

evenly spaced concentric diffraction rings

Ah. I thought you needed a slightly defocused star at high magnification for that (?). But anyway, don't forget the collimation myths.

HTH

[malc-c]

Last nights log file attached - only thing I changed was the exposure - reduced to 1s to see if the sampling rate made any difference.  PEC was running

I've also attached the PHD2 settings as they were.

Running a 1hr test on RA at the moment - seems to be running fine.

 

PHD2_GuideLog_2020-04-11_205622.txt