How can I improve this guiding?

[CraigT82]

I used tonight's lack of clouds to try out guiding for the first time ever.  I recently rebuilt my old EQ6 with the intention of long exposure imaging (I've only ever done lucky imaging up to now - bar a few attempts at widefield with a DSLR on a star adventurer a few years ago).

I didn't have the main scope imaging tonight, waiting on a new secondary mirror from TS. I set up the mount and scope, polar aligned using sharpcap, slewed the mount to M45 and then started up PHD.  Guiding calibration competed successfully and the rig is now guiding with the resulting table below.

Any obvious ways I can improve on this? I think PA suffered as I was stood next to the scope to set it but then once I moved away I think it shifts, due to my weight on the slabs.

Edited February 22, 2021 by CraigT82

[almcl]

Hard to tell from a screen shot (at least that's what the developers over on the PHD2  group say) but did you do a guiding assistant run?  It's always worth doing one at the start and accepting its recommendations.

It appears that your oscillations are down around small fractions of a pixel, so assuming that all the figures were correct when you ran the 'New Profile wizard' that's probably close to the limit of what you can expect.

I wouldn't sweat the PA too much, within 5 arc minutes is said to be fine and any Dec drift will be easily guided out.

If you want an expert opinion, load up the log files to the PHD server as per the help instructions and post the log ref on the PHD2 board https://groups.google.com/g/open-phd-guiding 

One of the devs there will give chapter and verse.

[CraigT82]

Thanks for that info, I will have a look at the PHD2 group and maybe post a log file there. Ive packed away now but tomorrow night looks clearish too so ill have another go hopefully.

I should have said that I did run the guiding assistant and I accepted the recommendations (though I must admit I didn't really have any clue what it was trying to tell me). One thing it did say was that I had tons of dec backlash and might want to consider guiding in one direction only. So actually that's something I can fiddle with for the time being and see if I can improve. 

[almcl]

3 minutes ago, CraigT82 said:

I should have said that I did run the guiding assistant .... One thing it did say was that I had tons of dec backlash and might want to consider guiding in one direction only. 

One recommendation (which may or may not help) is to nudge the mount north with  the controller before running the calibration.   

But I find the GA a bit variable on backlash: some nights it's 'guide in one direction...', others it's backlash compensation of 2000ms +' and others it's 'Backlash is small...'.  I've never figured out a reason and the mount usually hasn't changed apart from being dragged in and out of the shed, of course.

[CraigT82]

13 minutes ago, almcl said:

But I find the GA a bit variable on backlash: some nights it's 'guide in one direction...', others it's backlash compensation of 2000ms +' and others it's 'Backlash is small...'

Thanks that is quite reassuring. Do you run the assistant on every session? 

Edited February 22, 2021 by CraigT82

[almcl]

39 minutes ago, CraigT82 said:

Do you run the assistant on every session? 

Yes, my thought being that it accounts for seeing differences in setting min move.  

[michael8554]

Despite the Dec warnings your Dec guiding is fine, it's the RA error that is the problem, nearly 3 times that of Dec.

Looks like you have enough stars to tick the "Multi Star Guiding" option ?

Your PHD2 settings are burnt out in your image - it's best to take a Screen Grab with the Prt Scrn button on your keyboard.

If your exposure was 1 second, that would be "chasing the Seeing", which would account for the choppy RA result.

Michael

[CraigT82]

Thanks Michael that's very useful info, I completely overlooked the multistar guiding option so will  try that next time out.

Sorry should have mentioned the exposures were 2 seconds for that example. I tried everything between 1s and 3s but the guiding never really got below 0.9" RMS. 

The other values were:

RA agr 70

RA hys 10

RA minmo 0.26

Dec agr 100

Dec minmo 0.40

Scope max RA 2500

Scope max Dec 2500

[michael8554]

Having said RA is your problem, I see the RA guide error is only 0.82 arcsecs,  sounds good for an old EQ6, although I have no experience of the mount.

Perhaps an EQ6 owner can report whether better than 0.82 is a realistic goal ?

Are the stars in your Lights round ?

If elongated in RA you could try a small PA error to force Dec to guide, which could even out the error figures.

Michael

[vlaiv]

Belt mod.

As far as I can tell from the graph - you have fast oscillating component of about 10-15s in RA that is causing most of the issues. I just looked it up - stepper tooth period is 10.2s for EQ6

Either get belt mod, or try to adjust gear meshing in RA axis or maybe look into predictive guiding algorithm in PHD2.

Not sure if it is part of stable release, but I do remember it being in dev / beta test at some point. Algorithm tries to "sense" fast components and then does pre-emptive corrections.

BTW - 0.8" RMS is to be expected for stock EQ6. If you get that oscillation sorted - you'll be able to bring it down to say 0.5" RMS (if both DEC and RA are say 0.32" then total will be 0.45")

[CraigT82]

Thanks Vlaiv. Have had a fiddle with the spur gear meshing and the guiding has come down nicely... Now getting total RMS 0.65".

Still a 10s oscillation in the RA though so perhaps could get better still with a belt mod. 

Edited February 25, 2021 by CraigT82

[vlaiv]

2 minutes ago, CraigT82 said:

Now getting total RMS 0.65".

Notice that you have 0.45" and 0.47" in RA and DEC.

When these two values are so close - I think that is sign that you are seeing limited.

Bumping up exposure length to 4s or so does help in such cases.

4 minutes ago, CraigT82 said:

Still a 10s oscillation in the RA though so perhaps could get better still with a belt mod. 

Belt mod is certainly worthwhile upgrade, it can only help.

I switched to ASCOM driver instead of native for my guiding. That achieves much better SNR values. Not sure what your guiding resolution is (guides cope and camera), but it might help a bit with centroid precision.

[CraigT82]

Well it's guiding nicely all evening apart from the odd gust of wind, but now another issue has become apparent... all the subs show significant trailing in one direction (100% crop of 600s Ha sub below). The horizontal axis of the image is aligned with RA. 

I'm guessing this is differential flexure? I do have an OAG that I shall try for the next clear night.

Edited February 25, 2021 by CraigT82

[vlaiv]

2 minutes ago, CraigT82 said:

I'm guessing this is differential flexure? I do have an OAG that I shall try for the next clear night.

Most likely it is differential flexure.

I think you'll find OAG every bit as easy to use as guide scope - with added precision and lack of problems such as flexure. Only mind the spacing to avoid prism to stop down your aperture for guiding and depending on what guide camera you are using - you can select bin x2 in PHD2 - it will again improve SNR of guide exposures.

 

[CraigT82]

3 minutes ago, vlaiv said:

Most likely it is differential flexure.

I think you'll find OAG every bit as easy to use as guide scope - with added precision and lack of problems such as flexure. Only mind the spacing to avoid prism to stop down your aperture for guiding and depending on what guide camera you are using - you can select bin x2 in PHD2 - it will again improve SNR of guide exposures.

 

Thanks Vlaiv.  I'm using an Altair 290mono as a guide cam, does PHD2 software bin the guiding images on the fly?

[vlaiv]

Just now, CraigT82 said:

Thanks Vlaiv.  I'm using an Altair 290mono as a guide cam, does PHD2 software bin the guiding images on the fly?

Yep, it is under camera settings - noise reduction:

290 has small pixels - 2.9µm I believe? You'll probably want to bin them then. There are few options - 2x2 bin, 3x3 median. I think that 2x2 bin is the best one.

Even binned - you'll have plenty of guide resolution. You probably want your guide resolution to be at max 2"/px.

With that pixel size you only need 300mm of FL and I'm certain your scope has at least twice that if not much more (is that 8" f/6 on EQ6 there?).

[CraigT82]

5 minutes ago, vlaiv said:

is that 8" f/6 on EQ6 there?

Yeah it is. Currently imaging at 0.93"pp with the 383L and guiding at 2.66"pp with 60/225 guide scope.

With OAG guiding would be at 0.5"pp (or 1.0"pp if using bin2)

[vlaiv]

29 minutes ago, CraigT82 said:

With OAG guiding would be at 0.5"pp (or 1.0"pp if using bin2)

I guide with OAG with the same guide resolution - 0.48"/px native and 0.96"/px binned (3.75µm and 1624mm). I regularly use bin x2 without loss of guide precision.

[michael8554]

Vlaiv has pointed you to the Noise Reduction feature.

Binning is here:

[CraigT82]

Managed to get the OAG somewhat focused which took an hour, stars are boomerang shaped in PHD2 but it seems to be guiding around 0.8"pp so not too bad.

Currently shooting 600s Ha bin2x2 (1.86"pp) on Melotte 15 and stars are looking much better than last nights 'worms'.

Thanks for your help in this thread guys 👍

 

[CraigT82]

Just trying out multistar guiding and wowsers what an improvement! Its hovering around 0.40" total RMS but saw as low as 0.31" for a few minutes earlier. I think the seeing and transparency is currently excellent here in Bristol.

Currently shooting 1200s Ha subs at bin1x1 and stars are looking good. 

[CraigT82]

Here is the integrated result from 20x1200s subs at bin1x1 (0.93"pp) over the last two nights, I'm really pleased with how the mount is handling this big old tube, though must say there was very little wind over the last few nights, I think it would be a different story with anything more than the gentlest of breezes.

(Large file - 52mb)