"Star did not move enough" after moving to OAG

[kirkster501]

I am delving into the murky world of OAG and away want to move away from separate guide-scope because I was getting differential flexture on subs longer than about 15 minutes Scope is TEC140 with Atik460/EFW2/Atik OAG.

Got the camera focused and the guide cam focused and seeing stars.  Went very smoothly.  All autofocus re-adjusted.  Confidence sky high! However, I then met with a fall.... I keep getting "star did not move enough" PHD2 messages.  I deleted all previous calibration data from the guidescope and entered the new scope (TEC140) details to work out the step size of 500.  Not working...  I incremented the calibration step size upwards and upwards and upon not working at 6000 I gave up.  My balance is not perfect after some nighttime tinkering I had to do but surely this would not effect it that much?  But then again, maybe it could......  I need to work at this.  I am definitely seeing proper stars through the guidecam and not dead pixels.

Any thoughts please guys?

[daz]

Are the stars in the view definitely moving?

What's the guide camera?

How is the guide camera connected to the mount?

[kens]

Did you try a star cross test?

[kirkster501]

Hi guys, yes, the auto-selected calibration star is moving a bit but not far enough when PHD tries to calibrate.  It only just get out of the green square box after 60 steps and then errors.  When I do a slew new stars appear in the FoV of PHD from the guidecam so they are real stars. I also have a dark library to eliminate hot pixels being mistaken as stars. Previously, when I did this with separate guidescope (before implementing OAGing), the star would move loads WEST and then after 15 or so steps come back east to do the north and south then complete the calibration.

I am using pulse guiding and the guidecam is a QHY5 screwed into the Atik OAG.  The guidecam focus is good.

I realise I need to check balance further and will work on that this afternoon.  I had a bit of a faff when it was dark and had to start moving scopes about so the balance is out.  Would not have thought it would have made that much difference but I do realize that balance can upset many things.

[kirkster501]

13 minutes ago, kens said:

Did you try a star cross test?

^^^ Not sure what you mean Ken?

[almcl]

8 minutes ago, kirkster501 said:

^^^ Not sure what you mean Ken?

Think Ken is referring to this:

https://openphdguiding.org/man-dev/Trouble_shooting.htm#Star_cross_test

Star cross test is described about half way down but the rest may be worth a read, too.

[daz]

4 minutes ago, kirkster501 said:

 Previously, when I did this with separate guidescope (before implementing OAGing), the star would move loads WEST and then after 15 or so steps come back east to do the north and south then complete the calibration.

Yes, but now the image scale is very different.

What settings have you got in EQMOD for the pulse width?

What settings have you got in PHD for the camera and guiding?

[kirkster501]

Hi Daz,

Not using EQMOD - I am on Sitech with a MESU mount.

I have the QHY5 camera pixels and the focal length of the scope entered into the guide wizard that spits out a calibration step of 500 (from memory).

Even if balance and polar alignment are not spot on (they aren't) I'd expect the guidestar to move out of the box.....  So something amiss....

 

 

[RayD]

When I did this and went the other way (OAG to piggy back) I ended up having to make a whole new profile from scratch.

[kirkster501]

4 minutes ago, RayD said:

When I did this and went the other way (OAG to piggy back) I ended up having to make a whole new profile from scratch.

I will look into doing that Ray.

[RayD]

2 minutes ago, kirkster501 said:

I will look into doing that Ray.

I have no idea why I had to do it Steve, it seems it needed to calculate the pixel scale again in a fresh profile, and once I did this it worked fine.  As you know it only takes 5 minutes, so worth a bash.

As noted above, don't forget to make sure your guide scope FL is now that of your main OTA.

[daz]

2 hours ago, kirkster501 said:

Not using EQMOD - I am on Sitech with a MESU mount.

Sorry - must have missed that.

I don't know these mounts, but I'm assuming there will be a similar setting somewhere?

Ray's suggestion is sound though...

[kens]

Te first thing to do is check that you have comms to the mount hence the star cross test. You could also attach your guide log - there's a difference between not moving at all versus not moving enough but the message is the same.

You really should get a decent mount

[kirkster501]

Hi guys, I got PA and balance spot on, same problem.  So I deleted PHD completely, reinstalled it and it is now working a beaut!   I am now an OAG dude

Thanks for your help everyone!

[kirkster501]

Now with OAG I am doing 20 min subs with perfect round stars!!!!!!

[gnomus]

Glad to hear it's working for you Steve.

[daz]

Yay!

[kirkster501]

8 hours ago, gnomus said:

Glad to hear it's working for you Steve.

THanks Steve.  Yes indeed, made a big difference.  With the guidescope, even though mounted as solid as the Rock of Gibraltar, when I pushed  my subs out to 20 mins exposure I started to get eggy stars no matter what I did.  The thing was so solid you could hang an elephant off it; no matter, I got eggy stars with longer subs.

One thing I HAVE noticed since making this jump is a shadow down the right hand side of the sub.  Not at home at the moment to post the sub, and I am sure flats will sort this out, but I wonder if this is the prism in the light cone...?  I will post it later.

[gnomus]

19 minutes ago, kirkster501 said:

THanks Steve.  Yes indeed, made a big difference.  With the guidescope, even though mounted as solid as the Rock of Gibraltar, when I pushed  my subs out to 20 mins exposure I started to get eggy stars no matter what I did.  The thing was so solid you could hang an elephant off it; no matter, I got eggy stars with longer subs.

One thing I HAVE noticed since making this jump is a shadow down the right hand side of the sub.  Not at home at the moment to post the sub, and I am sure flats will sort this out, but I wonder if this is the prism in the light cone...?  I will post it later.

Prism in the light cone is probably right.  You can adjust the prism height - the 'correct' height will, of course, be dependent upon scope/camera combination.  I found the easiest way to judge the correct height was to shoot a series of flats - the prism shadow is obvious then.  It is a bit fiddly, but only needs doing once, of course. 

[ollypenrice]

Good that you're sorted. The shadow could be the prism. When I last used an OAG I set the prism depth while shooting a flats panel, so I could see when the prism began to intrude. Don't forget to orientate the system so that the prism enters the image in the middle of the long side, since this means you can have it deeper into the light cone without appearing in the frame.

The 'round stars' test can by passed by a mount guiding equally badly on both axes! Ideally set the system up in PHD so that it gives you an RMS in arcseconds. The generally accepted rule of thumb (which I can't justify, challenge or confirm...) is that your RMS in arcseconds should be no more than half your imaging pixel scale, so imaging at 1"PP requires an RMS of 0.5 arcsecs. Our Mesus regularly beat 0.3 arcsecs RMS.

Olly

Just crossed with Steve there.

[steppenwolf]

Good news on the guiding - whenever I have a problem like this, it's user error (me!) first then cables followed by software, rarely is it the real 'hardware'.

[steppenwolf]

5 minutes ago, ollypenrice said:

The generally accepted rule of thumb (which I can't justify, challenge or confirm...) is that your RMS in arcseconds should be no more than half your imaging pixel scale, so imaging at 1"PP requires an RMS of 0.5 arcsecs. Our Mesus regularly beat 0.3 arcsecs RMS.

I concur :-)

[kirkster501]

18 minutes ago, ollypenrice said:

Good that you're sorted. The shadow could be the prism. When I last used an OAG I set the prism depth while shooting a flats panel, so I could see when the prism began to intrude. Don't forget to orientate the system so that the prism enters the image in the middle of the long side, since this means you can have it deeper into the light cone without appearing in the frame.

The 'round stars' test can by passed by a mount guiding equally badly on both axes! Ideally set the system up in PHD so that it gives you an RMS in arcseconds. The generally accepted rule of thumb (which I can't justify, challenge or confirm...) is that your RMS in arcseconds should be no more than half your imaging pixel scale, so imaging at 1"PP requires an RMS of 0.5 arcsecs. Our Mesus regularly beat 0.3 arcsecs RMS.

Olly

Just crossed with Steve there.

Can you repeat that last paragraph in English please Olly?!

[ollypenrice]

8 minutes ago, kirkster501 said:

Can you repeat that last paragraph in English please Olly?!

Bien sûr! Si les erreurs sont comparable pour les deux axes... 

The first part just points out that if your mount has comparable and quite high errors in both RA and Dec you will get round stars because the errors will stretch the stars equally in both directions. Indeed we had exactly this when setting up the first Mesu on night one. The stars were round but quite big and soft. Refining the guide parameters made them smaller and brighter.

The second part is based on the fact that PHD allows you to inform it of your guider's focal length and pixel size. The FL is that of the main scope with an OAG and you can look up the camera's pixel size. By default the trace in PHD is shown in pixels but if you give it the FL and pixel size it will give you the trace information in arcseconds.  You can then directly compare the errors with the pixel scale of your imaging camera in the same units, arcseconds. Your 460 in the TEC is, like mine, working at about 0.9"PP which means your guiding will be effectively perfect provided your RMS value in arcseconds is 0.45 or better. If your Mesu is anything like the ones we have here you'll match or beat this requirement. 

(OK, PHD is measuring the difference between the mount's pointing position and the position of the guide star's image. The guidestar's image may not be exactly where the guide star really is because of the seeing, but if you use longish guide subs - we use 4 seconds - the seeing will average out. There is nothing more you can do on an autoguided mount.)

Olly

 

[kirkster501]

Here is the stretch image of a raw 20 minute sub pointing at Deneb area (strange artefacts around Deneb despite anti-blooming camera).  This is quite a result for me since I could never do  20 minutes before.  Note the band down the right hand side of the sub.  I did not have this when I was using guidescope so need to tinker with the prism a bit.  However, if flats remove it then I may not bother.  I have not done any flats as yet since I have gone to OAG.  This is the first (and only so far) sub that I have done.