I Think I Finally Understand My PHD2 Graph!

[StuartJPP]

You know when you have one of those eureka moments? Well I had one not so long ago and I think I finally get it now...I actually feel embarrassed that I didn't quite get it sooner (and will feel even more so if I have still got it wrong!). The reason I started getting stuck into this is that I was having slight elongation in my stars, but I have subsequently stiffened up my kit and reduced differential flexure quite a bit...but here goes anyway.

I commented previously that my PHD2 graph was pretty mountainous, but doing all the calculations on the focal length of my guide scope and the pixel size of the camera, it appears as though it is all just in the noise. The image in the guide scope is deviating on average by +/-0.5 pixels so not really jumping around as much as I had originally calculated/thought.

In the screen capture of a PHD2 session below, things look pretty rocky but the thing I didn't actually realise was that the pixel scale of the guide scope/camera is ~3.87 arcseconds per pixel (200mm FL and 3.75um pixels) so of course it isn't going to be anywhere near flat at this scale of +/- 3".

 

 

In this screen capture of PHD2 Log Viewer the guide star deviates by 4 arcseconds peak to peak, (I selected this portion as it is flanked by gusts of wind). It basically shows the same image as above, just more compressed in time but what the graph generally looked like except for those said gusts of wind.

 

And in this final screen capture, which is the same as the above except that the data is plotted in pixels rather than arcseconds. Which shows a deviation of ~1 pixel peak to peak.

 

What does this tell me? Well I think that this session was pretty good despite what the graph looks like as in reality it is as good as it can get with the guiding setup I am using. My 10 minute subs were pretty spot on (which is what counts) with only a small bit of drift over 30 minutes (~1 pixel) which I put down to differential flex which I am working on.

So is this complete gibberish? Am I understanding the numbers correctly now? Things aren't actually that bad and besides, the proof is in the final image.

I hope to put this one to bed now! Unless someone tells me I have still got it all wrong!

[Pompey Monkey]

That looks fine and dandy to me. The only thing I'd suggest is to up the guiding interval to 2 seconds or more - you might be chasing the seeing a bit, especially at lower elevations.

I was about to say that I would hope for a bit better than 1.24" rms from £2.5k mount (it is the Avolon M zero isn't it?). I get the same performance from my AZ-EQ6, and I'm not particularly happy with that at the moment. BUT the Avalon only weighs in at 5.5 kg, so that number is actually pretty impressive!  

 

 

[graemlourens]

Hey Stuart.

Sounds good to me too. I'd also go for higher exposure length in general, but as you are at a very low focal length i'm not sure if this will really help.  Just see that you don't guide on a saturated star (flat star curve at the top)

If you look at your Total RMS value (this is for me the thing i look at most) you will see that its around 0.32 pixels what is GREAT in my opinion. I think most of us are happy if we are at total RMS of around 0.5 - 0.8 pixels. Guiding at lower focal lengths is always easier, as seeing gets more critical the longer the focal length gets (talk to somebody guiding at 2000 and up mm, they will have long stories about how long it took them to get a steady graph )

Kind regards, Graem

[StuartJPP]

Thanks guys,

I'll take on board your comments. Regarding the imaging exposure length, I usually stop down my guide lens so that my exposure length is ~2 seconds, but on this occasion I was lazy and left it wide open (f/2.8) so even dim stars were being saturated at 2 seconds so opted for 1...

Luckily I don't intend on imaging at long focal lengths with this mount...so I should be okay...as I said before, if it wasn't for my flex issues I would never have looked into the graphs as much as I did...

 

On 27/01/2016 at 19:33, Pompey Monkey said:

That looks fine and dandy to me. The only thing I'd suggest is to up the guiding interval to 2 seconds or more - you might be chasing the seeing a bit, especially at lower elevations.

I was about to say that I would hope for a bit better than 1.24" rms from £2.5k mount (it is the Avolon M zero isn't it?). I get the same performance from my AZ-EQ6, and I'm not particularly happy with that at the moment. BUT the Avalon only weighs in at 5.5 kg, so that number is actually pretty impressive!

Yes it is the Avalon M-Zero...oh and the counterweight is 0.5Kg and the counterweight bar about 100g...that is with circa 7Kg worth of lens and camera attached.

 

 

 

[swag72]

Oh the joys of a guide graph ............. is it our best friend or our enemy? I can never quite decide! I try to ignore it if I can but sometimes I've just got to look at the graph instead of the resulting image ...... Damn!!!!  All looks fine to me

[ollypenrice]

I believe that you have to look at the guide graph because the 'round stars' test is bunk. If errors are random stars will be round. But how big are they? To put it another way, oval stars prove bad guiding but round stars do not prove good guiding, even if they are better than oval ones - which we'll all agree they are!

My Avalon runs at about 0.05 to 0.07 pixel error when the guide setup is at about 8.5 "PP. That means an error of around 0.5 to 0.6 ". The TEC is imaging at 1.8"PP so I feel I should be OK.

As for sub length, you'll chase the seeing less with longer subs and you'll get a better graph, potentially, with shorter ones. Remember that the graph measures the position of the star after it has been through the seeing, not the true position of the star. So with short subs a good graph can be a good graph relative to a false star position. For all that, and contrary to what Avalon recommend, I find my own mount does not thrive on long subs. I use 1 sec guide intervals. On the stiffer and more accurate Mesu I use 4 sec intervals. I reckon you have to treat every mount as an individual.

I also think, as do quite a few other users, that the Avalon's results are better than its guide trace might suggest because what the guide trace does not record is the speed with which a correction is applied. The only hard information on the guide trace is the series of positions from which each successive command is sent. What happens in between is just a computer generated smooth curve to join the dots. The computer has no information about where the mount really is between corrections.Because there is no backlash I suspect the Avolon gets back on target faster after a command than a backlash-affected mount.

Olly

[Ken82]

Looking good to me  it would be interesting to see your calibration data ? Could you post the PhD log ? 

[StuartJPP]

On 03/02/2016 at 23:07, Ken82 said:

Looking good to me  it would be interesting to see your calibration data ? Could you post the PhD log ? 

Should be here (if uploading a .txt file works):

PHD2_GuideLog_2016-01-22_210241.txt

Ignore the 1st two guiding sessions as well as the 1st two calibrations, I was messing about...

 

Here's a screen capture if the above doesn't work. Don't know what happened to the rogue W8 step. After this I reduced the window search area back to 15 (I think this was at 25) as well as upped the guide rate to 0.75x.