Dither Settling Params

[s3igell]

I've been successfully using BYE and PHD2 in concert for Guiding and Dithering since early BYE v2 and PHDv1.12, and with BYE v3.12 and PHD2 v2.4 I've successfully imaged and Guided two separate Mounts simultaneously, so I know that it all DOES WORK.

 

But, I've been using the recently released PHDLogView App to review many of these successful Sessions, and I'm perplexed...

 

According to the Dither setting for SettleAt (defaulted to 0.25), many of my BYE Sessions would have NEVER resumed Imaging after most of the Dithers as per the PHDLogViewer (using the Settings/Exclude Settling Parametrically and the Exclude Frames Settling settings that match those used in BYE) the Excluded-for-Settling portion often runs for several minutes at "0.25 pixels".

 

So, that leads me to ask:  Does BYE implement the two components of "Settle" sequentially (as indicated in the "Debunking Dithering" HowTo)  or are they evaluated simultaneously ??  All my experience sitting in front of BYE for whole Nights at at time say they are evaluated sequentially.  

But then, the PHDLogViewer App was written by one of the lead OpenPHD Devs, and we'd assume that it is also correctly interpretting the evidence in the Log Files.  And that indicates otherwise...

 

So, can you confirm which PHD API Property/Value is being evaluated for the Settle Distance ??

 

(I've posted an equivalent question on the OpenPHD forum in this Thread... )

[PeterCPC]

Hi

I have recently noticed that with the default dither settings that I am getting some images with double stars showing when using my 1200ED. What dither settings would you recommend for using a SW1200ED on an AVX mount? I have altered them to aggressiveness 1, settle at 0.15, calm down 10 and dither every 2 images - does that sound reasonable? This is all a bit of a mystery to me.

 

Peter

[s3igell]

Dithering causes a "Quick Shift" in the Mount's Drive Gear Loading - potentially triggering Backlash Issues in one or both Axes.

 

This can cause Oblong or even Double Stars as you gather Light for the next Exposure while your Drive Motors are still working to take-up any Backlash.

 

If you are getting "Doubled Stars" in your Exposures, then you may need to adjust the Calm-Down setting - extend the time to better accommodate your Mount's reaction to the Dither.

[admin]

I would start by increasing the calm-down period as suggested, try 30 seconds as a test then go down from there.  I suspect your sweet spot will be around 20-30 seconds.  This behavior is due to the mount not having sufficient time to recover to a stable guiding 'momentum' before the next picture is taken.  The calm down parameter is meant to extend that dead time so your mount can recover from the dither process.

[PeterCPC]

Thanks, I'll give that a try. What about the other settings? Should I go back to default?

 

Peter

[admin]

Leave the aggressiveness at 1 for now and only play with calm down.  You may want to increase the settle at to .20 or 0.25 if your imaging < 1000mm focal length... but only after you have found your sweet spot with the calm down parameter.  Only change 1 parameter between tests so you know exactly what is the impact.

[kenklein]

Using 3.5.1.  To turn on  dither, do I want the "dither" toggle to be dark or light?  Thanks.  

[admin]

Using 3.5.1.  To turn on  dither, do I want the "dither" toggle to be dark or light?  Thanks.  

 

You want it to be lit = on

[kenklein]

Thanks very much.  

[admin]

There is only one and only parameter in PHD to query the distance, that value is displayed in the progress wheel and has a value of 2.55 to 0.00.  The settled distance is confirmed when that number drops below the settled value of 0.25 in your example... after which BYE time will start the calm down period (in seconds), then resume imaging.

 

However PHD continues to race to the new intended position, that is what it does while guiding all the time and the same applies during and after dithering.  If you feel you need more settle time you can decrease the settle at to 0.20 and increase the calm period to 20 seconds and see if it does improve.

 

BYE has a exit loop strategy however, it will automatically stop dithering after 90 seconds.  This exit strategy was to mitigate a bug in PHD 1 several years ago where PHD would simply freeze and the calling app (BYE in this case) would wait indefinably for PHD to complete the dither process.  Maybe this is no longer needed with PHD2?

 

Hope this helps,

[s3igell]

I'm not having any troubles managing the Dithering, nor the Recovery, (although I do need to use a 60sec Calm Down for one of my less-responsive Mount/Scope Guiding Configs).

 

I'm looking to tie-in the Graphic Display / Analysis of the Guiding Logs, such that I can review :What Happened..." during those All-Night Imaging Sessions where I turn-in for a few hours of Warm-Up or Snooze.  And that is currently a bit Difficult, as I cannot always correlate the Guide Log to the BYE Log because the Dither SettleAt Parameter isn't readily matching to the Guide Log Analysis.

[admin]

Could it be that for one given guiding interval during dithering that PHP reported a value less that 0.25.  As soon as a value is less than 0.25 BYE considers the dithering complete.

 

Maybe a better implementation would be to have at least 5 seconds of consistent value < 0.25 to consider dithering complete, this would reject the out-liars for sure.

 

Regards,

[s3igell]

Maybe a better implementation would be to have at lease 5 seconds of consistent value < 0.25 to consider dithering complete, this would reject the out-liars for sure.

That is much more in line with OpenPHD's Dev Andy Galasso's interpretation used in PHDLogView and the PHD2 Server API (PHD2 itself makes no interpretation in the PHD1 mode):

 

 

Per Andy Galasso: The way I did it in PHDLogView is: guiding is settled when the distance remains below M pixels for N seconds.

 

If such an interpretation of "SettleAt" were used, we'd be more confident that the Guiding had indeed settled down after a Dither and not simply reporting a low value as the Guiding crossed the Axis while still attempting to overcome Backlash (the common "ringing" or "sawtooth" that less-than-perfect Mounts produce after a Dither).

 

If such a change were made, many of us users with less-than-perfect Mounts would likely need to change from parameters of "0.25 and 20" to something like "0.75 and 60".  Although nothing would be changed about the actual Guiding, our Images might well benefit from the assurance that we haven't started the Next Exposure before the Guiding has truly Settled Down.  And, it would be more in line with the Implementation OpenPHD has taken for the PHD2 Server API - if you ever consider moving to that API...