Meridian cross over guider question


#1

This is a little bit of a teaser. I have a Paramount MX whose standard “home” position points to the west side of the meridian. I chose to calibrate PHD2 at this point and then SGP sent it over to the East to center the object for the imaging run. After centering, the RA went off the scale as the guider polarity made the RA error worse.

I have not done a true meridian flip for a while but is this expected behavior? Does the SGP/PHD2 partnership only flip the RA polarity during the meridian flip operation or should it occur any time the PHD calibration is on one side and the mount is guiding on the other? If it does not - that would suggest one must always calibrate PHD on the same side of the meridian as the imaging sequence start point?


#2

Can you post the logs?


#3

I don’t quite understand what you are saying. Bottom line it’s better to start PHD calibration AFTER slewing and centering DSO.

Peter


#4

buzz,
If you have an ASCOM connection to the mount PHD2 will know which site of the meridian you are on relative to where you calibrated and will adjust the calibration accordingly. Make sure you have enabled Pier side reporting in the ASCOM driver settings.
Try View -> Display Stats and check what it says for Pier Side.
Without an ASCOM connection, PHD relies on SGP to tell it about the pier flip; if the pier flip did not happen during a sequence, SGP has no way of knowing that the flip happened after you calibrated.
Andy


#5

Thanks folks. All my settings are correct (I’m using TSX and I have the side of pier enabled in the ASCOM profile). PHD2 is working via ST4 and also has the ASCOM connection too to pick up info from the mount.

I think I have worked it out. I calibrated PHD near the celestial equator and pointing West. I started the sequence and told it to slew and center (to the East).
When PHD picked up again, the RA guider polarity was in the wrong direction.

I recalibrated PHD near the target and much later last night, the meridian flip was automatically carried out without a hitch. (I made a screen video to wind up my neighbor who uses Maxim DL)

So - it would seem that in essence PHD2 itself is not working out what side of pier is set but SGP informs PHD2 when it flips. I will check with the PHD guys to see if that is as it should be.

(By the way - I always try and calibrate any autoguider near the equator - it gives a more reliable result and if I had my way, would turn off RA / DEC compensation too. The system becomes too sensitive to seeing noise otherwise at high declinations.)


#6

Buzz, you may need to check the option to reverse ra on meridian flip in
the brain. Some mounts require that.
Andy


#7

Andy - according to the PHD2 help file, that button is a manual override. If I calibrated pointing west and started a sequence in the East, I could press it. This is not the entire solution however:

  1. If it is only a manual override, AFAIK it does not instruct PHD to look out for pier side and work out when to reverse polarity for itself.
  2. Not sure what would happen to the SGP meridian flip feature afterwards - would it flip again?

I guess I need to try it out and see what happens.


#8

Have you tried PHD calibrating on a guide star when your target DSO is already centered in FOV of your main camera and then run the sequence? If you are using ASCOM guiding, then Meridian Flip should not be an issue. I think running PHD calibrating where the target DSO is at would be better. That’s how I do it and works very well.

Peter


#9

Peter - yes, thanks, I have in the past and that would avoid this initial instance of the flip issue - but potentially has two other issues: With PHD(1) that did not give good guiding results when calibrated at high DEC - and I preferred to calibrate closer to DEC=0. (It also took an age to calibrate too). This is hinted in the original PHD instructions too and several papers on guiding best practice.

With PHD2 - the calibration time issue remains - but I’m not so clear about whether PHD2, like Maxim DL, normalizes the calibration figures, taking into account the DEC of calibration and guiding. With SGP being so effective at meridian flips, its worth fully understanding the behavior of the SGP-PHD2 relationship for start of sequence events too.

I have three questions then to resolve:

  1. Does the manual override instruct PHD to look out for pier side and work out when to reverse polarity for itself?
  2. Does 1) affect SGP meridian flip feature afterwards - will it flip again?
  3. Does PHD2, like Maxim DL, normalize the calibration figures back to DEC=0, taking into account the DEC of the calibration and the DEC of the guiding

Chris


#10

With PHD2 - the calibration time issue remains - but I’m not so clear about whether PHD2, like Maxim DL, normalizes the calibration figures, taking into account the DEC of calibration and guiding. With SGP being so effective at meridian flips, its worth fully understanding the behavior of the SGP-PHD2 relationship for start of sequence events too.
PHD2 takes into account declination based on where you calibrated and applies a bias. I don’t know the math. You might post over at the PHD2 Google Group if you really want to know what they’re doing exactly. They’re very active.

SGP has little input into what PHD2 is doing. It basically issues commands to the server and waits for a response back from the server. It used to do much more before they improved their socket communications; however once Andy/Jared/Ken started collaborating, things improved dramatically from there to the point that PHD2 is doing all of the work. If you’d like to see more about it, go to the Yahoo Group and do a search for PHD2. There are quite a few threads on it.

1) Does the manual override instruct PHD to look out for pier side and work out when to reverse polarity for itself?
Andy would be best to answer this. You might PM him.
2) Does 1) affect SGP meridian flip feature afterwards - will it flip again?
I flip back and forth every night, no issues with PHD2 guiding. It uses SideOfPier to figure out when to flip the data and uses the connection inside of PHD2 to the mount to figure that out.
3) Does PHD2, like Maxim DL, normalize the calibration figures back to DEC=0, taking into account the DEC of the calibration and the DEC of the guiding
PHD2 does take into account declination. How and why were spelled out in the PHD2 Google Group. I’d head over there and ask if you want the nitty gritty.

I don’t have issues calibrating PHD2 like I sometimes did in PHD1. I also find that PHD2 isn’t sensitive to where you calibrate as long as you’ve done it.


#11

Buzz,

I do not have issues with PHD2 calibrating and guiding at high Dec as compared to Celestial Equator (Dec = 0). I rely on PHD2 calculating for the best Calibration Step (msec). You fill in relevant information including target DSO’s Declination and returns the best result for Calibration Step (msec). No more guessing like in PHD1.

Peter


#12

It depends on how you’re guiding. If you’re guiding via ST-4 then you will likely run into problems if you manually flip the calibration data outside of SGP’s knowledge. However if your calibration is correct at the time SGP initiates a slew that results in a meridian flip SGP will flip your calibration data. So if you calibration data is valid then SGP will keep it valid. However if it is incorrect then SGP will actually keep it incorrect. Again this only applies if you’re guiding through ST-4.

If you’re guiding through ASCOM, with PHD2, then none of the above applies and your calibration is completely managed through PHD2. SGP may send a flip calibration command but I believe it gets ignored.

Unless your mount doesn’t support it I can’t think of many reasons to use ST-4 these days (not saying your are…just stating this). Pulse guiding can supply much better information to your mount for more rapid corrections than ST-4.

Jared


#13

Many thanks Jared, Mads. I am using ST4 as it happens on the Paramount MX. I did try pulse guiding once and it went wild after a meridian flip. I’ll will give it a try again and see what happens.

Topboxman - have you seen the following article? http://acp.dc3.com/McMillanAutoguiding11-2005.pdf

It gives food for thought on guiding algorithms at high declination. Normally the algorithm for DEC compensation is very straightforward. The DEC=0 coefficient is divided by cos(DEC). Working on the basis that seeing noise outstrips tracking error in the steady state at low declinations, this paper reasons that at high declinations, the RA error is so small in pixel terms that the seeing noise is an order of magnitude higher and it makes no sense to react to seeing noise with too much gain in the autoguider system. For example, at DEC=0°, a pixel error of 0.1 pixels may translate to a guider movement of 0.3 arc seconds but at DEC=80°, after compensation, it translates to: 0.3 /cosine(80°) or 1.72 arc seconds. There are two strategies - lengthen the integration time to make it less sensitive to seeing noise or keep the RA guider gain the same as for lower declinations.

I did this video for my friend who is working on a script to do meridian flips with the gold standard. I think I have saved another soul :smile:

I did not realise there was a dedicated PHD2 group and will go and join them. Clear skies.


#14

Thanks for the link. Since we both have high end mounts (mine is A-P1100GTO) meaning smooth tracking and low PE (high mount guiding SNR as the document stated), longer guide exposures to increase guiding SNR helps improve guiding performance. In the document you provided stated this:

“If the mount is very good with low PE and excellent polar alignment, it implies a low Guiding S/N ratio and will probably benefit from using longer guiding exposures to improve the Guiding S/N ratio.”

My typical guide exposure is 3 to 6 seconds and has worked pretty well.

I used to have CPC0800 and it had horribly high Dec backlash and PE (low mount guiding SNR). My typical guide exposure was less than 1 second and pretty often as low as 0.2 second. I also set Max Dec backlash to as high as 1500msec due to mount’s excessive Dec backlash. So what I have done with CPC0800 pretty much follows the document’s guidelines.

Thanks again for the link to the document.

Peter


#15

That was probably related to this setting in PHD2:

You have to experiment and see whether or not you need the option checked or not. You may need to change the setting when you switch between ST-4 and ASCOM pulse guiding. (You can use separate equipment profiles, one for On-camera and one for ASCOM.)

Andy


#16

Andy - I’m right in the middle of the experiment and it looks like ST4 does not need this box checked but pulse guiding through TSX and the Paramount does. If I leave the box unchecked, the RA is OK but the DEC goes off the page when I flip using pulse guide. Of course, it only flips polarity when I flip sides, so checking the box afterwards does not force a flip it seems.

update - the meridian flip worked using pulse guide and the DEC reverse checked - I posted on the PHD forum and I’m now doing some PHD2 experiments to see what conditions cause the occasional instability with a new exposure.


www.mainsequencesoftware.com