I have a TEC 140 refractor on a 1100GTO with encoders on a permanent pier.  I ran a 378 point model, and am getting excellent guiding with SGP and PHD2; mostly +/-  0.25 pixels, but getting very oval stars.  I'm not sure how that can happen, and don't know how to fix it.

Full-size pictures and additional information, including full setup, is at https://eccssw.com/AstroIssues/OvalStars/.

Here's the model results.  I have no idea if this is "good" or not.

Here's guiding with tracking corrections On, then Off.  Even when Off I'm getting pretty good guiding.  Note the scale is +/- 1px.

The right side of this is with guiding OFF.  I believe it shows polar alignment isn't perfect, but it's pretty close.  Note the scale is +/- 4px.
And hear are the oval stars:

How can I be getting very oval stars with extremely good guiding and pretty good polar alignment?

This just started happening recently.

Are you seeing the oval stars across the whole image ? If so, it could be that your imaging train is tilted ? If you are only seeing the oval stars in the corners you may have some field curvature.

Mike

--

Hi Eric,

The right side of this is with guiding OFF. I believe it shows polar alignment isn't perfect, but it's pretty close. Note
the scale is +/- 4px.

According to the pointing model, the mount's polar alignment is over 10 arc-minutes off in elevation and over 20 arc-minutes off in azimuth. Since you say that the polar alignment is now good, it might be that polar alignment shifted or was manually adjusted afterward, which would cause drift when tracking correction is enabled. I think you should improve polar alignment and then redo the model in either case.

Also, for a refractor, the pointing model's RMS is high. It should be closer to 10 arc-seconds for East and West. You might want to check that all fittings are tight and cables are not interfering or tight, which can cause random or unmodeled errors. Lastly, enable Dec Arc Tracking if your license allows for that. This will usually produce better tracking accuracy than the full-sky model does because it accounts for pointing errors outside of the full-sky model's pointing terms.

-Ray

On 4/24/2022 11:29 AM, Brian Valente wrote:

Finally, anytime I hear "great guiding, but oblong stars" my first reaction is differential flexure. Not sure what is your guiding
setup, but if you are guiding on a separate scope, that is definitely
a suspect.

I agree 100%. Any pointing model is made using the main camera only --
no guide camera. Then you can get "excellent guiding" with the guide camera, but that's unrelated to the main imaging train.

Question: are the stars elongated always in the same direction, regardless of the OTA's orientation to the sky? On both sides of the meridian? If not, look for differential flexure.

Here's a write-up of my saga to eliminate differential flexure: <http://astronomy.mdodd.com/flexure.html> It took a while, but I finally eliminated it. Well worth the effort.

Then Jeff wrote:

...when I look closely at some of the dimmer stars, they look like
very tight double stars to me.

Differential flexure again. Something shifted suddenly. With a refractor, you don't have SCT mirror flop, so something else is moving.

How is your guide camera attached to the guide scope? If it's attached to a focuser, check that first. As shown here: <http://astronomy.mdodd.com/flexure.html#Backrest> my guide camera is bolted to a dovetail bracket, and has no connection to the guide scope except a 1-1/4" nosepiece that slips into what's left of the focuser I removed.

(I focus by loosening the bracket and sliding the camera, as described here: <http://astronomy.mdodd.com/observatory.html#GuiderFocus>)

If you're using an off-axis guider, maybe something in that assembly is shifting.

Finally, don't ignore the main imaging train. Things can move there as well as in the guider.

Good luck.

--- Mike

Brian, the web page I linked to has all the hardware.  I have an SBIG STXL 16200 camera with the guider built into the filter wheel, so no differential flexure.
After I posted I checked the two boxes and basically got the same results.

I will be checking for looseness today, and will re-adjust things in case something is tilted.
Eric

On Sun, Apr 24, 2022 at 10:59 AM, Roland Christen wrote:

1) are you sure that you were guiding when you took the image?

2) If you did a model but the scope drifted in RA and Dec it appears that you did not activate the model

3) What is your guiding setup? Off-axis or separate guide scope?

4) We may want you to run a simple test on the CP4 to see what software version and internal settings are. Will check with you tomorrow during office hours.

1) Positive guiding was on - the picture of the oval stars was the first one taken in the "guiding On" part of the PHD2 graph.
2) Both "Pointing Corrections" and "Tracking Corrections" were on the whole time.
3) Per the link I sent, the guide chip is built into the filter wheel on the SBIG STXL-16200, so it's off-axis.

The link also has the .fit file and a screenshot from ASTAP's "CCD Inspector" that shows tilt.

If you are using APCC then after a slew wait about 10 seconds then in the AE tab on APCC disable the encoders.  Wait about a minute and then check to see if the stars are round.  If so, then wait a few minutes and turn the encoders back on, wait a few seconds then test again.  Are the stars round or oval?

Craig

I echo what M Hambrick and Mike Dodd wrote.

A simple way to check that your imaging train is square is to shoot a few very short exposures at very high gain (or at a very high ISO when using a DSLR or mirrorless camera). The short exposure time means there will be minimal tracking corrections from PHD, but at high gain/high ISO you'll get enough exposure of the sensor to record stars. If the stars are oval after only a short exposure time, the problem would have to be with the imaging train not being square.

Good luck and clear skies.

K Walsh

--
Roland Christen
Astro-Physics

Good news - I did the checks people mentioned and all seemed fine so I pointed the scope straight down, then loosened the three knobs between the Quad-TCC and filter wheel, then tightened them (all three a little, then more, then more). and have round stars again.  And ASTAP shows almost no tilt.

Pointing the scope down allows gravity to make sure the two parts are not tilted relative to one another.  The Quad-TCC and filter wheel interface seems to be the one that gets out of whack when I bump into the camera or it hits something, which is probably what happened.

I also did another polar alignment and am within 1-2 arc minutes.  I will be moving the pier in the observatory so didn't bother trying to get better alignment.  This time I turned off "Tracking corrections" before doing the polar alignment - I have the impression having it on throws off the alignment.  I will do another APPM run.  I had "Tracking corrections" OFF last night since I didn't have a chance to do an APPM run.

Thanks all for the suggestions.

Eric's experience with his Quad-TCC raises a good point. What is the best practice for attaching the Quad TCC or any other dovetail attachment (e.g. Doveloc) to the mating dovetail ? I tend to try to locate my Quad TCC in the same position on the filter wheel dovetail every time I attach it. Is doing this creating a "wear spot" where the lock screws are contacting the same spots on the dovetail that over time might result in the parts making less than ideal contact ?

Mike