Diagnostic Help: Progressive Frame Drift Despite Active Autoguiding

Mirror shift and/or differential flexure (if the guide scope is poorly mounted). I would also question the wisdom of using a focal reducer on a newton as your star shapes testify.

Matteo Luongo · Jun 17, 2026, 08:15 AM

Merci pour votre retour.

J’utilise pourtant ce réducteur depuis plusieurs mois et je n’avais jusqu’à présent pas constaté ce type de problème sur mes précédentes sessions. C’est pour cela que je m’interroge davantage sur un éventuel problème mécanique ou de guidage.

Concernant le décalage du miroir, pensez-vous qu’il pourrait également expliquer les pics récurrents en DEC observés dans PHD2 ainsi que la dérive progressive de la cible dans le champ malgré l’autoguidage actif ?

Je vais également vérifier le serrage du primaire, du secondaire ainsi que l’ensemble du train optique pour écarter cette piste.

I have looked at your images and they all show issues with the stars despite the obvious processing applied to them, but this has nothing to do with the present issue, clearly.

I don’t think the backlash issue in Dec has much or anything to do with the image drift you're experiencing although it does not help with the quality of the images either.

Excluding a case of really badly serrated secondary I don’t think the secondary is the culprit here but I’d check that the secondary spider vanes are tight and there is no wobble in the secondary mirror support at the touch.

The issue is most likely with the primary mirror in its cell. If anything like the one I have they sit on 3 pads (usually cork pads) and are restrained axially and circumferentially by 3 pegs. The gaps between the pegs and the mirror are the cause of the shift but they cannot be closed down arbitrarily as this will cause mirror pinch. Radially, you need no more than 0.015mm of gap to account for differential thermal expansion between mirror and cell and axially something of the order of 0.01mm. If they (the upper section of the pegs) are rubber made then you can safely put them in contact with the surface of the mirror but do not over-tighten them. Just enough to close the gap and no more, but leaving the 0.01mm gap is the safer approach. Use brass shims to close the radial gap (not the axial!) between the mirror and the pegs choosing the right thickness as they usually sell them (brass foils) with a number of different thicknesses.

Matteo Luongo · Jun 17, 2026, 09:41 AM

Merci pour cette explication détaillée.

Une chose m’interpelle toutefois : la nuit précédente, avec exactement le même matériel et sans avoir déplacé la monture entre les deux sessions, je n’avais pas ce problème de dérive du champ. Le principal changement a été le passage d’ASIAIR à NINA/PHD2.

Pensez-vous qu’un simple changement de logiciel puisse provoquer ce type de dérive progressive, même après une mise en station correcte, un centrage réussi de la cible et tout en refaisant un calibrage PHD2 vers le Sud au plus bas que je peux. ?

Dans tous les cas, je vais également vérifier le serrage du miroir primaire, du secondaire ainsi que l’ensemble des éléments mécaniques pour écarter cette piste.

Merci encore pour votre aide.

I have 3 rigs running NINA+PHD2 and never do more than polar align them every now and then. Sometimes I do not even bother with calibration at the meridian either. Never had anything as significant as you are experiencing but image drift is present in all my reflectors I’d say by design. Some locations are much worse than others (I can’t shot anything in Leo-Virgo for more than 60s because of that) so if your previous session was on a different object at different declination to start with, then the issue may not even be apparent at all. I would make certain that PHD2 is set up correctly to act as a server to NINA but again this is pretty automatic nowadays.

andrea tasselli · Jun 17, 2026 at 10:38 AM

And chose Green Swamp Server instead of EQMOD to be on more sure footing.

That’s exactly what puzzles me. It’s the same target, captured over the same time period and with the same equipment.

During the first night with ASIAIR, the sky conditions were actually quite mediocre, yet only about 5 out of 100 frames had to be rejected during stacking.

Last night, using NINA/PHD2, under better conditions, PixInsight only kept around 20 frames out of roughly 100 captured.

This significant difference between the two sessions makes me wonder whether the software change or some configuration difference could have played a role in the issue.

have you created a dark library in phd2?

also how confident are you in your polar alignment accuracy?

lastly… and this suggestion is definitely not best practice as you want to find and fix the cause of the issue ideally, but… consider running center on target in nina more often.

best to find and fix the issue though :(

TiffsAndAstro · Jun 17, 2026, 01:45 PM

Avez-vous créé une bibliothèque sombre dans PhD2?

aussi, quelle confiance accordez-vous à la précision de votre alignement polaire?

enfin... et cette suggestion n'est certainement pas la meilleure pratique car vous voulez idéalement trouver et résoudre la cause du problème, mais... envisagez d'exécuter le centre sur la cible dans NINA plus souvent.

mieux vaut trouver et résoudre le problème, cependant :(

Yes, I did create a dark library in PHD2.

Regarding the polar alignment, that's actually something that caught my attention. The first result I got with NINA was completely unrealistic, showing nearly 40° of error. I assumed something had gone wrong during the procedure, so I restarted the entire polar alignment process and only proceeded once I obtained a reasonable result.

With ASIAIR, on the previous night, I had achieved around 30" of polar alignment error using the same equipment and observing location.

That's why I'm struggling to understand the large field drift. After redoing the polar alignment and performing a fresh PHD2 calibration, the guiding appeared to be working correctly, yet the target still drifted significantly over the course of the session.

I'll run another session with ASIAIR to compare the behavior under the same conditions.

andrea tasselli · Jun 17, 2026, 02:26 PM

Matteo Luongo · Jun 17, 2026, 02:17 PM

andrea tasselli · Jun 17, 2026, 11:40 AM

À quoi ressemble le journal de PHD2 ? Quelle est l'erreur de PA mesurée? Je connais plusieurs cas où PHD2 guidait sur des pixels chauds au lieu d'étoiles, avec des résultats similaires. Vous pouvez vérifier si le drift s'aligne avec la direction verticale ou non pour exclure cela a posteriori. Utilisez-vous les options multi-étoiles dans PHD2? N'oubliez pas que vous devez configurer PHD2 vous-même pour obtenir les meilleures performances, y compris le guidage prédictif.

I checked the PHD2 log and generated the graph in PHD2 Log Viewer. Multi-star guiding was enabled, so it was definitely guiding on a real star and not a hot pixel.

I've attached a screenshot of the graph, and I can also provide the full GuideLog file if that would help with the diagnosis.

Does anything stand out to you that could explain the progressive field drift?

📷 image.png

PHD2_GuideLog_2026-06-16_230403.txt

Your guiding errors are HUGE!

But I don't understand how I had this kind of curve before and then ended up with the other one.

📷 image.png

andrea tasselli · Jun 17, 2026 at 02:35 PM

Have you checked that the image scale in PHD2 matches the focal length and pixel size of the guider? Obviously something is fairly wrong here in the way PHD2 operates with EQMOD (if at all).

I checked the PHD2 logs and noticed that the guide scope focal length was set to 160 mm, while the actual guide scope is specified at around 175–180 mm. The camera is an ASI120MM Mini with 3.75 µm pixels.

I’ll verify the profile settings and make sure everything matches the actual hardware before the next session.

Do you think such a discrepancy could be enough to cause this kind of field drift, or would you expect a different issue between PHD2 and EQMOD?

andrea tasselli · Jun 17, 2026 at 02:47 PM

Effectuez une résolution astrométrique sur le guideur pour vérifier la longueur focale réelle et l'échelle de l'image en premier, car un décalage a des implications sérieuses sur la précision globale du guidage. Je retirerais EQMOD de l'équation entièrement (avec NINA) et j'utiliserais GSS dès le début. Et oui, avec 5 heures d'exposition, une erreur accumulée de ce type ne peut être exclue. Vous avez également de très grandes pointes qui devraient être éliminées en modifiant la monture pour un arrangement de courroies et de poulies sans jeu et sans cogging.

Thank you for your help and suggestions.

I’ll check the guide scope image scale, verify the PHD2 profile settings, and run another test session tonight. I’ll also compare the results with ASIAIR using the same equipment and target.

I’ll report back with the results once I have more data. Hopefully that will help narrow down whether this is a software/configuration issue or something mechanical.

Thanks again for taking the time to look into this.

Because it doesn’t make a lot of sense since one run didn’t have the same errors of the other one and nothing changed between the two except the software.

Do you have the PHD dithering checkbox set to random?

Advanced Settings/Global/Dither Settings.

Make sure it’s set to Random and not RA Only.

Also, go back to the ASI Air for a night and see if the problem goes away. If not, the problem is not any settings in NINA/PHD2.

bonsoir

j’ai un 200/800 newton monté avec un starizona 0.75 donc f/3 et j ai l’impression que vous avec le même genre de config.

je dirais par mon vécu. mise au point pas top et un seeing affreux ce qui est le cas en ce moment selon où vous êtes.

ca provoque un guidage approximatif et donc des étoiles bizarre quant on a une mise au point dans les choux