Autoguiding, drift, diff-flex and all that.

[themos]

Hello,

When we are guiding with a side-by-side rig we expect some differential flexure which can build up, sub after sub, and make our subs gradually drift in a certain direction.

When we are using an OAG, should we expect some drift? I thought not but the numbers tell another story.

I got some numbers from a rig with a triplet wearing an OAG and a doublet just following the mount on guidescope rings. I coaxed the numbers into charts on Google Docs and here they are

The first sheet has a chart with the subs drift (as calculated by DeepSkyStacker) of the triplet-OAG. The second one shows the same thing for the doublet that was following. I expected differential flexure there, so no surprises. But how can we explain the drift in the triplet subs? And why is it in one direction only, the dy?

[Merlin66]

Hmmm

What were the sec/pixel ratios between the scopes? and the OAG guide camera?

What did the guide graph (PHD?) show as pixel movements?

I'd honestly expect zero movement using the OAG, but I suppose there's always the likelyhood of some field rotation ie the angular distance from the centre of the field and the position of the pick off OAG prism...

[themos]

Both scopes were 80/480 nominal, the triplet with the SW flattener. The triplet was imaging at 2.22 arcsec/pixel and the doublet at 2.45 (from astrometry.net). The 10% difference comes from the pixel sizes of the 450D and the 1000D (5.2 vs 5.7). The guider has a 5.2 pixel sensor so the same as the 450D, so that was running at 2.22 arcsec/pixel too.

Field rotation eh? I've added graphs for the angle and that idea makes sense. There's definitely field rotation going on. Before the meridian I get a lot more rotation than after. That's down to a different star being picked as the guide star, is that possible?

[themos]

Actually, that makes sense in terms of numbers. I used

Maximum Guide Star Angle for Polar Alignment Error Calculator

and put in 5 arcmin for PA error (I know that's typical because I've measured it before), 480mm for focal length, 15 minutes to see what happens after 3 subs, 61 degrees for the declination of the target, and 5.2 microns for the pixel size. The calculator then says that the guide star should not be further than 3 degrees away and my DSLR field is 2.6x1.7 degrees so the OAG could be picking a star close to the edges of that.

I'm happy with that understanding. Time to see if I can get my PA error below 3 arcmin.

[themos]

I have to keep in mind that another problem can raise its ugly head if Polar Alignment is too good: Declination corrections tend to go in either direction and with stiction and backlash in the gears, guiding can get harder. Hope to counter this by biasing the scope weight distribution in Dec.

[MG1]

Have you considered come error?

Doesn't this manifest as field rotation between subs?

I'm still getting my head around this so apologies if it's a daft suggestion. CE is a bit of a nemesis of mine ATM!

[themos]

No, cone error comes in when you're trying point at things and you assume that RA and DEC axes of your scope are at right angles when they aren't. I'm just following the scope as it rotates around the RA axis and guiding in both directions, it doesn't matter if the two directions are not at right angles as guiding is a closed loop.

[FrankieValley]

Ok, probably out of my depth here, but here goes..... Would there not be differential flexure between the OAG pick up and the star you are measuring? Would the flexure in dx/dy be dependant on the orientation of the cameras in relation to the centre of rotation? To put it another way, if your OAG and star image was (coincidentally) lined up along the radian (I mean the wheel spoke if radian is not the correct term), would you get flexure in dx (or dy) only? Could the following doublet flexure, that is more prominent in dx, be explained by the fact that it is at (approx) right angles to the spoke/radian (OAG/star alignment). Would rotating the OAG around the optical axis have shown a different result, or perhaps pointing elsewhere would dis-align the OAG and star from in line with the spoke and produce flexure in both?

Just thinking out loud and sorry, no answers, just more questions. Interesting test and results though. Thanks for stimulating the grey matter.

[MG1]

The other thing to consider is that it may not be only in one axis...Just that you are quantum limited by your apparatus.

[themos]

Theoretically everything flexes! But I'd be surprised to see such a large flexure between the imaging sensor and the OAG sensor.

Yes, the amount of flex you get in dx/dy depends on which way gravity is pulling relative to the orientation of the sensor. On the other hand, I know I am getting field rotation ( because DSS measured it). And the plausible explanation for that is the polar alignment error coupled with the distance of the guide star from the star at the center of the imaging sensor. I will use Occam's razor and suppose that all of the drift and rotation I see in the triplet is down to that, with any extra shown by the doublet being down to the fact it flexes more.

What I need to do is measure my polar alignment error each time I set up and then try to correlate that error with the field rotation abd the drift. So, I need to think of a quick way of measuring PA error and I think I know of one.