Binning when autoguiding to bin or not to bin?

[gregbradley]

Its common practice to bin 2x2 when autoguiding. Mainly because guide stars are often dim.

I have found in the past binning 1x1 I got more accurate autoguiding. Not surprising really as the centroid of the guide star would be a more accurate result as its higher rez.

What have you found? I think I might experiment with 1x1 binning tonight.

Greg.

[Shiraz (Ray)]

according to Cyanogen, "In most cases, you can set Binning to 2 or 3. Binning 1 is only required when you are using a separate guide scope that is a much shorter focal length than the main scope. The guide camera can have 1/10th the resolution of the main camera and still guide very well, since the software can measure the centroid of a star to a tiny fraction of a pixel".

http://www.cyanogen.com/help/maximdl/Autoguiding.htm

FWIW, my guide system works fine with pixels that are about 1/5th the angular resolution of those in the imager.

[DJT (David)]

Hi Greg

I recently went to binning * 2 for guiding and its made quite a difference in terms of having bright enough stars to guide on with the MMOAG. Star shapes are fine

[niharika]

If you have a suitable guide star with good SNR bin 1 is fine, that's what I do most of the time. But at times there are no suitable guide star for reasonable guide exposure and in those cases I do Bin 2x.

[AlexN]

I bin 2x2 when I can't find a bright enough star. I'm more likely to bin my guider than change the framing of the subject.

With ccdsoft and adaptive optics the guider is set to bin 2x2 by default in an attempt to get stars bright enough for fast guiding. I haven't had too much experience with it yet but in the 7 or 8 tests I've done binning 2x2 has allowed 0.5 - 0.1sec guide exposures...

[gregbradley]

Originally Posted by Shiraz

according to Cyanogen, "In most cases, you can set Binning to 2 or 3. Binning 1 is only required when you are using a separate guide scope that is a much shorter focal length than the main scope. The guide camera can have 1/10th the resolution of the main camera and still guide very well, since the software can measure the centroid of a star to a tiny fraction of a pixel".

http://www.cyanogen.com/help/maximdl/Autoguiding.htm

FWIW, my guide system works fine with pixels that are about 1/5th the angular resolution of those in the imager.

Yes SBIG claimed their lens guide kit will guide great. Similar argument. I did not find that to be the case at 1260mm focal length in a good seeing environment on a good mount well polar aligned. It was very hit and miss and mostly miss. So I would modify that datum to it would depend on the PE of the mount.. More accuracy is better than lesser accuracy overall but high PE mounts would have PE that is easy to detect whaereas high end mounts would have very low PE that would require a bit more accuracy to detect.. I notice the AP guide scope is F9 and quite long that they sell for their refractors. But I get that the guide scope may not need to be the same focal length of the main imaging scope. Although the best guiding I get is from a MMOAG which gives the same focal length but smaller image scale with an SBIG STi 7.4 micron pixels.

1x1 binning not always be practical as brightness of guide stars can be an issue. A workaround is to put a reducer on the front of your guide cam to help with star brightness. Guiding with a reducer and a sensitive guide camera with an IR 750nm cutoff in a strong no flex OAG may be the ultimate guide cam setup. If you have enough guide star brightness then 1x1 binning may add an additional level of accuracy. In some cases it may be easier if you only have a really bright over saturated guide star it could bring it down to more manageable levels.

Originally Posted by DJT

Hi Greg

I recently went to binning * 2 for guiding and its made quite a difference in terms of having bright enough stars to guide on with the MMOAG. Star shapes are fine

Originally Posted by niharika

If you have a suitable guide star with good SNR bin 1 is fine, that's what I do most of the time. But at times there are no suitable guide star for reasonable guide exposure and in those cases I do Bin 2x.

[AlexN]

Would it be advantageous to have a 750nm ir cut filter on a guide camera? I guess it would reduce star bloat a little increasing the accuracy of the centroid position calculations.

All these factors only reaffirm my love for self guided cameras. I am really keen on getting an STT 3200ME + FW8G + AO8T as my next camera setup. Best of both worlds. The whole camera/fw/oag/AO setup bolt together into essentially one flexure free box of awesome.

[codemonkey (Lee)]

The key thing about shooting in the near infrared is the reduced turbulence, which enables you to get more accurate guiding. The ONAG is great because (a) you're guiding in the near infrared, (b) no flex and (c) full FOV.

Since the ONAG is out of my budget at the moment, I've just recently bought an IR-pass which I plan to use in conjunction with an OAG and the QHY5L-II. I think it's still going to be a few weeks before that gets here, but I'll post my findings when I do.

Edit: Back on topic.

I've only ever used 2x2 binning a couple of times on the QHY5L-II and I can't really remember my motivation to be honest. I'd only ever think to do it now if I couldn't find a star with 1x1. Whatever you do, make sure your software detects the new effective pixel size properly, or you update it... I forgot to set mine back initially and was like "wow! This is the best guiding I've ever seen!" I was a bit disappointed with the results ;-)

[rally]

If the mount can track accurately then a longer guide exposure time will improve the SNR considerably.
Thats what I did on Tarantula.

[jase (Jason)]

From tests I conducted a couple of years ago there was no difference when guiding 1x1 versus 2x2 in terms of accuracy providing SNR was ample. I wish I had kept the results but in summary how efficiently the software could calculate the star centroid was based on the SNR. 2x2 will often get you a good SNR using short exposures.

The calculation of multiple star centroids on a guide chip (multi-star guiding) will provide even greater guiding accuracy.

[Paul Haese]

I use 2x2 on both systems with regard to traditional guiding. Better star selection with binning and I have not seen any reason to use 1x1. My recommendation is to stick with 2x2.

[gregbradley]

I have a couple of IR pass filters. One from ebay and one from Edmunds Optics. The Edmunds Optics one is not quite the right cutoff of 750nm that the ONAG uses and I think it was more like 720.


If you have bright enough guide stars I did notice a slight improvement in guiding. It wasn't much but then I suppose it was good seeing at the time.

Greg.

Originally Posted by codemonkey

The key thing about shooting in the near infrared is the reduced turbulence, which enables you to get more accurate guiding. The ONAG is great because (a) you're guiding in the near infrared, (b) no flex and (c) full FOV.

Since the ONAG is out of my budget at the moment, I've just recently bought an IR-pass which I plan to use in conjunction with an OAG and the QHY5L-II. I think it's still going to be a few weeks before that gets here, but I'll post my findings when I do.

Edit: Back on topic.

I've only ever used 2x2 binning a couple of times on the QHY5L-II and I can't really remember my motivation to be honest. I'd only ever think to do it now if I couldn't find a star with 1x1. Whatever you do, make sure your software detects the new effective pixel size properly, or you update it... I forgot to set mine back initially and was like "wow! This is the best guiding I've ever seen!" I was a bit disappointed with the results ;-)

[gregbradley]

Interesting Jase. I have been using a few softwares recently. Sky X autoguider shows the HFD of the guide star every time it downloads. It can vary quite a bit. Also the shape of the star can vary a little bit.

Makes me wonder if better rez gets a better calculation. Anecdotally I noticed better guiding in the past using 1x1 at times. Hence the topic.

I think I will check this out a bit more and post my results. Perhaps you are right.

Does Maxim multi star guiding work yet? Or are you referring to some other plug in?

Greg.

Originally Posted by jase

From tests I conducted a couple of years ago there was no difference when guiding 1x1 versus 2x2 in terms of accuracy providing SNR was ample. I wish I had kept the results but in summary how efficiently the software could calculate the star centroid was based on the SNR. 2x2 will often get you a good SNR using short exposures.

The calculation of multiple star centroids on a guide chip (multi-star guiding) will provide even greater guiding accuracy.

[rally]

What is the guide interval you are using ?

[gregbradley]

Originally Posted by rally

What is the guide interval you are using ?

Guide exposure is a variable and it changes depending on the setup as well as seeing but generally 4-6 seconds on the PME and 5-6 seconds on the AP1600. I used to run 1 second with a Tak NJP, 3 seconds with a 1 second delay on PMX, sometimes just 3 seconds no delay or simply 4 seconds.

There is a great astrochannel interview with the developed of PHD2 and he goes over guiding in great detail.

Too short a guide exposure and you can chase bad seeing and lose SNR and make it hard to get a guide star. Too long and you are missing correcting PE especially any step segments of the PE. So guide exposure is not a set amount for a mount but more of a range depending on the current conditions.

Greg.

[PRejto (Peter)]

My experience, guiding through an ONAG on my TEC140 with an ST-i, is that I had improved guiding with 2x2. For me it wasn't about brightness of guide stars as much as finding that with increased resolution unbinned the guide star was too much effected by seeing and would jump all over the place. Perhaps reduced aggression etc might have tamed it but things just generally worked so much better with 2x2 binning that I've not looked back. I've thought of it in relation to average seeing of 2 FWHM. Wouldn't guiding at 1x1 produce spiky results that varied by more than a pixel on a regular basis? If the mount tried to follow that I imagine guiding would be rather rough.

Greg, it will be interesting to see how good your results are. On a very good night of seeing perhaps there will be a dividend.

Peter

[gregbradley]

Yes good point Peter. I have only used it at my dark site where the seeing is often very good.

So perhaps its like undersampling, best when you've got the seeing to support it.

But in choppy conditions you'd want to smooth that motion out to filter out some of the seeing.

Greg.

[rally]

Greg,

Why not guide for a lot longer ?
Have you checked the SNR of your guide image ?
Worth checking every time.

The P-ME can handle an awful lot longer and it really doesnt need to be pushed around with the seeing that often.
Assuming you have good polar alignment - try an unguided exposure of a few minutes (or more) just to assure yourself it can handle more than a few seconds !
Unless something is not right, I bet you couldnt tell the difference.

Rally

[gregbradley]

Its worth a try Rally. I am guiding usually at 5 or 6 seconds which is pretty long. But I can try longer and see what happens.

Greg.