Guiding camera exposure time

[-Joe_]

I was reading through Kiers 100 Best Astrophotography Targets today and the section at the back on guiding mentions the ideal guiding camera exposure is 5s, and that under 2s may cause guidance problems due to ‘chasing the seeing conditions’.  How does that reckon with the consensus here? 

‘I’ve been getting about 0.9” RMS on my guidance  (with the 72ED),  and setting the exposure to 1s, thinking quicker is better, as long I can see the star. What’s your opinion? I can sort of see the logic in Kiers opinion. 

is my 0.9 good/bad given the 72ed?

Edited July 12, 2019 by -Joe_

[vlaiv]

Depends on the mount you are trying to guide.

Longer is better, provided that your mount supports it - it minimizes seeing influence, and indeed you won't be chasing the seeing. Some mounts are capable of being guided with 10+ second exposures - but not many. You need to have a smooth running mount with slow periodic error (and almost no random jitter) to be able to do that.

You can figure out what is the best guide exposure for your mount if you record periodic error (or periodic error residual after applying periodic error correction) and examine it for max error change (or speed of error) - this is given in arc seconds per second or perhaps percent of sidereal speed - in either case unit is the same "/s. From that you can figure out how long your guide exposure can be - by setting upper limit on how much you want your mount to deviate from intended position. If your mount for example has 0.1"/s as max error rate and you allow for 0.4" absolute error - then you can use maximum of 4s guide exposure (or rather guide cycle - if you want to be very precise about it - it is exposure + latency + guide pulse duration, but in reality you don't need to be that precise because most other things related to this aren't).

If this is too complex for you (don't have PE/PEC done and don't want to do it) - another measure is sub FWHM vs guide exposure length. With EQ6-R you should be able to do at least 3-4s guide exposure and that is what you should be using. Use shorter than that if you asses that by using such exposure duration you are constantly getting higher star FWHM then with shorter durations. Try going higher as well and use same measure - if FWHM gets better - use longer guide exposure.

On my belt modded HEQ5 I usually use 4s guide exposure - sometimes less if seeing is particularly good. In different conditions, I don't see much difference between 4s and 6s. I went up to 8s, but that is causing another problem - sync with dither, and waiting for settle time increases. That is ok for CCD style imaging where one uses 10 minute subs, so dither can be 30s or so, but I use CMOS and short exposures (1-2 minute), so having 30s dither and settling will cause as much as 20% of imaging time wasted.

[Ronclarke]

I vary mine, if I'm imaging low down I use 2-3s but near the zenith I would use 1s max.

[Starwiz]

I was using 1s, but recently upped it to 3s.

I'm still trying to decide whether it's better or not.  Perhaps I need to do some experiments in the same conditions.

John

[spillage]

I usually set mine between 2-2.5s but will make small adjustments as and when I think its needed. If my guiding gets worse I just move it back. I usually get 0.6 on average. I tend to rely more on the target in PHD2 and use this to check for any balance issues or cable snags. Sky condition on the night will also make a huge difference.

[rickwayne]

I've seen a recent article that seemed to debunk the "chasing the seeing" concept pretty thoroughly and argued that you should use the shortest cycle that gave a sufficient SNR. I will try to rediscover it.

[ollypenrice]

I don't agree at all with the advice in the book. On our two premium mounts, Mesu 200s, we use what I consider to be long guide subs of 4 seconds. These mounts have very low and very slow PE and don't move far in 4 seconds.  When guiding our EQ sixes we guide at sub-second intervals because there is rapid PE. If we guided at five second intervals with these mounts we'd be all over the place.

However, the guide trace itself cannot be entirely relied upon to tell the whole truth: if you go for ultra-short guide subs the mount won't get far along its error before being corrected, so you'll get a nice RMS figure but you may be chasing the seeing in doing so. However, I hope Rickwayne can find that link because I've long suspected that short subs work well. I only use long ones on the Mesus because I can and I've been prepared to accept the idea that longer subs will give a more averaged-out stellar centroid.

To test it properly we would need to make a good set of exposures guiding on short subs, another good set guiding on long ones, then compare the FWHM values of the resulting images.

Olly

[vlaiv]

For all those that think that chasing the seeing might not be a real thing - just look at this or similar videos:

Also note that in this particular video seeing is quite astonishing (you can see incredible level of detail in raw recording!).

Jupiter has angular diameter of about ~40' - yes that is 40 arc seconds. Compare ripples and distortions in this video, or rather their magnitude to diameter of Jupiter (and keep in mind that this is as good as it gets most of the time in terms of seeing).

My estimate is that ripples and distortions create local displacement of about 1" or so. If one was doing short cycle guiding - like 1/30th or similar - graph would jump around and if for every displacement one tried to make a correction (magnitude of error warrants correction) - their mount would be all over the place, and it would be definitively chasing the seeing as mount can't respond in such short period of time.

Let's do some basic math and try to "guesstimate" star position error vs integration duration. If every 1/30th of a second there is 1" in magnitude in excellent seeing, "stacking" 64 of those (provided they are random) will give you x8 improvement in position accuracy - so for about 2 seconds of exposure you get 0.125" error - that is acceptable error for most mounts and will probably be under "min mo". Thus in excellent seeing one can use 2 second exposures to be below "chasing the seeing" threshold. In average seeing, I reckon that 1/30th ripple will be at least 2", so you really need 4s or more guide exposure to smooth it out. In poor seeing, you need to go as long as 8s or more (even 10s sometimes won't help if there are local thermals that cause large ripples - happens to me sometimes in the winter when a lot of people is heating their houses around me and some of that heat escapes and forms local thermals and very poor local seeing).

On the other hand, each mount can be measured for max PE (PE residual after PEC) change and you can determine how much real error - one due to motion of the mount and not seeing / wind and other stuff that you don't want to correct for, there is per one second - you will find most often that you should be able to guide 3-4s to be within 0.1-0.2" for most decent mounts (if not - you have mechanical issue and you should fix it in that domain rather than guiding).

[andrew s]

I think the issue may well depend on if the guiding optic is smaller or larger than the atmospheric turbulence cells. Below about 300mm  Optic the image tends to be displaced by turbulence while above 300mm it causes bluing.

I had a different problem when guiding on a spectrogarph hole as the star could spill out one side or another causing a jump in the centroid. I had to use short exposures << 1 s as the star was bright and direct from the main 400mm optic. I countered this by using a very low aggression which essentially integrated out the jumps but countered any long term drift. 

PE was not and issue for me.

Regards Andrew

[rickwayne]

Sorry, I can't find it. My best memory of their analysis -- which was supported by data, but you only have my word for it -- was that "seeing wobbles" are such high-frequency phenomena that you'd need absurdly fast exposures, AND processing, AND mechanical response the pulses, in order to chase them in the first place. Certainly the Jupiter image is wiggling at much more than 1 Hz -- I'd guesstimate more like 10 or more. So even 1-second exposures would be integrating quite a few of those.

Mind you, I've no interest in flogging an angels/pinhead argument here, especially since I don't have the data myself. I do advocate is that one be open to experimentation. If 3 seconds works for you, more power to ya. It's not as if I've got my own guiding dialed in yet, tell you that for free.

[alacant]

Hi everyone

PHD2 has a guiding assistant which recommends an exposure depending on the seeing so removing some of the uncertainty. Using a range of telescopes, it's never recommended anything over 4s.

After a while, you get an feel for what the seeing is like. A good indication  of atmospheric conditions can be had when focusing zoomed on a bright star.

HTH

[Starwiz]

10 hours ago, alacant said:

Hi everyone

PHD2 has a guiding assistant which recommends an exposure depending on the seeing so removing some of the uncertainty. Using a range of telescopes, it's never recommended anything over 4s.

After a while, you get an feel for what the seeing is like. A good indication  of atmospheric conditions can be had when focusing zoomed on a bright star.

HTH

Thanks, I didn't know about that, but will give it a try.

John