Attention HEQ5 Pro owners, what is your max exposure times and focal lengths guided?

[Ryan_86]

Hi guys, I currently have a HEQ5 Pro belt modded, polemaster, 100 Ds Pro Evostar all guided with an Antares 60mm guide scope/Altair Gpcam2 AR0130M. 

I was wondering how long people reliably guide for with similar setups/focal lengths. 

The focal length of the 100 DS Pro scope with the reducer is 765mm. 

Thanks all

Ryan

[swag72]

I had a standard HEQ5 years ago and used a 500mm focal length scope with a smallish pixel camera at 1.87" pp (this is a more important figure than just focal length as you can use camera's with different pixel sizes that will give you all sorts of combinations of imaging scale) - I used to always do 30 minute exposures with this combination.

[vlaiv]

I've got belt modded HEQ5 and OAG, imaging an 1600mm FL for resolution of ~0.5"/pixel - but I always bin results x2 (1"/pixel) because it is pointless to go that high.

My exposures are 4-5min at most, but not because of guiding, rather because I use CMOS sensor and have strong LP (no point in going for longer exposures). Regardless of that, I'm fairly confident that there would be no mount limit (except for meridian flip) in exposure length.

That was not the case when I still used guide scope, although I did not do long exposures, I was confident it would not work - there was some sort of drift during imaging session. I can probably give rough estimate on max exposure time - in 2h session, this drift would be about 20px long. Keeping it below 1px would imply exposures less than 6 minute long. Never really fully understood the cause for this drift but I think it might have been due to some sort of flex between OTA and guide scope.

[Spikey]

I've done 1 hour exposures with a HEQ5 and ED80 combination.  Guided of course.

[Ryan_86]

Thnaks for your replies guys.

Currently I am using a modded Canon 600d with a resolution of 1.16"pp. Can anybody tell me what sort of RMS I would need to be achieve good guide results?

Am I correct in saying I cant change the binning of a dslr?

Thanks again all???

[PhotoGav]

The theory of guiding is that your RMS error should be a max of half your pixel scale. So with your set up at 1.16”/px, you should be aiming for a guiding error of less than 0.58”. That is do-able on a night of good seeing with an HEQ5. 

I used a belt-modded HEQ5 for quite a few years until just recently and could confidently guide well within 1” error. My Edge 8” and camera are 0.55”/px and while I took 30 minute subs for narrowband with that set up, I was never happy with the sharpness, which I think was down to the guiding not being good enough. I have recently upgraded the mount and hope to get better results... yet to test!

[Ryan_86]

Hmmm? 0.58 is a pretty ambitious number to achieve. I presume that's the total RMS is it or 0.58 on RA & 0.58 DEC? 

Now I am thinking maybe a shorter focal length scope for example the 80ED would result in 2.61"pp, therefore anything below 1.3RMS would be fine. 

Thanks pal??

[carastro]

I have done 30 min subs with a shorter focal length scope.  WO ZS71

Carole 

[Ryan_86]

Hi Carole, 

That's amazing!!! What camera are you using? I love the WO scopes, I had a Z61 when I had a Star Adventurer. I may sell the 100ed and go for something a little smaller.

Ryan

[carastro]

Atik460EX

Have previously done 30 min exposures with an ED120 and Atik314L on the same mount HEQ5.

I don't make a habit of doing such long exposures as there is always the risk of planes flying through the sub, so only do it when I am trying to image something really faint. 

In answer to your other question:

Quote

Am I correct in saying I cant change the binning of a dslr?

I was never able to and so I think that is correct. 

Carole

[PhotoGav]

19 minutes ago, Ryan_86 said:

Hmmm? 0.58 is a pretty ambitious number to achieve. I presume that's the total RMS is it or 0.58 on RA & 0.58 DEC? 

Now I am thinking maybe a shorter focal length scope for example the 80ED would result in 2.61"pp, therefore anything below 1.3RMS would be fine. 

Thanks pal??

Yes, total RMS.

Remember, that is only theory that I was spouting! The best thing to do is give it a go, with the theory in mind, that way you can assess the results with some understanding. The HEQ5 might just surprise you! The other thing to consider to improve guiding is an off axis guider rather than separate guide scope. That can eliminate all sorts of weirdness, especially where longer focal lengths / smaller pixel scales are concerned.

You are right, an 80ED with the larger pixel scale would be more forgiving, but use what you have first!

Good luck and I look forward to hearing / seeing how you get on.

[carastro]

Recent guide graph:

click on the link, hopefully it will work. If not let me know and I will post it in the normal way. 

[Ryan_86]

Thanks for the graph Carole, however it displays an error when I click it. 

PhotoGav, I am patiently waiting for a gap in the clouds to record a PE curve. I am hoping it will help me further. 

I will stick at it especially now I understand image scale etc thanks you everybody's help??

Much appreciated guys.

Ryan

[carastro]

I had a feeling it might not open for any-one except me.  By the way I still use PHD1

attached again. 

[vlaiv]

Yes you should aim for 0.5" Total RMS for that kind of resolution.

If you can't achieve such guiding, then bin your DSLR images. It can't do hardware bin, but in software it is easily done. If you want to go for lower resolution, first thing that you can do is use super pixel mode for debayering - that will be very similar to x2 bin - since you are using OSC camera that is even better than to debayer and then bin x2 in software.

Btw, 0.5" RMS is achievable on tuned and belt modded HEQ5, but with such small guide scope, you are lacking star position measurement precision to be certain that you are guiding that good. With only 227mm FL (if I'm right about your guide scope) and 3.75um pixel camera, your guide resolution will be ~3.41"/pixel.  About 1/16th - 1/20th of that will be measurement error in star position - and that is 0.17"-0.2". You probably need your guide RMS to be at least 3 times that to be able to measure it correctly - 0.5" is at bottom limit of "precision". This does not mean that you won't be able to guide at 0.5" RMS, it just means that if PHD2 is telling you that you are guiding at 0.5" RMS (or other low value) - that might not be quite correct - it might in fact be a bit less or a bit more than that.

I've found that sometimes you can go lower than 0.5" RMS with HEQ5 but such nights require best of conditions - absolutely no wind, and steady atmosphere. Here is my "personal best":

[Ryan_86]

Hi Vlaiv,

Thanks for your reply pal. Ok so all in all what is my best option? I am keen to image galaxies so would prefer not to change for a shorter FL imaging scope. Maybe a guide scope with a longer focal length, or what about a OAG. I'm not familiar with OAG and what the advantages a nd disadvantages are. 

My current payload is 7kg so adding a larger guide scope will put me close to the limit. 

Hmmm decisions decisions, what do you guys advise?

Thanks again all.

Ryan

[vlaiv]

Many people believe that OAG guiding is harder than guide scope for some reason. Often quoted "problems" are - awkward to setup and use, trouble finding suitable guide star and such.

In my personal experience OAG was just as simple to use as guide scope and maybe even better. Focusing OAG is maybe harder than guide scope, by a hair - just because there is small screw that you need to undo and then move camera / prism assembly by hand, but once adjusted you really don't need to refocus OAG that much. With sensitive CMOS guide cameras, there is always at least few stars that you can choose from - above screen shot is probably representative of a bit more crowded part of sky, but I still have not had a case where I needed to change OAG position because there were no guide stars in FOV.

It is lighter than guide scope, you don't have to worry about mounting it and any flex that can come with your choice of mounting rig. You do need to make sure you have enough distance between sensor and FF/FR to accommodate OAG. There are different models with different optical lengths - there is even very short one - 9mm of optical path. I use one with 16mm and have been able to use it with FF/FR that requires 61mm. My filter drawer is also very short. If I was to use filter wheel, than I'm not sure I could fit everything and still have 61mm - guess it depends on filter wheel optical length.

OAG is probably better suited to dedicated AP cameras rather than DSLR, but I've seen OAGs specially designed to go with DSLR, so that could be one concern - you would be getting non reusable component (if you switch to dedicated camera, you'll probably need different OAG as well).

Another plus side that I can think of is guide measurement resolution - guide cameras have small pixels, usually smaller than imaging camera - and they will be both working on same FL, so you don't have to worry if your guide setup is providing enough resolution - it certainly will. I even bin my guide cam because natively it is ~0.48"/pixel (1/16th of that is 0.03" so good precision for 0.1" RMS or higher - something HEQ5 just can't do) and I don't need that much precision for HEQ5, so I recover some SNR for guide stars by binning it.

There is one additional thing - most people when switching to OAG say that their guide graph looks worse, but stars are tighter - this is because of added measurement resolution of such setup, so if you choose to go with OAG - rate your guide performance on stars rather than graph (and graph and rms readings will probably be more precise).

As you could have guessed - +1 for OAG

 

[Stub Mandrel]

Here's one for you @vlaiv

I have often noticed looking at the 'target' in PHD2 that the apparent centre of the dots is off-centre.

Presumably this is because the mount is playing continual 'catch up' so even if the errors are normally distributed it is around a point which is NOT the target point.

In your example above, for example, I'd estimate the mean RA error (not RMS error) to be about -0.2. I have seen the distribution even more heavily skewed, especially in times of poor transparency.

If the RMS error was calculated from the mean position, rather than the target position, it would be somewhat smaller.

I suspect this is why people's results are often somewhat better than the RMS figures would suggest.

Neil

[vlaiv]

26 minutes ago, Stub Mandrel said:

Here's one for you @vlaiv

I have often noticed looking at the 'target' in PHD2 that the apparent centre of the dots is off-centre.

Presumably this is because the mount is playing continual 'catch up' so even if the errors are normally distributed it is around a point which is NOT the target point.

In your example above, for example, I'd estimate the mean RA error (not RMS error) to be about -0.2. I have seen the distribution even more heavily skewed, especially in times of poor transparency.

If the RMS error was calculated from the mean position, rather than the target position, it would be somewhat smaller.

I suspect this is why people's results are often somewhat better than the RMS figures would suggest.

Neil

I'm actually not quite sure how RMS is calculated in Phd2 - will need to check the code to find out (or maybe there is reference somewhere on internet for this), but you are right - RMS as a measure should not include fixed/assumed position, and it should be calculated based on average of all samples taken into account when measuring RMS.

If it is indeed calculated based on fixed position then we can look at particular extreme situation to see how it might skew results - just imagine perfectly flat line in RA deviation (ideal mount that does not react to guide commands) displaced from 0 by some amount. For duration of exposure RMS calculated based on "standard" algorithm would be 0 - as there is no change of position, while one calculated based on absolute reference point would still have certain value (if I'm not mistaken, of the top of my head it would be precisely equal to line offset from 0).

I think that PHD2 actually measures based on average and not fixed / absolute reference. It is measure of jaggedness of line rather than it's position on graph (DC offset), at least I think so. I remember thinking about this couple of times when my guide graph was rather smooth but offset - reported RMS error was smaller then I would expect if it was calculated from absolute baseline.

Btw - such case where line has DC offset and is smooth is indicative of poor guiding. Lagging / trailing can happen for various reasons - mount can have periodic error component that is both longer in period than base worm period and higher in amplitude - this case is easily guided out in normal circumstances. It can also be due to poor PA, or refraction of atmosphere - when guiding low, close to horizon.

In any case, effective guiding should have saw tooth pattern and be of low amplitude - each time pulse is issued it should return mount to absolute 0 position and due to constant drift it can happen that mean position is not the same as absolute position. What can be poor guiding is if guide pulses don't return close to absolute 0 (it will always have some error because of measure precision and mount response, but I would expect such error to be random, so sometimes it would undershoot 0 or overshoot 0 - what should not happen is undershoot most of the time - this means that issued pulse is not enough to correct for one or several guide cycles).

Btw - skew in mean position can happen even if there is no chasing - simply by virtue of random numbers. Given a set of random numbers centered around some value - their average will also deviate slightly from central value. One needs infinite number of samples to have average be exactly center of distribution (this is why there is still some noise in image regardless of the number of subs - it does go down with increased number of samples but never reaching exactly 0).

Increasing number of samples in PHD2 "target" screen can help detect any offset with greater precision.

While we are at it - I still don't know, nor have I found reference at exact characterization of PSF resulting from guide error graph or RMS figure alone. This complicates things when we are trying to relate guide RMS error to resulting star profile / FWHM value. What can be said with certainty is that those two are correlated - larger RMS will result in larger FWHM, and rule of thumb about keeping RMS at or below sampling resolution is just that - rule of thumb - empirical figure. Don't know how would I go about explaining it analytically. Simplest way would of course be to somehow translate RMS value into Gaussian PSF of certain sigma / FWHM. I know how to proceed from that point, but this first step is troubling me. I have no idea if we can make such assumption and how accurate that approximation is. Central theorem is pointing us to idea that we should be able to do such approximation - given enough samples. I'm just not sure if 50 or 100 samples is enough.