58 replies to this topic

[syyntax]

When selecting a guide rate on the mount, in my case I have 0.5 to 0.9 as the range I can choose from.

 

I understand that once the mount is on target, it's tracking at the sidereal (or King) Rate.

 

Is this measured as a percentage of sidereal? or is it some other rate altogether?

 

I'm trying to get an understanding of why you might choose a slower/faster guide rate for use when guiding?

 

Thanks!

 

 

[CharlesW]

You’ll want to set it at .5 sidereal. Here is a bit from TheSkyX Pro’s manual, “The right ascension axis is always moving to keep up with the apparent motion of the stars. This rate of motion is called the sidereal rate, and is very nearly one revolution per day. You will often see reference to some fraction of the sidereal rate, such as 1x (exact sidereal rate), 0.5x (one-half sidereal rate), etc.

Because the right ascension axis is always moving, you have an opportunity to eliminate many (but not all) sources of error by simply making guiding corrections at speeds slower than the sidereal rate. For example, it is common to make right ascension adjustments using either 0.5x sidereal (slow the mount down to half speed to move it further east), or 1.5x sidereal rate (speed the mount up to 150% of sidereal rate to move it further west). Depending on the focal length of your telescope, you may get better results at 0.75x/1.25x (shorter focal lengths), or at 0.25x/1.75x (longer focal lengths).
The declination axis, on the other hand, is stationary until a correction needs to be made. If there is substantial looseness in the declination axis at any point (motor bearings, reduction gears, worm endplay, or worm mesh with the worm gear), the mount may literally be unable to guide adequately in declination.”

[syyntax]

so If I needed to make quicker RA corrections (to correct high frequency movement), I would want this set higher? (I've been using .9x, the highest I have).

 

It sounds like my corrections would take longer if I set it to a slower rate, am I thinking correctly?

[Lite2]

Glad this came up. I never knew what it meant and just stuck with .51, my AVX default.

Maybe that's the time I had such bad tracking with my old RC when I set it to .90

Edited by Lite2, 17 July 2018 - 08:04 PM.

[Jon Rista]

so If I needed to make quicker RA corrections (to correct high frequency movement), I would want this set higher? (I've been using .9x, the highest I have).

 

It sounds like my corrections would take longer if I set it to a slower rate, am I thinking correctly?

Guide rate sets a lower and upper speed for how quickly guiding moves the mount relative to normal tracking rate.

 

Your mount tracks at 1.0 by default. If you use 0.9, then that sets the lower speed to 0.9x sidereal, and the upper speed to 1.1x sidereal. If you use 0.5, then that sets the lower speed to 0.5x sidereal, and the upper speed to 1.5x sidereal. As such, 0.9 means guiding is in effect "slower", while 0.5 means guiding is in effect "faster". 

[Mike7Mak]

Guide rate sets a lower and upper speed for how quickly guiding moves the mount relative to normal tracking rate.

 

Your mount tracks at 1.0 by default. If you use 0.9, then that sets the lower speed to 0.9x sidereal, and the upper speed to 1.1x sidereal. If you use 0.5, then that sets the lower speed to 0.5x sidereal, and the upper speed to 1.5x sidereal. As such, 0.9 means guiding is in effect "slower", while 0.5 means guiding is in effect "faster". 

I don't think so Jon. The guide rate number is the amount of sidereal that the mount either slows down by or speeds up by. A guide rate of .9 slows the sidereal speed to .1 and speeds it up to 1.9. A .5 rate means slower corrections than .9.

Edited by Mike7Mak, 18 July 2018 - 02:27 AM.

[spokeshave]

I don't think so Jon. The guide rate number is the amount of sidereal that the mount either slows down by or speeds up by. A guide rate of .9 slows the sidereal speed to .1 and speeds it up to 1.9. A .5 rate means slower corrections than .9.

This is correct.

 

Tim

[syyntax]

After reading a *lot* of guiding related posts here, I've seen several instances where the information was conflicted.

 

I can see logic in both of the above explanations, which was my initial confusion that led me to ask.

 

So, how would we go about determining this empirically?

 

Say, if I set my guide rate to 0.5, what can I monitor that I could then compare against using the 0.9x setting?

 

Could this possibly be mount specific?

Could Ascom factor in some way (is there source code that can be examined for example).

 

I don't mind spending some time (if I can get some breaks from the clouds) to run some experiments, I just need to know where to start.

[freestar8n]

If it is tracking well then objects should remain centered as long as you don't make a guide command in RA.

 

If the rate is 0.1x then pressing a guide button E or W should make an object move about the same speed across the view or camera - just opposite directions.  That tells you it is doing 0.9x in one case and 1.1x in the other - because the star itself is always "moving" at 1x.

 

It's good to keep in mind that if your guide errors are small and the guide rate is 0.5x, your typical correction pulses will be milliseconds.  The normal rate of motion is 15 arc-sec per second, so a correction of 1 arc-sec would take 1/15 of a second at 1x and 2/15 of a second at 0.5x.  And 1 arc-sec is a pretty big error to correct during normal guiding.

 

I see some mount manufacturers saying to guide with 1x for some reason - but that is very odd to me since it means actually stalling the mount in one direction.

 

I use 0.5x for RA - but for dec. it is completely different and will depend on how the guide software behaves and how the mount backlash behaves.  The rate has little connection to how the mount actually moves if it is winding out backlash.  Dec. errors tend to be harder to correct - but they are also less fast - so they can be chased more casually.  You just don't want to overshoot.

 

And normally you also have aggression to set - and a value of 1 for aggression says how much you trust the error value and if you want to correct it exactly - while the guide rate sets how quickly those error corrections will happen.

 

Frank

Edited by freestar8n, 18 July 2018 - 07:24 AM.

[Jon Rista]

I don't think so Jon. The guide rate number is the amount of sidereal that the mount either slows down by or speeds up by. A guide rate of .9 slows the sidereal speed to .1 and speeds it up to 1.9. A .5 rate means slower corrections than .9.

Are you sure about that? And is this the same with every mount?

 

I use EQMOD, so maybe it is different than most mounts...but 0.9x sidereal seems to guide more slowly than 0.5x sidereal, and 0.3 seems to guide faster than both.

[entilza]

I found this whole guide rate even more confusing when switching to EQMOD. The developer defaults the guide rate to 0.10x. The reason stated is to do more smaller corrections instead of larger ones.

I locked down my EQMOD to 0.5x just from habit and haven't really looked back. I felt by locking down one field and becoming familiar with it, I can later judge if a change makes any noticeable difference, however I haven't really gone back to do more through testing.

Now the aggression factor in PHD2 also seems to have an effect.

I now have my RA/DEC aggregations down to 60 in PHD2, perhaps with a lower guide rate these could be higher.. I really don't know. I just go with what seems to have worked, too many settings in this hobby.

[syyntax]

Are you sure about that? And is this the same with every mount?

 

I use EQMOD, so maybe it is different than most mounts...but 0.9x sidereal seems to guide more slowly than 0.5x sidereal, and 0.3 seems to guide faster than both.

 

Jon,

 

How do you tell if it's guiding "more slowly/faster"? Do you watch the duration/number of the correction pulses? 

 

I think if I set up a scenario where I do the usual PHD calibration/GA routines using a .5x value and guided for an amount of time, then do the same for .9x and then compare the logs, I should see some differences that would be useful.

 

What would be the useful statistics in the phd logs to examine?

[terry59]

I now have my RA/DEC aggregations down to 60 in PHD2, perhaps with a lower guide rate these could be higher.. I really don't know. I just go with what seems to have worked, too many settings in this hobby.

We want RA and DEC aggression numbers to be smaller, no? 60 implies pretty good to me 

Edited by terry59, 18 July 2018 - 10:52 AM.

[entilza]

We want RA and DEC numbers to be smaller, no? 60 implies pretty good to me

Yeah, based on what I've read here, but if we were to reduce the guiderate speed, would an increased aggression just offset the difference here?

[entilza]

Jon,
 
How do you tell if it's guiding "more slowly/faster"? Do you watch the duration/number of the correction pulses? 
 
I think if I set up a scenario where I do the usual PHD calibration/GA routines using a .5x value and guided for an amount of time, then do the same for .9x and then compare the logs, I should see some differences that would be useful.
 
What would be the useful statistics in the phd logs to examine?

The best result would be a reduced total RMS " error.

Just be sure to re-calibrate after changing guide rate as well.

[syyntax]

That is definitely the end goal, but I'm also trying to "learn the tools" to the best degree that I can.

 

I want to get to a point where I can look at my graph/logs, and given my current observing conditions, and my current image data (fwhm, etc) be able to reliably "tune" PHD in the right direction..

Without flailing about / shooting in the dark so to speak (pun intended).

 

Sort of like what a good mechanic does when working on a car's engine.. if it's idling rough, he'll know to look at the timing, or fuel mixing etc...

 

+1 on the re-calibrate after changing guide rate, as it's used to calculate stepsize in PHD.

 

So, given my scenario in post #9 , if I slowed down my guide rate, I should likely see: increased # of corrections  and/or increased pulse durations.

[terry59]

Yeah, based on what I've read here, but if we were to reduce the guiderate speed, would an increased aggression just offset the difference here?

Good question, I always have to tweak my settings since the environment is different each time. When I can use 60 aggressiveness, have a good RMS error and subs have a good FWHM/eccentricity it is an exceptional night along the front range for me

 

Edit: I've never adjusted the guide rate in Gemini 2 so....dunno what that would do

Edited by terry59, 18 July 2018 - 10:10 AM.

[bobzeq25]

The best explanation I've seen is in this superb book.

 

https://www.amazon.c...h/dp/1138055360

 

Crucial points.  Autoguiding is plagued by very poor signal to noise ratio, the signal consisting of things like PE and atmospheric dispersion, the noise things like seeing moving the guide star even though tracking is just fine.

 

A key to improving the situation is that the signal is generally lower frequency than the noise.  So, contrary to your intuition (as is so often true in AP), less rapid response of the system can improve the situation.  The data input to the autoguider unavoidably contains both signal and noise, applying a low pass filter can improve SNR.

 

Many of the available tools serve as low pass filters.  Exposure is the obvious one.  Everyone uses longer exposures in poor seeing.  Agressiveness is another, it tells the mount not to take PhD2's suggestions too seriously right now.  <grin>

 

And so is guide rate.  Faster is not necessarily better.

 

Getting the most out of guiding is based on figuring out how to adjust the various parameters, which interact.  You can't universally say a faster or a slower guide rate is better, it depends on everything else, including the equipment, the situation, and the other settings.

 

Main thing to realize.  The less you correct the system with the necessary but annoyingly imprecise autoguiding, the better.  Why people like good mounts.  Why polar alignment is important.  PhD2 has a setting which turns off DEC guiding and just lets your excellent polar alignment handle it.  That's saying "we give up".  <smile>

Edited by bobzeq25, 18 July 2018 - 09:57 AM.

[spokeshave]

There are some misunderstandings here,I think. First, guide rate is the fraction of the sidereal rate that is used for guiding moves. For DEC, this is simple. If your guide rate is 0.5x, then the guide pulses will be at half of the sidereal rate. A rate of 0.9x would guide at 90% of the sidereal rate. For RA it is a bit more complicated since the RA axis is already moving at sidereal speed. Here, if your guide rate is 0.5x, it means that the RA guide rate will change by half of the sidereal rate. So a pulse to the west will move the mount at 1.5x the sidereal rate, and a pulse east will move it at 0.5x the sidereal rate. Similarly, if your guide rate is 0.9x, a west pulse would be at 1.9x sidereal, and a pulse east would be at 0.1x sidereal.

 

There is no relationship between guide rate and guiding aggressiveness. The aggressiveness setting determines the amount of correction to send to the mount. The guide rate lets PHD2 translate that amount of correction into a pulse length. Guide rate will not affect the amount of correction needed, only the amount of time the correction takes. For example, if PHD2 determines that a 0.5 arc-second move is needed to the north, it calculates the pulse length needed based on the guide rate. If you are guiding at 0.5x, that translates into 7.5 arc-seconds per second. To make a 0.5 arc-second correction, PHD2 would need to send a pulse that is 7.5/0.5 = 67ms in duration. If your guide rate is 0.9x, PHD2 would calculate a pulse length of 37ms to make the same correction.

 

Tim

Edited by spokeshave, 18 July 2018 - 10:36 AM.

[syyntax]

Thank you Tim!

 

This is the type of information I'm looking for, this should be reflected in the PHD log correct?

 

RAGuideDistance, DECGuideDistance are the distances of the correction in arc-seconds.

RADuration, RADirection, DECDuration, DECDirection are the direction/duration of the pulses made to attempt to correct.

 

So, when comparing a .5x to .9x phd run, I should expect to see (for a given correction distance) the .9x pulse will be shorter (faster).

 

The next few nights that I get, I'm working on improving my guiding (trying to get it to remain relatively stable through several hours), and I'll be running various guide rates (and various phd settings) to attempt to find a good base to start with on a given evening. That way I'll only have to deal with adjusting based on the nightly conditions and make relatively small changes to adjust.

 

Edit, corrected my comparison example above for better wording.

Edited by syyntax, 18 July 2018 - 10:15 AM.

[spokeshave]

In theory, guide rate should not materially affect guiding. The only thing it should affect is the duration of the guide pulse. However, in practice, there are always complicating factors. Things tend to get really complicated with ST-4 guiding - which is the primary reason that mounts allow you to set the guide rate. With ST-4 guiding, a guiding correction is performed by first sending a command to the mount to start the guiding move, and then sending a stop command to the mount when the move is complete. The mount is continually polling the guide port waiting for these commands. However, the polling frequency is crucially important. Most older mounts have a relatively sparse polling frequency - sometimes as low as 4 times per second. While that may sound fast, those 250ms between polls are actually quite long. Imaging that you are guiding at 0.9x sidereal, and PHD2 decides that it wants to make a correction of 0.2 arc-seconds. It would calculate a guiding correction duration of 15ms. With ST-4 guiding, it would work like this. PHD2 would send a "guide" command to the mount and 15ms later, it would send a "guide off" command. However, the mount only polls the ST-4 port once every 250ms. So, the mount receives a "guide" command from the port and starts to make a move at the guide rate. It then waits (while still moving). Even though PHD2 sent the "guide off" command 15ms later, the mount won't receive the command until the next poll 250ms later. So what should have been a 15ms movement turns into a 250ms one and potentially even longer depending on the timing. Now this is a worst-case example, and most modern mounts poll faster than that but even at a polling frequency of 100 times per second, the interval between polls is still 10ms. Depending on the timing of the "guide" and "guide off" commands, the correction can be as much as 10ms too long. Most of the puny processors inside mounts aren't capable of polling much more frequently than that while still performing the other mount functions. For this reason, a slower guide rate is recommended for ST-4 guiding since a longer guide correction will be less susceptible to polling error.

 

With ASCOM pulse guiding, things work differently. The PHD2 software does not send individual "guide" and "guide off" commands. Instead, it sends a single command that says "guide for XXXms" and that's all. Here there is no polling error. The mount receives the command and executes it without waiting for a second command to stop. So guide rate is far less important. However, there are a few things to keep in mind. For heavily loaded mounts, or mounts that are not particularly vibrationally stable, a fast guide rate may induce some "bounce" particularly if several pulses come in succession. On the other hand, fast guide rates will make settling after dithers go much faster. Slow guide rates have disadvantages as well. The longer pulses take longer to complete the correction, so the deviation remains in place longer (though we're still only talking about milliseconds). Additionally I have read (though this seems dubious to me) that slow guide rates tend to have more difficulty overcoming stiction in the mount. 

 

So, here's my advice (worth every penny paid for it). If you are using ST-4 guiding, the first thing you should do is stop and switch to ASCOM. Failing that, you should pay close attention to the guide rate and guiding performance when short pulses are used. You want to set the guide rate slow enough such that you don't see overshoot from polling error. 0.25x is a good starting point. If you're using ASCOM guiding (if you're not, you should be for this very reason) don't worry to much about guide rate. You typically can set it as high as you want up to the point of inducing mechanical bounce. When people ask me about this, I typically tell them to start at 0.5x and work your was up until you notice a degradation in guiding, or until you reach your mounts guide rate range limit - whichever comes first. If you don't want to do that, just start at 0.5x and leave it there. You are not likely to see mechanical bounce at that rate.

 

Finally, when using ASCOM guiding, the guide rate should have little to no effect on guiding performance provided you aren't seeing bounce. Even long pulses will almost always be completed before the next guide image is taken since even long pulses are still only a fraction of a second long. As a result, the correction is almost always in place before the next guide image is taken. So, no harm, no foul regarding guiding performance. However, a higher guide rate can produce better images even when the guiding is unchanged. That's because the corrections happen faster so the imaging sensor is collecting less time with the tracking deviated. However, this effect will likely be tiny and not noticeable. If you are having guiding problems, there are at least half a dozen other genuine gremlins you should be chasing before you even consider changing the guide rate.

 

Tim

[Jon Rista]

Jon,

 

How do you tell if it's guiding "more slowly/faster"? Do you watch the duration/number of the correction pulses? 

 

I think if I set up a scenario where I do the usual PHD calibration/GA routines using a .5x value and guided for an amount of time, then do the same for .9x and then compare the logs, I should see some differences that would be useful.

 

What would be the useful statistics in the phd logs to examine?

I was going by correction pulses. But I could have been misreading or misunderstanding something. With 0.9, IIRC the guide pulses were usually longer, while at 0.3 (or even lower) the guide pulses were usually too short (EQMOD has a minimum guide pulse of 20ms, and at smaller guide rate values, I often seemed to be spitting out guide pulses of only a few milliseconds). I usually use 0.5 these days, and rely on other means for tuning guiding. Seems most people use 0.5 as well.

[spokeshave]

I was going by correction pulses. But I could have been misreading or misunderstanding something. With 0.9, IIRC the guide pulses were usually longer, while at 0.3 (or even lower) the guide pulses were usually too short (EQMOD has a minimum guide pulse of 20ms, and at smaller guide rate values, I often seemed to be spitting out guide pulses of only a few milliseconds). I usually use 0.5 these days, and rely on other means for tuning guiding. Seems most people use 0.5 as well.

Jon:

 

 Are you sure that's what you saw? You should be seeing just the opposite. I know that's what I see.

 

Tim

[syyntax]

Keep in mind, in my case specifically, I'm trying to tame an unruly RA performance on my mount.. so far my best success has been at fast exposure (1s), and my guiding rate has been at .9x, and my RA aggression is usually 60-70% hysteresis at 5.

 

I've been able to maintain .6" to .8" total RMS with 2" to 3" peaks, getting round stars and decent average fwhm's between 2" and 3".. which is a success in itself for me!

 

However, maintaining that for more than an hour or two before it goes off the rails is proving more difficult. eventually large oscillations build up and I have to stop guiding and re-start.

 

It may be that I've just reached the proverbial limits of circumstances..

[entilza]

Keep in mind, in my case specifically, I'm trying to tame an unruly RA performance on my mount.. so far my best success has been at fast exposure (1s), and my guiding rate has been at .9x, and my RA aggression is usually 60-70% hysteresis at 5.
 
I've been able to maintain .6" to .8" total RMS with 2" to 3" peaks, getting round stars and decent average fwhm's between 2" and 3".. which is a success in itself for me!
 
However, maintaining that for more than an hour or two before it goes off the rails is proving more difficult. eventually large oscillations build up and I have to stop guiding and re-start.
 
It may be that I've just reached the proverbial limits of circumstances..

What's the gain set on your guide camera? I've been seeing a lot of improvements by reducing the gain. Try to go as low as you can. I'm at 40% with me ASI290mm