Monday, February 22, 2021

How To Adjust ASIAir Guide Aggression

 V1.5.3

Why do we guide at all during astrophotography?  The ideal situation would be that we have perfect polar alignment (no drift in declination) and perfect RA drive gears (no periodic error in the worm gear).  In this utopian instance we would not need to guide.  Why you ask?  Because the most ideal guide ignores the effects of seeing and just corrects for the error I mentioned previously.  We don't want to make guide corrections when the star jumps around.  This often makes the image worse than doing nothing.  The solution is to tune the guide pulse and we do this by adjusting the guide exposure, guide pulse multiplier and aggression settings.


Lets talk about a few guide camera settings that will help get to the next steps.  Calibration step and Max dec and RA Duration. 

The calibration duration is simply a value that ensures that a star moves enough during guide calibration so that the axis orientation can be determined.   This should really only take a couple of minutes to complete for both axis.  A value too low will result in an error that the star did not move enough. A value too high will result in a star lost or too few steps completed.  The correct value is dependent on a lot of factors.  The largest factor is the focal length of the guide scope.  Wider field guide systems require the value to be higher so that the movement can be detected.  Conversely if your guide system uses a higher magnification such as when you use an off axis guider or use a barlow lens on your guide scope, this must be reduced.  The goal is to get enough movement in 4 or 5 steps to complete the axis test.   To start you can use this as a guide for some common configurations of guide scope focal length:

ZWO Mini Guide Scope (120mm FL) = 8000ms to 10000ms
Orion 60mm helical focuser guide scope (240mm FL) = 1500ms to 3000ms
Off axis guider (FL same as main scope) = 250ms to 500ms

Again these are just starting points and depending on cameras used it could be different. These should be used as starting points and adjusted so the axis completes calibration in 4 or 5 steps.

Max Duration settings should in most cases be left at the default settings.  PHD2 defaults are 2500ms for each.  Remember these values are Maximum values per guide pulse.  Pulses will more often be smaller in actual use.  So why would you want to change this?  The most frequent use case for reducing this value would be when there are transient issues occurring like wind gusts which would give a short term burst sending the guide star off very quickly but temporarily.  By reducing this value you can prevent the guide from over correcting on such events.   Normally I would just leave this at the default or around 2000ms to 2500ms.  On windy days you can try around 1000ms or less but be mindful that it could take more pulses to get the normal correction back on track. Obviously this is a bit of a trade off and focal length of the guide system will have an impact on the degree of change.  Using an OAG this could come way down to 500ms or less depending on the amount of wind.  Just try different values and see how the system recovers from wind or touching the tube for a second to simulate a wind gust. 


Now that you have your mount and guide session calibrated and have selected a star and started guiding, you will notice the guide graph at the bottom of the guide screen.  There is a lot of information here.  You have options to select the Declination Mode (upper left corner of the graph) and Corrections show or hide (upper right corner of the graph).  Dec mode is used to change how asiair corrects for drift in declination.  Options are off, north, south, auto.  If you are using a tracking mount that doesnt have a motor for declination corrections this should be off.  Most will start with Auto which will correct in both north and south directions. However once you figure out which way your declination drift is trending you can set it for north or south.  This will help by avoiding dec corrections in the wrong direction caused by large movement in seeing.   Corrections should

always be set to show.  I can’t think of any reason not to show corrections on the graph.  This is an important visualization to determine if guiding is too aggressive or not aggressive enough. 

Before we get into aggression settings, lets discuss what we can control.  The first thing to set is the guide exposure.  When we have really good seeing on the order of 0.25” we can use a shorter guide exposure on the order of 1-2s.  When seeing is really poor, the only way to help average seeing is to use a longer exposure like 2-4s.  This is because as the star moves around on a longer exposure it will make a larger star blob tracing out the movement of the star on the guide image.  Phd2 which is what ASIAir uses for guiding uses a centroid algorithm that calculates the center of the blob.  This is the best guess of where the star should actually be on average and where the measurement is made from.  This alone can greatly smooth the graph.  Its ok to start at 1s guide exposure to calibrate, get set up, and evaluate seeing.  But when its show time you should increase this 2-4s to get better results.


The next thing we can control is the guiding rate or guide pulse multiplier.  In ASIAir this is found under mount settings as .25x, .5x(default), .75x, and .9x.  The baseline guide pulse is the pulse sent to mount to keep things moving at the sidereal rate.  When a corrective guide pulse is issued, this setting will send a pulse as .25x the sidereal pulse, .5x , etc. Every mount is different and responds differently.  Most will find that .5x is perfect but not always so.  My Orion mount corrects best at .75x.  For some mounts these settings are not offered in the app and must be set on your hand controller.  You can determine which setting is best by setting both RA and Dec aggression to 50%.  Then try each multiplier setting and watch the total rms number.  Wait for the graph to measure the full width of the chart using the selected setting.  You can use 1s guide exposure for this part to speed up the process.  Choose the setting with the lowest rms error.  Note that as of version 1.5.3 this setting is not persistent and will reset to .5x at boot up.  Another thing to file in the back of your mind is to consider the guide pulse multiplier as a course setting and the aggression sliders as a fine tune setting.  As you try larger or smaller guide pulse multipliers you may need to decrease or increase the aggression to get the best guide.  If you find your aggression is too high or low you may need to adjust the multiplier to keep your aggression in the middle of the range.

Before we get too deep into aggression though, lets talk numbers.


Many folks starting out become fixated on RMS numbers.  So what is RMS. Its stands for Root Mean Square and in this instance its a measure of the seeing and mount error from a perfectly still star measured over a particular period of time.   You can Google RMS to learn the mathematical wizardry here.  But we will keep this simple.  RMS error is important but in most cases these numbers are limited by seeing conditions.   They translate to the approximate size of stars to expect over the given period of time. 

For reference there is a practical limit of seeing anywhere on the planet due to the atmosphere.  At high elevations (observatories) it may be possible to obtain seeing below 0.2 seconds of arc or 0.2”.   For most of us though 0.25” is pretty much the absolute lower limit.  Good seeing is about 0.5” to 0.7” and average is between 0.7” and 1.0”.   You can still image up to 2.0” but stars will be larger and you might benefit from binning.  That will be discussed another time. ZWO has a white paper on their site that discusses binning properties with their CMOS cameras. Any time you get seeing below 0.5” consider yourself lucky.

Now that you have an understanding of target numbers note that these can change over hours minutes or even seconds.  Atmospheric seeing is what makes the guide graph move randomly over time.  When we guide we are not correcting for seeing conditions we are correcting for error in Polar Alignment and in periodic error of the main RA worm gear.  Guiding on seeing changes can actually make your stars and image resolution worse. Therefore we want the guide corrections to be effective but not overly so.  To help average seeing we use longer guide exposures over several seconds sometimes to average the blob that is being measured. 


Lets focus on Declination first.  This is because the rms error for dec is that due to Polar alignment, seeing, and corrrections.  At any time if we have dec guiding tuned properly and assume we have good enough polar alignment, this number should be really close to the practical limit we can achieve due to seeing conditions.  Therefore this should be the first parameter to tune and minimize as we will use this as a guide to set RA.  Its important to tune Dec so corrections mostly are made to correct for polar alignment error.  This will ALWAYS be in one direction north or south and will automatically switch after a meridian flip.  Once you figure out the trend you can set the dec mode to north or south to keep the trend in check. You can lower Dec aggression down to the minimum and watch the graph. The red line will eventually drift up or down the rate of which will depend on how good your polar alignment is.  Select north or south dec mode and see if the rate increases or decreases. You may need to put dec aggression back to 50% or more to see this effect more quickly.

Once you have the mode set, now we can focus on aggression.  Remember we just want to correct for the trend.  If the aggression is too low the trend line will drift away slowly as it did before.   If aggression is too high the correction will cause the line to overshoot the zero line.  If dec mode auto were selected we might see a correction in the opposite direction.  For this we just watch the line.  The correction should not cause it to cross zero line by much.  It takes some trial and error to get this right. But take some time to practice and watch what happens when aggression is set really high and really low so you can spot the error quickly.

Once declination is ironed out its time to tackle the RA setting.  There are more things that can go wrong here but the process is similar with a catch.  Note the rms error we previously obtained in Dec tuning.  Don't worry too much about the number. Just note that we want RA to be about the same as close as we can get.  Since RA is always correcting one way or the other we want to start by moving from 50% aggression to 5 or 10%.  If you see that RA pulses are happening in the same direction a lot one after another then it means it can’t get the line back to zero and the aggression needs to be increased.  Work it back up a bit and wait for the line to come back to zero.  If the aggression is too low you will see multiple corrections in the same direction.  If the aggression is too high you will see a corrective pulse and then immediately a pulse in the opposite direction. Its better if the first one falls a bit short and a smaller second one takes it to zero.  You will see the overshoot happen once in a while even on a properly tuned guide.  But it should be random and seldom.  There should be nothing on the guide graph that looks like a pattern.  If there is, your aggression is likely too high. 

So now that you have things tuned, you can now increase your guide exposure to 2-4s.   This takes some trial and error.  A well tuned mount can take 4s or more but some just have too much periodic error or stiction to tolerate that long period without a guide pulse.  As you go from 1s to longer exposures you may need to choose a new guide star or reduce the gain as the bright ones will start to clip.  You will almost certainly notice that as you increase your exposure, the rms error starts to drop as there are fewer corrections due to seeing.

Now the catch.  Nearly every worm gear in existence has what is called periodic error.  The worm gear in the RA axis rotates once every few minutes.  Mine has a 10 minute cycle.  This has an impact on tracking, guiding, and rms error.  Its best to watch the guide graph over a 10 minute period to make sure the aggression is high enough to correct for the error.  This has a limiting effect on the rms error value that you can obtain.  You may not be able to match the Dec rms error because your aggression must be increased to compensate.  To further complicate it, a long guide exposure may not be able to detect the start of this error soon enough to correct in time.  Its something  you have to watch for before you start imaging.  Understand how your RA works over the full cycle of the worm gear so you don't have any surprises. Better mounts have hand picked gears to minimize this effect. 

Some mounts have what is called PPEC or permanent periodic error correction.  This is a learned cycle of the worm and can be played back within the motor controller to correct for periodic error. It requires the hand controller to be operational.  However, PPEC is not compatible with INDI which is what is used on the ASIAir.  Some have had success but its believed to be a disaster waiting to happen.  I have witnessed conflicts using PPEC.  The issues are a matter of timing.  You can either get a double corrected guide pulse or a software glitch which turns tracking off in the hand controller.  Either way its a disaster to any imaging session.  Either use PPEC by itself or just use ASIAir guiding with PPEC disabled.  Just a note that there are some limited ASCOM drivers designed to allow both to work.  ASCOM requires a PC to control everything and at this time there are no INDI drivers that safely allow the use of PPEC with PHD2.

I hope this sheds some light on getting the aggression settings set properly.  With some practice you can quickly get these settings adjusted based on the current seeing conditions.  Just remember we are always seeing limited and one session to the next can have vastly different settings and results. Take some time to practice during full moon and eventually you will master these settings.
 


No comments:

Post a Comment

Note: Only a member of this blog may post a comment.