Getting the best performance from my AM5

StarzLite

Let me begin by saying a big THANK YOU to Chen!

Chen gets all of the credit for explaining how to tune a strain wave mount for optimum performance and I simply applied what Chen describes and the results for my mount are nothing short of incredible!

Chen, @w7ay gives a very detailed explanation in this thread. I highly recommend that you read this:
https://bbs.astronomy-imaging-camera.com/d/15711-very-large-backlash-in-new-am5/13

I received my AM5 about three months ago. The QC report seemed reasonable, although I really had no idea at the time what the important data was that the graph was showing. Chen explains all of that in the recommended post.

I was very happy with the AM5 performance initially. I'm using it with the ASIAIR+, ASI120mm mini, ASI2600MC pro, WO Redcat 51 and WO Uniguide 32. I started off using the default guide settings of the Air+ and the results were acceptable. I really didn't spend much time changing anything other than tuning the aggression settings and guide camera exposure time between 1 and 2 seconds. Typical settings I was using and their results looked like this.

As I spent more time using the mount for many hours of astrophotography sessions, I noticed intervals when the guiding numbers would become erratic. These rough spots where somewhat short in duration and didn't seem to effect data quality, but they also seemed periodic in nature. They would look something like this:

Looking around on the net, I found that many others were experiencing similar issues with this. Discussions covered the usual trouble shooting suggestions. Accurate polar alignment, seeing conditions, wind, calibration, tripod stability and vibration, etc. I was meticulous checking off all of the variables, but the erratic patches continued regardless. The periodic and repeatable nature of the behavior is what I kept coming back to. I concluded that there must be something mechanical about the mount that was causing the issue. I also figured that maybe it was something that was just luck of the draw. Maybe I should just accept what I was getting and be good with it. Was I ever wrong!

Browsing around here on the ZWO forum a few days ago I came across the post where Chen explains the idiosyncrasies of tuning a strain wave mount. I was skeptical, but optimistic to try something different. Besides, Chen clearly backed up every claim with solid facts, examples and even citations! I went out a few nights ago with no other agenda but to tinker with guiding accuracy. Working strictly from memory of a first reading of what to do, apparently I got it wrong. Nothing much changed. I decided to read everything again the next day very carefully. As the old saying goes, the devil is in the details. I decided to go out last night and try again, This time starting from scratch with settings that seemed so extreme, surely it would be comical. Spot On! I could not believe this was the same mount!

As you can see, I used settings that are quite different than default. I slewed to Procyon for the initial calibration because of its current proximity to just east of the meridian and along the equator. I initially set the aggression of each axis to start off at 5 (the lowest possible) and it only took a few seconds to dial in the numbers you see here. I guided here for an hour and it remained very stable. The total stayed between .2" and .5", averaging around .35"! Unfortunately, I didn't know that PHD logging on the Air+ only saves when using a plan or autorun, so I don't have that data to show the hour long run. I'll get more complete data the next time out.

I then slewed to three more general quadrants of the sky to check for consistency. I went west of the meridian along the equator, east near the north polar area and west near the north polar area. They all looked good and I made a few minor tweaks to the aggression settings along the way to dial in things a little better.

West of the meridian along the equator:

Northeast:

Northwest:

Finally I ended the test with a five pixel dither between each exposure to check settling repeatability. Again, I made some minor tweaks in aggression, but nothing major. Small changes in aggression settings can make a significant difference, so I found a slow and methodical approach here is beneficial.

I hope this information can help some others get the most from their AM5. I believe this mount has near unlimited potential within its designed capability and I'm very happy and impressed with mine. There are some pretty significant differences in the approach to tuning a strain wave mount and current software is not optimized for this technology. By having a better understanding of how this technology works, dialing in the best performance becomes more intuitive. Many thanks again to @w7ay for taking the time to explain the details behind the design, function and optimization of this amazing technology.

Dave

MrAstro

StarzLite
Happy for you 🙂
I dont use ASIair but I am using PHD2..
What exposure time did you use.. I couldn't figure it out from the ASIair screenshot ?
Is it 0.5 Sec. ?
Thanks

fotosgrafiabymiguel

@w7ay StarzLite Thank you Chen for your well-written description "getting the best performance on the AM5" I used your suggestions and got the best guiding on my AM5, AA+, RedCat 51, ASI 2600MC Pro and ASI220mm Guidescope.

Do you have any guiding suggestions for AM5, AA+, Celestron EdgeHD 8, OAG w/ASI174mm and ASI533MC Pro?

Thank you in advance.

Jimfmiller

StarzLite Thank you for the article. I am using PHD2 vs ASI Air. Why are there 2 screen shots for each of your guiding examples, with the
lower screenshot seemingly having better results?

Many thanks

Jim

raawe

I have to try this. Amazing performance. I am still adjusting to ASI Air+ and AM5 mount and even with all the default settings I am really satisfied with guiding. I have to note that after v2.1 my mount started to misbehave, and I had to restart it a couple of times to get things back to normal.
My AM5 report states PEC 15.5/6.3 Arc-secs, and here is a guide screenshot after some 3h on the target. It shifted from 0.3-0.5 during my 6h imaging session yesterday.

StarzLite

raawe

That looks very good. Your experience would suggest that the slope(rate of change) of the various harmonic curves of your mount may not be as steep verses other AM5 mounts that experience wider fluctuations in guiding accuracy. The amplitude(Max and Min numbers on that report) of the fundamental PE is not really of much importance with regard to guiding. The zoomed graph at the bottom of the report will give a better idea of predicting which mounts will be more or less challenging to guide. That graph depicts the various harmonic curves and most important, how steep(slope) the rate of change in those curves is over time. This is where more frequent sampling(guide camera exposure rate)/correction pulses can dramatically improve the guiding of certain AM5 units.

See @w7ay for a great explanation here:
https://bbs.astronomy-imaging-camera.com/d/15711-very-large-backlash-in-new-am5/16

Dave

w7ay

StarzLite That graph depicts the various harmonic curves and most important, how steep(slope) the rate of change in those curves is over time.

Dave, I have started to write up a white paper on guiding strain wave mounts. I want to make sure all the i's are dotted and t's are crossed before writing a single line of code in INDIGO.

Here are three figures to illustrate what you described. This first one for zero harmonic distortion (never going to happen with stain wave gears, I can assure you). I have given the fundamental periodic error (blue curve) a 50 arcsec p-p (25 arc sec amplitude).

The green curve is the first drivative with is ordinate scale on the right. This will in the final analysis (you will have to wait for the white paper) determine the minimum possible RMS error of the guiding -- in this case 0.37*0.707 = 0.26 arc sec/sec. So, 0.37 arcsec for 1 second guide exposure.

If you look at a high end mount, you will see a "smooth" curve, in which case, the smaller peak-to-peak periodic error curve wins.

However, higher harmonic terms could destroy that premise.

The above shows (again blue line) a periodic error that really does not have that much large harmonic terms (that I also scaled to 50 arcsec p-p, to be fair. But look at how ill behaved the derivative has become. The peak of the first derivative is now more than twice the one of the "smooth" mount.

Here is a worse case (but still nowhere near some of the curves ZWO published as representative of their mounts):

Notice that the peak derivative is now over 1 arcsec. Good luck guiding such a mount. It may work fine most of the time, until you hit that part of the hour angle where the first derivative shoots up.

Especially with traditional guiding methods where you "tune" the parameters to match the mount. The mount does not even behave the same from one 200 second to the next 200 second.

Don't worry, there is a closed form solution to all of this.

For now, just use a ruler-and-pencil way to estimate the worst case slope. There is no way to beat that; so no need to bang your head against the wall to try to get better guiding with the mount that your have.

Chen

w7ay

By the way, the above shows that even the worst mounts possible don't need a long correction pulse.

Even in the case where the slope of the periodic error curve is 1 arcsec/sec, you will need less than 1/7 second pulse to correct it. I.e., assume that you choose 0.5x sidereal rate as the guide rate, that is equavalent to 7.5 arcsec/sec. To correct a 1 arcsec error, you would need only a 133 ms pulse, not anywhere close to a 2000 ms pulse.

Any max pulse that is longer will just cause problems (attempts to correct errors that are not caused by periodic error curve).

Slow mechanical flexure can cause some error, and so can atmospheric refraction if the mount does not support King tracking rate. But neither are large enough to cause harm (any large mechanical flexure should be fixed independently, in any case). Most mounts support King rate, but since ASIAIR does not support KIng rate, I don't know if ZWO's mount does:

https://canburytech.net/DriftAlign/DriftAlign_3.html

Chen

StarzLite

w7ay
The 150ms max is working fine for me. It was exactly what you said!

Those additional illustrations above really get your attention. The magnitude of the accumulative peak when several of the harmonics all cycle in to sync is impressive. Thanks for the additional discussion.

It would be really great to see this integrated in to guiding software as a strain wave mount option/driver or however that would work. Keep at it, this is really appreciated.

Dave

w7ay

StarzLite The magnitude of the accumulative peak when several of the harmonics all cycle in to sync is impressive.

Yep Dave, I had also seen that and written this in the white paper (of course, a little bit more "math jargon" than what I would post on this forum :-):

Notice that the derivative has an "N" term in the multiplying factor. I.e., the firth harmonic term creates 5 times the problem from the fundamental. Fortunately, the fifth harmonic distortion of most gears are also down by a lot -- it is up to QC to reject the high harmonic content ones (everything boils down to QC).

Here is a close up look at the dynamics of one guide frame that has a single pulse per exposure:

In the part shown here, the periodic error always exert a positive "pressure," even when there is a pulse. The pulse just happens to have a much larger negative slope (of the order of 7.5 arcsec/sec, vs the periodic error (of the order of 0.1 to 1 arcsec/sec).

When the pulse is running, the actual telecope is moved downwards (in this case) at a slope of the slope of the periodic error - the slope of the guide rate (7.5 arcsec/sec). Once you let the foot off the pulse, the telescope now moves again at its periodic error rate, for the rest of the exposure frame.

Notice that the guide star is also making the identical movement. I.e., during a 1 second exposure, the guide star will move up and down between 0.1 and 1 arcsec (so the centroid computing part had better be ready to handle that). For a 0.5 second exposure, the problem is halved. Thus the need to use the shortest exposure possible.

At some point, you can no longer use shorter exposures (unless you use larger aperture guide scopes). But here is where the N-pulse per exposure come in. Instead of ptting all the correction above at the start, we apply only half of it at the beginning, and apply a second one at the center of that exposure. The sawtooth doubles in frequency, while the sawtooth amplitude halves! If the slope does not change that rapidly (I have the also shown the second derivative in my white paper to show that it is indeed small) this is equivalent to halving the exposure time, without having to actually shorten the exposure time.

Even better, use 4 pulses, or 8 pulses, etc. Each doubnling of pulses will half the amplitude of the corrected curve.

The penultimate solution is to use a single pulse, whose durration is the same as the exposure time, and whose slope is the exact opposite of the mount's slope! The amplitude of that green curve above collapses to zero (or somehing that is very small). That makes a strain wave mount as smooth as any worm geared one. Not all mount protocols can do that (control the guide rate to such fine grain). My RST-135 can control the guide rate from 0.01 x sidereal rate to 0.99 x sidereal rate, so I have at least something to experiment with in the future.

FWIW, I have been using a Borg 55FL objective (55mm aperture,250 mm focal length) as my guide scope. Seems adequate for use with a camera like the ASI178MM for 1/2 second exposures. But if I can go to multiple pulses, I can perhaps start using 1 second or 2 second guide exposures and move back to as smaller, say 40 mm guide scope.

Chen

raawe

@StarzLite Just wondering did you do anything to the guide rate of the mount or is it still set to the default 0.5x sidereal? Also, what gain/Bin do you have for guide cam and how many pixels did you set in dither? I have to test everything for my setup because I'm using OAG-L with 290MM Mini. Hopefully I'll get enough data for 0.5sec exposures. My previous post with 0.39 total RMS was with 2 sec exposures on guide cam.

Going over all this, I am actually starting to understand these graphs and explanations are spot on. This got me thinking about guiding software, and it looks like we need to have a variable Max RA/DEC duration rates and RA/DEC aggression percentages depending on mount and what part of gear train is currently locked in (in the full 360 degree). The program could analyze RA/DEC gears and see what kind of periodic errors are presented, then adjust the values accordingly.

w7ay Would this avoid the unavoidable guide spikes?

w7ay

raawe Would this avoid the unavoidable guide spikes?

The only way to avoid the quide star from moving is to apply that last paradigm which I mentioned above -- ie, to guide by using different guide rate (Pulse Amplitude Modulation; if you are an EE) instead of pulse duration (Pulse Width Modulation) as we do today. I.e., if the correction slope (we come back to slopes again :-) is the same as the periodic error's slope, so they cancel each other out, and the guide star becones stationary -- in practice, the slope is not exact, and you will notice the slight movement, and apply a slightly different guide rate in the next frame, etc.

It is a large leap, and I don't know how many mount protocols can accept variable guide rates. However, that second paradigm, which is to issue multiple distributed shorter pulses, whose accumulated pulse duration is equal to that of the original pulse. Think of it as PWM at a higher pulse rate.

This would reduce the movement of the guide star (lower amplitude sawtooths), but not eliminate it.

Once you get used to the appearance of that sawtooth, it will become obvious that it is impossible to avoid the guide star from tracing a line while a guide exposure is being taken. Once you grok that, the solution(s) become obvious.

Whatt you descibed if a good way, too. I.e., changing the guide dynamics. Albeit, requires the periodic error curve for the entire revolution of the RA axis -- about 24 hours worth of recorded periodic error curve.

Chen

StarzLite

raawe
I did not change the guide rate. As you said, I believe the default is set to 0.5x sidereal.

The guide camera is the ASI120mm mini. The gain is set to High(48) and Bin1(Bin2 is off) mode.

Exposure is set to .5sec(2FPS).

Good luck with tweaking in a little more guiding precision.

Dave

StarzLite

MrAstro

Yes, ASI120mm mini set to High gain(48) and .5sec exposure.

Dave

StarzLite

tempus

I haven't used PHD2, but when you say "complaining", are those warnings something you can disregard and still proceed with using the settings you would like? If so, I can see no harm in experimenting with the settings we are discussing.

The entire premise of what w7ay is presenting here is that the current guiding software platforms are not optimized for use with strain wave mounts, so the settings discussed here are a work around.

Dave

tempus

I use Phd2. When I try to lower the Max RA Rate and Max Dec Rate below 1500ms, Phd2 keeps complaining that those values are too low to make the needed corrections. Especially after a dither. I can't see Phd2 working well at the low values suggested here. What am I missing?

mikenoname

I think tempus meant max duration not max rate because I am getting the same error messages.

I got out last night briefly and tried the 150ms max duration settings, lower aggression and 0.5s guide pulses and did not see much improvement. In fact, in some cases it was worse. PHD also intermittently complained that with the duration set so low it could not keep the guide star in the FOV.

Only an hour of play time is not enough to draw any conclusions, so I am looking forward to the clouds going away again...

w7ay

@mikenoname

What is the worst case slope on the periodic error curve that ZWO included with your mount?

From the first posting in this thread, Dave has one of the better ZWO mounts out there, with magnitude of slopes similar to my RST-135. My RST-135 has larger p-p periodic error, but smaller harmonic distortion. That is why I jumped into this discussion with more "color."

As in everything in this hobby, don't blindly use some other person's number. Measure that slope carefully and then set up the max pulse duration. If you have one of the worse ZWO mounts, as I mentioned earlier, there is nothing you can do.

For example, if the slope is over 1 arcsec/sec and you are using 1 second exposures, then the RMS guding error (just of the RA axis) could never be always less than .35 arcsec. If the slope is worse than that, the RMS error will also be worse.

The only thing you can do is to reduce the exposure time to 0.5 sec by using a reasonably sized guide scope. And if the slope is still too large to match your plate scale, then you will just need to move to a different mount.

By the way, also check where your error is coming from, the RA axis or the Declination axis. There were some postings in this forum complaining about the large backlash in their ZWO mounts (the strain wave gears do not have much backlash, but the rest of the mechanical design can add backlash).

If you are driving the guide rate at 0.5x sidereal rate, you can never get a backlash in the RA axis (except when you dither)since the RA axis is being tracked at 1.0x sidereal. A 0.5x sidereal pulse will only change the track from 0.5x sidereal to 1.5 x sidereal, and since both are positive numbers, the pulses can never produce a backlash.

The story is a little different in the declination axis. If you are seeing a large declination guiding error, introduce some polar alignment error (that's right, you read that correctly) so that there is a declination drift rate that is larger then the slope of your declination axis. Unfortunately, ZWO does not publish curves for them, so you will have to use some trial an error. If you remember, a polar alignment error can be measured by watching the drift of a star in declination (check out "drift alignment" on the web). So, just add enough polar alignment error so that the declination drift is greater than the curve of your declination gears. Your exposure duration may be limited by the field rotation, but that is what one has to sacrifice when there is lots of error in the mount.

Chen

mikenoname

w7ay

I seem to remember you detailing how to determine the slope of the curve from ZWO's diagram, but can't seem to locate it now. Can you point me to it?

Mike

w7ay

mikenoname I seem to remember you detailing how to determine the slope of the curve from ZWO's diagram

Place a ruler at the steepest part of the curve. The ordinate of the chart gives the vertical scale in arcseconds. The horizontal scale can be approximated (as units of seconds) by assuming that the fundamental period is about 400 seconds long.

The ruler line then gives you units of arcsecond per second.

Chen

mikenoname

w7ay

Don't know what 'fundamental period' means in this context. Since the mount rotates 360 deg in one day ideally, then one degree is 24*3600/360, so basically 240 seconds. From my zoomed in section (hopefully my chart will upload below) from the first peak to the bottom of the next trough looks like ₁.9 degrees to me (30.7-31.6). Or should I just be measuring one individual short slope?

In any event, by my calculations that comes out to be 456 seconds. The peak looks to me to be ~+6.9 and the trough at -7.7 for a difference of 13.6 arcsec.

Therefore my resulting number would be 13.6/456 or 0.03 arcsec/sec correction required.

Am I close? [Waits anxiously for his paper to be graded]

StarzLite

mikenoname
This will give you a starting point.
It's an example where w7ay did a worst case slope calculation on @Kevin_A's report:

https://bbs.astronomy-imaging-camera.com/d/15711-very-large-backlash-in-new-am5/19

Dave

mikenoname

StarzLite Thanks Dave. I knew I had seen him post it somewhere but I couldn't find it again when searching.

Mike