HEM15 and PHD2 Guiding Settings

[Ben Schafer]

Hi Guys,

I"m reasonably experienced with PHD2 with my Mach 1 mount, but I just got an HEM15 and I'm trying to figure out the best settings.

Do you think Predictive PEC would work with this mount?

I've tried a lot of different settings in terms of guide exposure and minmo, but I'm sure I'm limited by my usually bad seeing.

Are you all working on PPEC for the Dec axis too for Harmonic Drive mounts? :-)

I can get the RMS down under one arcsecond, but I can't seem to reduce the peak-to-peak errors, and they are causing elongated stars on the fast part of the PE curve.

Any help would be greatly appreciated!

[bw_m...@earthlink.net]

Hi Ben.  It would be better if you would include both the guide and debug logs so we can do a more thorough job.  Please use the built-in PHD2 Upload Log Files feature, explained here:

 

https://openphdguiding.org/getting-help/

 

To answer your other question, there’s no periodic error to worry about in Dec because the Dec motor and axis are mostly idle during guiding and tracking.  There are only intermittent very small moves in response to guide commands.  RA, of course, is a different matter because the RA drive system is running continuously in order for the scope to track the sidereal movement of the sky.

 

Bruce

--
You received this message because you are subscribed to the Google Groups "Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guidi...@googlegroups.com.
To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/d3c090c7-4e21-47b7-943d-cf03e740facen%40googlegroups.com.

[bw_m...@earthlink.net]

Also, could you please provide some details on the gear you have riding on the mount?  Are you really guiding with something that has only a 100mm focal length?  The guider image scale is really coarse if that’s correct.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/000f01da3109%24f9d1e4c0%24ed75ae40%24%40earthlink.net.

[Ben Schafer]

Hi Bruce!

Here are the logs, I didn't see a place to upload the debug file so I'm assuming it's in the zip file with the guidelog:

https://openphdguiding.org/logs/dl/PHD2_logs_hc5X.zip

My setup:

Mount: iOptron HEM15

Guidescope: Kowa  36mm of aperture, 100mm of focal length, and a focal ratio of f/2.8.  SBIG used to sell this cope with the STi autoguider and Michael Barber says it can guide down to one arcsecond with the STi, which had 7 micron pixels I think.  

Guide Camera:  QHY5LIIM camera with 3.75 micron pixels in a 1280 x 960 array that is 4.8mm x 3.6mm in size.

Imaging Scope: Sony 200-600mm zoom at 600mm at f/6.3.

Imaging Camera: Sony A7MIII

I thought I read somewhere that the DEC has periodic error from some type of mechanical wobble in the RA gear. I see this crazy dec movement even when I had dec guiding turned off. 

Do you think PHD2's predective PEC would work with this mount?

Thanks for helping!

Ben

[Bruce Waddington]

On the second page of the Upload UI, it will tell you what logs are being uploaded.  It will always include the debug log file unless you’ve moved it or deleted it.  Your upload included only the guide log.

It may be possible in some circumstances to get acceptable guiding with these coarse image scales, and maybe 1 arc-sec RMS is good enough for you. It assumes, of course, that you are working with stellar SNR values that will routinely allow centroid accuracy of 0.1 px or better.  What you may not be thinking about is the mechanical requirements it places on the guiding assembly.  In your case, a 3.8 micron shift of the guide camera is going to create a 7.7 arc-sec guiding excursion.  That’s 1/10 the thickness of a human hair.  Is that little finder-scope mounting arrangement and associated guide cables going to meet that requirement?  Here's a look at a typical guide session in your log (green is Dec):

There's very little difference here between RA and Dec even though the Dec motor is idle about 90% of the time.  You can see there are a lot of big excursions of 3-4 arc-sec on both axes, including the mostly-idle Dec axis.  I have never heard that this mount has problems with mechanical cross-talk between the axes so that would be a question you should direct to iOptron.  All of this unwanted movement could be due to seeing if it's really poor; or to wind, vibration, etc.  But it could also be caused by a lack of rigidity in the guiding assembly, remembering that a 4 arc-sec move here will be caused by a 2 micron movement of the guide camera.  If you want to isolate possible sources of this movement, you can run some test sessions guiding through the main scope.  You can always try using PPEC for RA but I really doubt that it will help much.  One of the common problems with these strain-wave mounts is that their tracking errors aren't truly periodic so PEC models often don't work very well.  I think that's one reason manufacturers of these kinds of mounts offer some kind of absolute encoder option in RA.

Regards,

Bruce

[Ben Schafer]

Hi Bruce,

I don't know why the debog log was not available on the second page up the upload dialog, and I don't know if you want to see it anymore, but here it is:

https://drive.google.com/file/d/1dp_oiryspqGZvAVsZTwPCjkMjGGRHtFi/view?usp=sharing

My guidescope was very well mounted, but I will use it in my refractor next time so there is no question. I don't think it was wind or cables or anything like that.

If the Dec movement is coupled to the RA, what settings do you think might work to reduce this problem?

I don't know how to combine really short sub-millisecond corrections for the peak-to-peak excursions while wanting to use longer exposures to average out the seeing.

Thanks,

Ben

[Bruce Waddington]

You can't do "sub-millisecond" exposures with PHD2.  Going below 500 ms isn't recommended for many reasons - you will chase the seeing and the SNR of whatever guide stars you have will drop linearly.  I notice now that your guide star sizes are huge which indicates that the guide camera isn't well-focused.  I really doubt that there's any evidence of mechanical cross-talk between axes, but if there is that's a problem that iOptron needs to address.  There's no magic elixir for that via guiding so you would just leave the guiding parameters alone.  I think at this point you and I have very different views of your situation.  It sounds like you're convinced you have exotic problems that demand equally exotic guiding solutions and I'm strongly inclined to think you have more basic problems just like hundreds of others before you have dealt with.  Perhaps running the guiding tests through the main scope will help to clarify things.  By the way, the debug log file shows numerous cases where you lost connection with the guide camera so that's another complication to the puzzle.

Good luck,

Bruce

[Ben Schafer]

Hi Bruce,

I'm sorry I made a mistake. I meant to say sub-second exposures, something under one second, something I've never done before so I didn't chase the seeing, but that has been suggested by iOptron and some others who have iOptron strain wave gears to try to stay on top of the fast parts of the PEC curve.

I don't think I have exotic problems. I don't know what kind of problems I have. I'm trying to figure that out, which is why I really appreciate you trying to help me.

I would love nothing more for that dec behavior to be a loose guidecamera, or the wind, or those crazy excursions in RA and Dec to be some operator error and not some inherent design property of the gear.  

It's just that I have read in multiple places that these harmonic gears have strange periodic error and that the dec ends up with the same periodic error.  I believe that is what I was seeing, but you seem to be convinced otherwise and because I value your expertise, I'm not dismissing it.

I know you have seen these kinds of problems a million times before and they are in fact caused by something being loose, or operator error, so I understand why that is what you always think it is, because it usually is. I would like maybe a tiny benefit of the doubt that I'm not a complete moron though.

I'm going to work hard to try to see if there is a problem or if it is operator error. God knows I've made enough mistakes in life and astrophotography before. 

I will revert back to the stock settings, and use the suggestions from the guide wizard.

Sorry for the bother, and thanks for your help.

Ben.

[Bruce Waddington]

It isn't a bother and I'm not suggesting you're crazy or that you can't possibly be right.  I just think it makes sense to go at this systematically and not immediately assume you have some kind of outlier situation, especially when there are other things in your setup that look dodgy.  If you have heard from iOptron that they have mechanical cross-talk problems, that would be interesting although there would be little we could do about it.  Hearing that from end-user forums might need to be treated with a degree of skepticism.  We have seen quite a bit of data from other strain-wave mounts and yes, people often need to use a fairly fast cadence (e.g. 500ms) to stay on top of what is often poor sidereal tracking.  With multi-star guiding, that is not such a problem and many of these setups enjoy reasonable guiding in the 0.6 to 1 arc-sec range.  If you want to experiment with a different Dec guiding algorithm, you could try Hysteresis, the algorithm usually used for RA.  Since your mount presumably has little or no backlash, there shouldn't be any penalty in trying that.  But there isn't going to be any solution to guide star deflections of multiple arc-sec that occur at intervals of 0.5 to 1.5 secs.

[Ben Schafer]

Hi Bruce,

You are absolutely correct that I need to approach this from the basics first.

I'm going to check everything else first before blaming the mount. With me, it's usually operator error.

I'm going to put the guidescope in my refractor and carefully focus it.

Should the dec backlash be visable visually in ahigh-power eyepiece? Would it just look like slop if I push the scope with my hand?  Or is it only after a reversal? That should be visible in a highpower eyepiece also, shouldn't it?

Thanks again for your help!

Ben

[Ben Schafer]

I forgot to ask, how would this backlash be apparent visually? 

Can you see it in a high-power eyepiece?

Does it look like slop where the field will move if you just push the scope with your hand, and then move back and forth in dec?

Or does it only show up after a direction reversal? Shouldn't that be visible visually in a high-power eyepiece?

Thanks again,

[bw_m...@earthlink.net]

Run the Guiding Assistant, let it measure the Dec backlash.  There shouldn’t be any to speak of.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/1051e7ab-07c1-47e8-98a9-131411229a87n%40googlegroups.com.

[Bruce Waddington]

Also, if you're going to guide through the refractor - a good idea - don't forget to run the new-profile-wizard for the new configuration.  Otherwise, it will really be a mess.

[Ben Schafer]

Hi Bruce,

I had 5 clear hours last night with decent seeing and I did a bunch of tests.

Both logs are here:

https://openphdguiding.org/logs/dl/PHD2_logs_pnki.zip

The backlash is terrible.

I did get the rms down to around 0.5 at one point. I think that is about as good as it's going to get.  I got that low with both Hysteresis and PPEC in RA. 

Here is my setup: AT65mm refractor with 65mm of aperture, 422mm of focal length at f/6.5 on top of the HEM15 mount, on a Losmandy tripod.

Why do you think there is so much movement in dec even when dec guiding is turned off or in one direction?  There was little to no wind, all cables were secure, etc.

Thanks!

Ben

[Ben Schafer]

I forgot to mention, I wasn't on the meridian and celestial equator, I'm blocked by a big tree due south.

And I wanted to let PPEC run a lot longer, but got clouded out.

Ben

[Bruce Waddington]

Hi Ben.  I think guiding through the main scope makes things a bit more clear at least for me.  If we look at the Dec behavior you're asking about, here's a typical view without guiding (Dec in green):

Unless there is some very odd vibration in your setup like from a fan - which I doubt - you are looking at seeing effects.  It is made worse because you are using 500ms exposures to try and tame the problems with RA tracking.  You mentioned that you thought the seeing was "decent" but I don't know what means on an absolute scale.  Are you looking at seeing forecasts?  What general area do you live in?  Anyway, it is what it is and this is a good example of why we tell most people to use longer exposures.  You can see that this puts a lot of pressure on settings like min-move and the damping terms in the guiding algorithms to avoid chasing these seeing deflections with guide commands.

 I see that you let the Guiding Assistant measure the Dec backlash and it was pretty bad.  I didn't understand that this mount doesn't use a strain-wave drive for Dec, so yes, you are very much in the camp of many imagers who have to deal with large Dec backlash.  One option might be to increase the guide speed in Dec because that should produce a proportional decrease in this reversal-delay time assuming it is pure backlash and not something affected by stiction.  For example, if you increased the Dec guide speed from 0.5x to 0.9x sidereal, you might reduce the reversal delay time by nearly 1/2.  That would probably put you down in a region where the PHD2 adaptive Dec backlash compensation would help you out.  I notice that you experimented with uni-directional guiding but you ended up not doing any Dec guiding at all for most of the sessions.  You should read the section in the User Guide that explains how this works, noting that the direction of the needed corrections changes with sky location.

https://openphdguiding.org/man-dev/Supplemental_Info.htm#Declination-Uni-directional_guiding

This most recent look of the RA tracking unfortunately paints a somewhat grim picture.  Here is an FFT analysis showing the periods of the various tracking errors in RA (RA in red):

The problem children here are the two peaks to the left.  You have a 1.3 arc-sec peak-peak error occurring every 2.4 sec and a 1.4 arc-sec peak occurring every 17 sec.  These are too rapid for guiding to handle very well and they unfortunately place a limit on the guiding results you can expect.  I really think you should ask iOptron what can be done about them.  The tracking error component at the right is about 3.4 arc-sec peak-peak and occurs every 120 sec.  This one is more amenable to guiding and you could set the PPEC algorithm to use a period length of 120 sec and disable 'Auto-adjust period'. 

I noticed that you weren't using the new Calibration Assistant (Tools menu) despite upgrading to 2.6.12.  You should always do that going forward because it will do a better job of clearing your Dec backlash and reducing its effect on the calibration.  Since you have sky obstructions you can use the CA feature for setting a custom sky location for calibration and then use that.  Your goal should be to get as close to Dec=0 as you can while worrying less about how far from the central meridian you have to be - just don't let the scope start pointing down close to the horizons.  Once you get a calibration that is at least 'acceptable', keep using it - don't re-calibrate over and over.

Good luck,

Bruce

[Ben Schafer]

Bruce, Thank you for taking the time to do this analysis, I really appreciate it.

No fans blowing, no heavy equipment rumbling, no major traffic nearby. I always have mediocre to bad seeing here. I'm in New Jersey suburbs, just across the river from Philadelphia. Our seeing here is usually bad because of the jet stream being overhead about 80% of the time.  MeteoBlue's seeing prediction for my area was for under 2 arcsec that night while I was testing, which was really good for my area.

So I'm faced with the fundamental dilemma of using very short exposures to try to correct for these multiple fast tracking errors,  but needing a long exposure because of the bad seeing. 

How much does multiple star guiding help with the seeing? Not enough to help here I'm guessing? Has it been quantified how much multi-star guiding helps? I'm curious.

I will increase the dec guide speed to 0.9x as you suggest. When you say "adaptive dec backlash" I'm not sure which specific settings control this. I don't recall seeing a PPEC for dec?   

You said the mount was "not doing any Dec guiding at all for most of the sessions."  I am puzzled by this. I thought I was. I remember turning Dec guiding off for one experiment, but I thought it was on for the rest. Sounds like operator error.

I will definitely talk to iOptron after I stop making all these stupid OPIE errors (Old People's Intellectual Errors). 

I think I ran the calibration assistant, but if you see I didn't in the log, it's operator error again.

This is getting embarrassing. I'm going to go back and re-read all the documentation. I even took notes the first time. I guess I just need to follow them instead of thinking I can remember them. 

If I just switch to 3 or 4-second guiding exposures and ignore the 1.3" and 2.4" errors,  will this damage the final RMS too much if the seeing is 3 arcseconds in a 422mm scope that can resolve about 2 arcseconds?  I think iOptron is recommending short guiding exposures for these mounts.

Thank you, Bruce; I will certainly incorporate all of your suggestions, try to learn from my mistakes, and report back.

Ben

[Bruce Waddington]

1.  Multi-star guiding does help to keep PHD2 from chasing seeing effects but it can only do so much.  Other strain-wave mount users have done ok with 0.5 sec exposures but I wouldn't go below that.

2.  Read the User Guide on PHD2 Dec backlash compensation.  Once you've changed the Dec guide speed, let the GA measure your backlash again.  If the GA doesn't offer to set the Dec backlash compensation for you, you can try it manually.  On the algorithms tab to the right under Dec, there are controls for Dec backlash compensation.  The amount is in units of milliseconds so you might try 2000 and click on 'Enable'. 

3.  Your mount wasn't guiding in Dec for the most part because you chose the wrong guide direction.  The mount was drifting in the opposite direction to what it was doing when you ran the GA because the pointing position had changed.  Watch the guiding graph when guiding is active - if the Dec is drifting steadily away from the x-axis and there are no Dec guide pulses being generated, you have probably chosen the wrong Dec guiding direction.  But hopefully you won't have to fool with this if the increase in Dec guide speed works.

I can't estimate how much the short-period RA tracking errors will affect your imaging results.  You can just try it and see how you do, perhaps by keeping the main camera exposures reasonably short.  I don't see why you're hesitating to talk to iOptron - I think the frequency/period and backlash graphs  pretty much tell the story.  The reality is that you ended up with a mount that has tracking problems, everything else is pretty much just trying to work around them or mitigate them a bit.

Bruce

[Ben Schafer]

Hi Bruce,

I hope you had a good holiday. 

Thank you for all your help. It's been cloudy here so I haven't had a chance to get back out for testing.

I intend to talk to iOptron, but I want to understand what I'm talking about when I complain about the mount, and I don't feel like I completely understand yet (my fault, not yours! :-).

Have you seen a lot of data from other strain-wave mounts? I know the dec backlash has been around on the other mounts that they use a belt drive on, I'm not that worried about the dec.

Have you seen other mounts like this that are well-behaved? 

Thanks!

Ben

[Brian Valente]

Hi Ben

The relevant information to share with Ioptron is what Bruce wrote below.

There's not a lot more to that story that IOptron needs to know. I suggest you share the picture below along with Bruce's comments 

Those are rather large errors and unfortunately of such a short period that it is difficult or impossible to guide them out. Those are the periods you want to bring to their attention

The problem children here are the two peaks to the left.  You have a 1.3 arc-sec peak-peak error occurring every 2.4 sec and a 1.4 arc-sec peak occurring every 17 sec.  These are too rapid for guiding to handle very well and they unfortunately place a limit on the guiding results you can expect.  I really think you should ask iOptron what can be done about them.  The tracking error component at the right is about 3.4 arc-sec peak-peak and occurs every 120 sec.  This one is more amenable to guiding and you could set the PPEC algorithm to use a period length of 120 sec and disable 'Auto-adjust period'. 

--
You received this message because you are subscribed to the Google Groups "Open PHD Guiding" group.
To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guidi...@googlegroups.com.

To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/310e3b6b-cc28-487d-9066-381c78b32fadn%40googlegroups.com.

--

Brian 

Brian Valente

astronomy portfolio https://www.astrobin.com/users/bvalente/ 

nature portfolio brianvalentephotography.com 

other astronomy https://www.brianvalentephotography.com/astrophotography/ 

[Peter Rusling]

Hi, I have a HEM15 and also have RAW RA peaks at 2.4s and 17.1s although the associated peak-peak values are less at 0.1" and 0.5" respectively. At c. 120s I have 0.8" and at c. 360s (the mount's period I believe) I have 3.5". This was over a 100 minute guide run during which the total RMS was 0.83" (OK at my pixel scale). I can provide any further info that would be useful. Hope you get this resolved; please keep us informed.

[Ben Schafer]

Thanks for that information. I will let you know what iOptron says.

[Ben Schafer]

I had a couple of clear hours last night, so I played around some more.

Logs are here:          

https://openphdguiding.org/logs/dl/PHD2_logs_D3gh.zip

The first run was with the PHD defaults, except I lowered the aggression. Exposure was 500ms. I also reduced the maxpulse on RA to 200ms and changed the Dec maxpulse to 1504ms. My theory was to keep both maxpulses to less than 200ms each because of the 500ms exposure time, but I let the dec go with the recommended backlash compensation from the guiding assistant. Run one didn't look too bad, what do you think?

Run two was 75 minutes with PPEC and period length set to 120s as suggested by Bruce.

The third run was 13 minutes with the Z-Filter on both axes and then the clouds came.

The first run seemed to produce the best results, but I will let those with more knowledge than me interpret the results.

The fundamentals looked considerably different on PPEC. Why would that be me?

Note that I have planes that fly through my images here all the time, so some craziness in guiding is occasionally undoubtedly caused by them.

Thanks!

Ben

[Brian Valente]

Hi Ben

How did your discussion with iOptron go?

To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/50d9186c-4027-4e64-9538-e7cd2630a01dn%40googlegroups.com.

[Ben Schafer]

After the holidays.

[Brian Valente]

I think your logs confirm everything discussed previously

Your first run is the best because PPEC was targeting the primary PE and doing a good job of reducing it

However, your very short period (17 seconds as measured in your first run) remans the primary constraint to good guiding .RA is nearly 2x Dec, so I would anticipate star elongation, how visible that is depends on the image scale of your imaging system

So.... pretty much the same recommendation as before

To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/b7b7d1a2-a328-4e6e-bea9-cd3804498fd6n%40googlegroups.com.

[Brian Valente]

I would also add your Declination backlash is still problematic, not sure why you do not have auto dec backlash compensation enabled

You can see issues of Dec having a rapid jump in the first set of arrows, and once it gets back on track, quite a bit of time to get it reversed when it wanders off.  That's about 220 seconds to reverse dec in 500ms exposure increments

Fiddling with max pulses is not going to have the effect you want. If it needs to turn it around quickly, you've just forced it out by quite a number of attempts, as you can see

Planes satellites etc. flying through will not have any significant effect on guiding

[Ben Schafer]

Hi Peter,

Thank you very much for the information on your mount, it is helpful.

Ben

On Thursday, December 28, 2023 at 8:13:56 AM UTC-5 peter....@gmail.com wrote:

[Ben Schafer]

Hi Brian,

The first run was not PPEC, it was Hysteresis on RA and Resist Switch on Dec.

The 1 hour 16 minute run was PPEC.

But the fundamental chart you posted is from the first run.

I get it though, the problems are still there.

Thank you for your help.  I just have to call iOptron now.

Ben

[Ben Schafer]

Hi Brian,

This Dec graph is from the PPEC run of 1 hour 16 minutes. 

I'm not sure why I switched off the dec auto backlash. I think I was thinking it was longer than the exposure time, so lets try one run with it off.   And now I know. Not good. 

What does PHD2 do when a correction pulse is longer than the exposure and the maxpulse is set to allow it? Does it just delay the next exposure, or is it exposing 500ms every frame and cutting off corrections?

These are not tiny little planes at high altitude flying through the frame. I'm under the flight path for planes landing at an International airport about 10 miles away. When these commercial airliners fly over me, they are at an altitude of about 3,000 feet and fly directly through the frame occasionally and their jet wash trashes the seeing for a little while.

Why are the fundamentals for the PPEC so different than the others?

Thanks for your help, Brian, I really appreciate it.

Ben

[Brian Valente]

Hi Ben

i'm glad you wrote back

I went back and re-examined the runs, particularly the one with ppec active

Regarding airplane flights, given your description it's possible that would account for the few lost star events but i does not appear to impact your guiding overall. It will probably impact your imaging results far more given your description of the wash. 

The overall guiding results actually look pretty good. total rms 0.87" with (importantly) ra and dec about equal. 

You never answered my question about your imaging setup image scale, but if you are 1"/pix or above, this should produce acceptable-to-good results, notwithstanding the plane traffic

So at a bigger picture, maybe this is a reasonable guiding approach that will  produce good results.

 Looking at the total 76 minute run, you can see the basic challenge with this mount:

1. RA (in blue) rapidly oscillates on that 17 second period, and Dec (red) wanders about due to backlash. This is a noisy mount with a lot of dec backlash.

RA also has ocassional big jumps of 3" or more (green arrows), they are instant and PHD does a good job correcting for these. but it still is just not good. RA is just a very noisy result. Some of that is obviously seeing-related when you are guiding at such fast rates but there is also a cyclical nature to these that points to the mount (note the rise and fall

>>>What does PHD2 do when a correction pulse is longer than the exposure and the maxpulse is set to allow it? Does it just delay the next exposure, or is it exposing 500ms every frame and cutting off corrections?

It will make the correction and take as long as is needed. so if it's a 3 second correction, it will take 3 seconds before the next exposure. You have a bit of a 'pick your poison' decision. On one hand, if reversing dec takes 3 seconds, it will need that amount of time to clear the backlash. If you decide you don't want to use backlash and let it correct over time, it will take longer (due to overhead of guiding) and your dec will wander about more. There may be some middle ground you can find that helps clear it more quickly but maxes out at some point. you can try experimenting by setting the min and max for auto backlash compensation

To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/8ccbed2f-f710-433f-9ce5-aec6b401798bn%40googlegroups.com.

[Ben Schafer]

I got the mount back from servicing at iOptron.

They reduced the Dec backlash from about 5 seconds (32 acrsec) to about 1.5 seconds (9.5 arcsec).

They did not do anything to the Right Ascension.

They said the error at 2.4 seconds was within the margin of error, and the 17 second one was typical of strain wave gears.

Here is the service log.

1. DEC large backlash: fixed by adjusting end cap

2. DEC axle adjusted

3. RA stopper checked: ok

4. Firmware upgraded

5. Wifi: ok,

6. iPolar: ok

7. Guiding emulation: ok