PHD2 backlash measure test failed: "mount never established consistent south moves:test failed"

243 views
Skip to first unread message

andrea arbizzi

unread,
Feb 29, 2020, 3:58:01 AM2/29/20
to Open PHD Guiding

Hi.

My new Cem40 works like a charm for some nights and during my past nights tests i always did guiding assistant and measured backlash almost always: just once it failed but in generally guiding assistant did a complete measurements of my backlash.


I didn't change anything important except some parameters just to experiment a bit: since 2 nights guiding assistant isn't able to finish a backlash test, it always abort it saying "mount never established consistent south moves:test failed". I reset everything, did 3 new calibrations: didn't help. Tried to delete phd profile, i created a new profile, did a new calibration, did GA twice: same result, backlash test failed.


So backlash test was completed correctly for some nights, but last couple of nights it always failed.

 

This happened with my previous Cem25 too: for the first times the backlash test worked, after that it no longer has been completed..


Attached you can find:

- example of backlash test completed

- example of backlash failed

- debug file of backlash failed test: https://drive.google.com/file/d/1MZdhgAxHohtD2O6ZYP0MMpvdKvzbHHx9/view

- log file of backlash failed test: https://drive.google.com/file/d/1no_K7LHtTZtpvj-kF6v5Fyb9mUmqbLWA/view


Thanks for your help


Andrea

Immagine.jpg
Immagine1.jpg

andrea arbizzi

unread,
Feb 29, 2020, 7:52:09 AM2/29/20
to Open PHD Guiding
Attached, how everything is connected...

cable.jpg

Brian Valente

unread,
Feb 29, 2020, 10:43:03 AM2/29/20
to Open PHD Guiding
Hi Andrea

In general i would not worry about that test not completing. I suspect the reason is because sometimes when you calibrate your DEC axis' last move was north, so it took up the backlash. in cases where it did not complete, it may have been the backlash was not yet taken up

The backlash test results is more of an informational warning rather than a failure. it just says if you have problems with DEC guiding it may be because of backlash.

Looking at your guiding, it looks fine. your DEC and RA are very close in RMS, so your backlash isn't a major issue here.

You could improve your DEC performance by enabling auto backlash compensation. you might try starting with a value of 600ms and let it auto adjust. 
 

one last thing - while it's a good idea to run GA every night, you really don't need to run the backlash test every night. Once you have done it, chances are the existing settings you have will work fine. you do have a fair amount of backlash in DEC, but PHD is handling your overall guiding reasonably well. 

--
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/89f3d73d-9287-48af-898a-2e4ff8e7b8e3%40googlegroups.com.


--
Brian 



Brian Valente

andrea arbizzi

unread,
Feb 29, 2020, 11:51:22 AM2/29/20
to Open PHD Guiding
Thanks, Brian for your complete and exhaustive answer: now i'm more relaxed that "backlash failed test" isn't a real failure but it can happen in certain circumstances as you explained above.
Now as i'm experimenting a lot about parameters and different settings to achieve better guiding, i'm doing almost every night GA as an help: but i know i don't need it in the future.

After some tests, for me, until now, the most effective settings are:
- RA: PPEC 80% aggressivness
- DEC: Resist switch 100% aggressivness
- 0,75X guiding rate

Do you advise me lower the guiding rate to 0,5X?
Use maybe Hysteresis in RA?
Lower the DEC aggressivness maybe to 80%?

Thanks for your help

Brian Valente

unread,
Feb 29, 2020, 11:54:40 AM2/29/20
to Open PHD Guiding
Hi Andrea

your guiding settings look okay

i would not recommend reducing your guide rate - generally the faster guide rate will produce better results because guiding is more responsive

for DEC, i suggest you turn on backlash compensation
regarding PPEC for RA, that's a good choice, but you need to find your mount's primary worm period and enter that. once you have that I recommend going with the defaults for that algorithm and see how it performs



--
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/f392f53f-eac9-4a67-9e4c-72194d35d7df%40googlegroups.com.

andrea arbizzi

unread,
Feb 29, 2020, 12:03:54 PM2/29/20
to Open PHD Guiding
Thanks, Bryan for your valuable advices.
I obviously already know cem40 worm period, that is 400s. Because i did some unguided long run tests when i received the mount, just to measure PE and tracking accuracy: my purpose was to know that mount was perfect and not a bad sample to ship back to factory.

Do you advise to set Aggressivness for Dec Resist switch to 100%? Or lower to 80%? 

Brian Valente

unread,
Feb 29, 2020, 12:11:35 PM2/29/20
to open-phd...@googlegroups.com
Usually resist switch at 100% is fine. In your case it’s fine too. I do recommend you look at auto backlash compensation

--
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.

andrea arbizzi

unread,
Feb 29, 2020, 1:10:32 PM2/29/20
to Open PHD Guiding
Even if backlash seems "small" and GA often didn't recommend to compensate it, I will follow your advice and next time i will try to enable backlash compensation: 600ms, as you suggested above.
Thanks Bryan
Message has been deleted
Message has been deleted
Message has been deleted
Message has been deleted
Message has been deleted

Brian Valente

unread,
Feb 29, 2020, 7:35:47 PM2/29/20
to open-phd...@googlegroups.com
Well, Ga also said it couldn't measure the backlash because there wasn't southern movement and looking at your guiding results it shows multiple guide pulses to reverse Dec direction. 

Gives it a try and see how it goes. Worst that happens is you turn it off again

--
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.

bw_msgboard

unread,
Feb 29, 2020, 11:52:13 PM2/29/20
to open-phd...@googlegroups.com

Hi folks.  I’d like to try to clarify what this message really means and why it shows up.  To start, if you get a message like this or the backlash test (BLT) results don’t make sense, you should always look at the graph.  Even if you forget, the numerical results and the BLT graph are still available to you.  Just run PHD2, don’t start guiding, then click on the ‘Guiding Assistant’ menu.  At the bottom of the GA window, you’ll see a button that says ‘Review previous’.  Click on it and select the date when you ran the BLT that got the unexpected result.  Then look at the graph to see how your mount behaved.

 

The BLT is not dependent on how the mount was last moved or what happened in the last calibration.  The BLT goes to great lengths to clear the backlash in the mount for all reasonable cases, and I haven’t yet seen a situation where backlash of less than 10 seconds wasn’t cleared.  Typically, even after the backlash is judged to be clear, the mount will be moved at guide speed for *another 8 seconds* in the north direction.  That’s why its very unlikely that there’s any lingering backlash by the time the north moves are completed.  The test then reverses direction and issues 8 seconds of moves in the south direction.  No attempt is made to clear backlash going south because that’s the point of the test, to see how the mount behaves after a direction reversal, backlash included.  Backlash will appear as a “flat spot” at the top of graph, a time period during which the mount is receiving south guide commands but isn’t moving the Dec axis.  It might look like this:

 

 

Now look at the south moves once the backlash has finally been cleared.  The mount again starts moving in a consistent way, now in the south direction, at a rate equal to or nearly equal to the north moves.  In this case, the mount is a good candidate for PHD2 Dec backlash compensation because it is responding predictably other than the period in which it didn’t move at all.  

 

Now consider a somewhat extreme (but real) example where the test ends with the message you got: “mount never established consistent south moves”:

 

 

This is similar to the graph PHD2 produces, it’s just coming from a different source.  The red points are the north moves, the green points are the south moves.  We can see the mount took a long time to start moving south, and even when it finally got going,  the rate of south movement was quite small, nothing near the rate of movement in the north moves.  So this doesn’t look like simple backlash, it looks like backlash plus some other problem – some type of static resistance or binding, major imbalance, whatever -  that is keeping the mount from coming up to speed after the direction reversal.  This means it isn’t practical to make a recommendation for backlash compensation and it isn’t possible even to estimate what the backlash amount is.  What guess would we make in this example and how confident could we be about it?  For my part, just from looking at this graph, I have no idea how much backlash is in the mount and I’d say there’s a more important issue that needs to be understood.  

 

Now let’s look at your recent result:

 

 

In this case, there’s a better definition of the “flat spot” and the rate of south movement looks much better although it’s still not as high as the rate of north movement.  This is a more nuanced result.  In your shoes, I would judge the flat spot to be a reasonable estimate of backlash – in this case, the length of one guide pulse or about 800ms.  Of course, the measurements aren’t quite as simple as all this because they must factor out any drift that happens during the course of the test and do some linear regression to compute the best estimates of north and south rates.  It’s quite possible we may need to relax the algorithm a bit and accept south rates that differ from the north rate by a greater amount.  In this case, I think you just barely “missed the cut.”  L

 

Getting this error message doesn’t mean that Dec backlash compensation can’t help, it just means you’ll have to make your own estimate for how large the compensation should be (using the graph) and then keep an eye on things to be sure you haven’t triggered instability and oscillation.  I think that’s what Brian has done and he’s evidently gotten good results with that approach.

 

Hope this helps,

Bruce

 

 

image002.jpg
image004.jpg
image005.jpg

Brian Valente

unread,
Mar 1, 2020, 12:30:14 AM3/1/20
to Open PHD Guiding
thanks for that clarification Bruce. I had thought it came from the calibration phase, but i obviously didn't read close enough

I think your analysis still supports the recommendation of enabling backlash compensation - i took a stab at 600ms and yours says 800ms, so in the ballpark



To view this discussion on the web visit https://groups.google.com/d/msgid/open-phd-guiding/7240F41F210B4F179F1EA13B8A997BE6%40HomeDesktop.

Brian Valente

unread,
Mar 1, 2020, 1:09:51 AM3/1/20
to open-phd...@googlegroups.com
Yes except for backlash compensation maybe start at 800ms

On Sat, Feb 29, 2020 at 9:58 PM andrea arbizzi <andrea....@gmail.com> wrote:
So Bryan to sum up everything, do you confirm and suggest these settings? (see attached)
Thanks

--
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.
Message has been deleted
Message has been deleted

andrea arbizzi

unread,
Mar 1, 2020, 6:01:14 AM3/1/20
to Open PHD Guiding
Thanks, Bruce for your comprehensive, and very exhaustive explanation. 
But i don't still completely understand the reason why Phd2 is able to complete a bl tests 2 times out of 10, or maybe none out of 10.
Everything has been left mounted permanently for 4 nights: maybe just first night BL test completed and other 3 nights failed. 

I attached an image with 3 previous tests failed vs 1 test completed.
And another image shown my actual settings for the next guiding test: as Bryan confirmed, might these settings be fine?
Is 0,75X guiding rate ok? Or better to increase to 0,9X?

Thanks

To unsubscribe from this group and stop receiving emails from it, send an email to open-phd...@googlegroups.com.


 

--

Brian 

 

 

 

Brian Valente

--
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...@googlegroups.com.

backlash.jpg
Clipboard01.jpg

bw_msgboard

unread,
Mar 1, 2020, 10:55:26 AM3/1/20
to open-phd...@googlegroups.com

Hi Andrea.  As I said, you’re right on the margin of declination motion that the algorithm is looking for.  The measurement test wants to see two consecutive south moves that are at least 90% of the median size of the north moves.  At the time, that didn’t strike me as an unreasonable requirement. There will always be variations in these measurements because of seeing, amount of Dec drift, scope pointing position, mechanical loading, etc.  I haven’t gone back to your log to compute the actual sizes of the moves, but yours are probably bouncing around that 90% value – some just below, some above – such that you don’t get two consecutive above 90%.  As I said, we may relax this constraint if we see enough results like this.  When I get a chance, I’ll go back and look at each one of your tests and make sure we understand exactly what’s being calculated.  But taken at face value,  I think it’s likely that your mount doesn’t move south quite as quickly as it does north in most cases, but probably not by an amount that will affect your guiding.  Perhaps it’s a load imbalance, there’s no way to know.  As Brian said, there’s really no reason to keep doing these tests.  If you got a test that completed, that tells you what you need to know and there’s no reason to keep doing this unless you make changes in the mount mechanics.  Or you can make the visual estimate we talked about previously and let the backlash compensation algorithm auto-adjust it.  Personally, that’s what I would do at this point.  Your guiding is generally quite good, there’s probably no good reason to worry about this.  

 

Cheers,

Bruce

 

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of andrea arbizzi
Sent: Sunday, March 01, 2020 2:47 AM
To: Open PHD Guiding
Subject: Re: [open-phd-guiding] PHD2 backlash measure test failed: "mount never established consistent south moves:test failed"

 

Thanks, Bruce for your comprehensive, and very exhaustive explanation. 

But i don't still completely understand the reason why Phd2 is able to complete a bl tests 2 times out of 10, or maybe none out of 10.

Everything has been left mounted permanently for 4 nights: maybe just first night BL test completed and other 3 nights failed. 

 

I attached an image with 3 previous tests failed vs 1 test completed.

And another image shown my actual settings for the next guiding test: as Bryan confirmed, might these settings be fine?

 

Thanks

 

Ps. As soon as i click "publish", i can see my post just once: after reloaded page, it suddenly disappeares, i cannot see it no longer, and i can only see "1 message has been deleted"...

 

Il giorno domenica 1 marzo 2020 05:52:13 UTC+1, bw_msgboard ha scritto:

On Sat, Feb 29, 2020 at 12:58 AM andrea arbizzi <andrea...@gmail.com> wrote:

Hi.

My new Cem40 works like a charm for some nights and during my past nights tests i always did guiding assistant and measured backlash almost always: just once it failed but in generally guiding assistant did a complete measurements of my backlash.

 

I didn't change anything important except some parameters just to experiment a bit: since 2 nights guiding assistant isn't able to finish a backlash test, it always abort it saying "mount never established consistent south moves:test failed". I reset everything, did 3 new calibrations: didn't help. Tried to delete phd profile, i created a new profile, did a new calibration, did GA twice: same result, backlash test failed.

 

So backlash test was completed correctly for some nights, but last couple of nights it always failed.

 

This happened with my previous Cem25 too: for the first times the backlash test worked, after that it no longer has been completed..

 

Attached you can find:

- example of backlash test completed

- example of backlash failed

- debug file of backlash failed test: https://drive.google.com/file/d/1MZdhgAxHohtD2O6ZYP0MMpvdKvzbHHx9/view

- log file of backlash failed test: https://drive.google.com/file/d/1no_K7LHtTZtpvj-kF6v5Fyb9mUmqbLWA/view

 

Thanks for your help

 

Andrea

--
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...@googlegroups.com.


 

--

Brian 

 

 

 

Brian Valente

--
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...@googlegroups.com.

Il giorno domenica 1 marzo 2020 05:52:13 UTC+1, bw_msgboard ha scritto:

On Sat, Feb 29, 2020 at 12:58 AM andrea arbizzi <andrea...@gmail.com> wrote:

Hi.

My new Cem40 works like a charm for some nights and during my past nights tests i always did guiding assistant and measured backlash almost always: just once it failed but in generally guiding assistant did a complete measurements of my backlash.

 

I didn't change anything important except some parameters just to experiment a bit: since 2 nights guiding assistant isn't able to finish a backlash test, it always abort it saying "mount never established consistent south moves:test failed". I reset everything, did 3 new calibrations: didn't help. Tried to delete phd profile, i created a new profile, did a new calibration, did GA twice: same result, backlash test failed.

 

So backlash test was completed correctly for some nights, but last couple of nights it always failed.

 

This happened with my previous Cem25 too: for the first times the backlash test worked, after that it no longer has been completed..

 

Attached you can find:

- example of backlash test completed

- example of backlash failed

- debug file of backlash failed test: https://drive.google.com/file/d/1MZdhgAxHohtD2O6ZYP0MMpvdKvzbHHx9/view

- log file of backlash failed test: https://drive.google.com/file/d/1no_K7LHtTZtpvj-kF6v5Fyb9mUmqbLWA/view

 

Thanks for your help

 

Andrea

--
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...@googlegroups.com.


 

--

Brian 

 

 

 

Brian Valente

--
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...@googlegroups.com.

--

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.

andrea arbizzi

unread,
Mar 1, 2020, 1:02:10 PM3/1/20
to Open PHD Guiding
Now i completely understand the meaning of "inconsistent sout moves" and mainly the technical reason and the specific behaviour behind such a failed tests.
Thanks for your detailed and technical and exaustive explanation

Cheers 
Andrea

bw_msgboard

unread,
Mar 1, 2020, 2:12:12 PM3/1/20
to open-phd...@googlegroups.com

Hi Andrea, I’m glad it makes sense.  I still plan to go through some of your tests in detail so I can be more specific about what happened.

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/b304447d-943d-46d0-850e-47f0432797a2%40googlegroups.com.

andrea arbizzi

unread,
Mar 1, 2020, 2:29:20 PM3/1/20
to Open PHD Guiding
Bruce your previous explanation made full sense about what happened during failed tests.
I like to understand the meaning and causes and reasons about everything related with hardware and softwares, expecially strange and meaningless behaviour...and try to understand WHY...
So if you want me to do some more tests let me know: for me it's a pleasure find and try to solve problems, even more if my efforts are useful to help you and Andy to improve much better Phd2 (that is already a wonderful piece of software)...

Thanks a lot Bruce

bw_msgboard

unread,
Mar 1, 2020, 3:11:29 PM3/1/20
to open-phd...@googlegroups.com

Thanks for the kind words Andrea.  If you really do have a borderline situation that we want to handle differently, I’ll let you know and you can perhaps help out by testing the change before we release it.

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/84214d26-d130-41fc-895a-f6c887ce0edc%40googlegroups.com.

andrea arbizzi

unread,
Mar 1, 2020, 3:17:00 PM3/1/20
to Open PHD Guiding
Sure, i'm glad to help you and to contribute, for my infinitesimal contribution and my little little part, to improve software and its alghoritm.

Cheers

Andrea

bw_msgboard

unread,
Mar 1, 2020, 11:28:25 PM3/1/20
to open-phd...@googlegroups.com

Hi again Andrea.  I think your results are quite instructive and we can probably make some changes that will handle a broader class of mount behaviors while still producing a usable measure of backlash.  Here’s a table that shows the gory details of one of your tests.  I chose one that I thought looked pretty good, one that I would have expected to produce a usable backlash estimate:

 

 

What we see is that your Dec moves (the Dec Delta columns) – in both directions – seem to follow a pattern of smaller-larger-smaller-larger…  I don’t know why this happens, I think it’s somewhat unusual.  The current algorithm uses a median statistic to avoid a situation where a single outlier or the initial backlash period dominates the result.  But that approach is vulnerable to the pattern we see here – if there are 11 “smaller” moves and 10 “larger” moves, the median value will be one of the “smaller” ones. I think we can improve the measurement algorithm to deal with this sort of behavior because we know from your data that the mount actually guides pretty well.  I want to think about this some more and probably discuss it with Andy, but I think there’s room for improvement.  Once we have something to try out, we’ll probably be in touch. J

 

Thanks again for bringing this up,

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/84214d26-d130-41fc-895a-f6c887ce0edc%40googlegroups.com.

image002.jpg

andrea arbizzi

unread,
Mar 2, 2020, 2:52:14 AM3/2/20
to Open PHD Guiding
Hi Bruce.
With this data extracted from my tests, now it's so clear what happened and why BLT failed.
I'm glad to help you more to improve algorithm: i will do any test you want, as soon as weather will cooperate...

I take this opportunity to ask you this question...Do you normally prefer to  use high guiding rate value (as you told me above) or low ones? Better to use 0,5X with smoother and just necessary corrections? Or 0,9X with more reactive, instantly but maybe over corrected corrections?

<font size=3 face="

bw_msgboard

unread,
Mar 2, 2020, 10:54:09 AM3/2/20
to open-phd...@googlegroups.com

Hi Andrea.  Unless the mount manufacturer says otherwise, we advise guiding with high guide speed settings.  This makes the mount respond more quickly, obviously, but it also helps it push through small static resistance that can be present – small manufacturing errors, stiff grease, small imbalance, dragging cables, etc.  We’ve never seen evidence that this reduces the accuracy of the guiding.  Some older mounts or inexpensive mounts may not guide as well at the higher speeds, mostly for software reasons, but that shouldn’t apply to your situation.  You could even try running at a higher guide speed if you wanted to see if this “bumpiness” improves.  Just to remind you, whenever you change the guide speed, you’ll need to recalibrate.  If you then find that the calibration uses too few steps or takes too long, you can adjust the calibration step-size and repeat the calibration.  The calibration step-size is accessed under the ‘Advanced’ button on the ‘Guiding’ tab of the Advanced Dialog.  If you decide to experiment this way, please let us know what you see.  But don’t do it just for us.  J

--

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.

andrea arbizzi

unread,
Mar 2, 2020, 11:44:16 AM3/2/20
to open-phd...@googlegroups.com
Thanks Bruce.
I like to experiment a lot and to try to find the best approach to partially solve and understand problems: so i'm happy to experimenti a new settings, eithe for me as for you...No problem: trust me, for me it's a pleasure...
I usually calibrate every night now, because i'm doing many tests and experiments, and many guys advice to always calibrate every night: it only take 2 minutes...
So may i try 0,99X for DEC and 0,9X for RA?


You received this message because you are subscribed to a topic in the Google Groups "Open PHD Guiding" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/open-phd-guiding/K75yP6ufYDQ/unsubscribe.
To unsubscribe from this group and all its topics, 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/7DF37A34FD4C43298D31F2A0BA2C1358%40HomeDesktop.

Bruce Waddington

unread,
Mar 2, 2020, 9:59:56 PM3/2/20
to Open PHD Guiding
Hi Andrea.  I'd probably set the guide speed on both axes to 0.9x sidereal.  PHD2 will handle any combination of guide speed settings, but having them be equal can make it easier for us mere mortals to see what's going on. :-)

Bruce

On Monday, March 2, 2020 at 8:44:16 AM UTC-8, andrea arbizzi wrote:
Thanks Bruce.
I like to experiment a lot and to try to find the best approach to partially solve and understand problems: so i'm happy to experimenti a new settings, eithe for me as for you...No problem: trust me, for me it's a pleasure...
I usually calibrate every night now, because i'm doing many tests and experiments, and many guys advice to always calibrate every night: it only take 2 minutes...
So may i try 0,99X for DEC and 0,9X for RA?


Il giorno lun 2 mar 2020 alle ore 16:54 bw_msgboard <bw_m...@earthlink.net> ha scritto:

Hi Andrea.  Unless the mount manufacturer says otherwise, we advise guiding with high guide speed settings.  This makes the mount respond more quickly, obviously, but it also helps it push through small static resistance that can be present – small manufacturing errors, stiff grease, small imbalance, dragging cables, etc.  We’ve never seen evidence that this reduces the accuracy of the guiding.  Some older mounts or inexpensive mounts may not guide as well at the higher speeds, mostly for software reasons, but that shouldn’t apply to your situation.  You could even try running at a higher guide speed if you wanted to see if this “bumpiness” improves.  Just to remind you, whenever you change the guide speed, you’ll need to recalibrate.  If you then find that the calibration uses too few steps or takes too long, you can adjust the calibration step-size and repeat the calibration.  The calibration step-size is accessed under the ‘Advanced’ button on the ‘Guiding’ tab of the Advanced Dialog.  If you decide to experiment this way, please let us know what you see.  But don’t do it just for us.  J

 

Bruce

 

To unsubscribe from this group and stop receiving emails from it, send an email to open-phd-guiding+unsubscribe@googlegroups.com.

--
You received this message because you are subscribed to a topic in the Google Groups "Open PHD Guiding" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/open-phd-guiding/K75yP6ufYDQ/unsubscribe.
To unsubscribe from this group and all its topics, send an email to open-phd-guiding+unsubscribe@googlegroups.com.

andrea arbizzi

unread,
Mar 3, 2020, 2:24:26 AM3/3/20
to Open PHD Guiding
Hi Bruce.
I was wondering (while i was awake in the middle of the night, thinking...:)) about my mount "pattern of smaller-larger-smaller-larger" and its sort of "bumpiness" movements.
Could it be the extremely precise balancing of all the stuff the reason of above behaviour? I usually spend hours and hours achieving precise balancing of all: nothing moves in every sort of position combinations, also in some unusual positions everything stay still (IOptron mounts are extremely susceptible to very very small imbalance, eg. something positioned 2 mm out of usual place can be found out easily). On the contrary, might a very little little very small DEC imbalance help to improve or remove  its"bumpiness"?

Thanks

<font size=3 face="

andrea arbizzi

unread,
Mar 11, 2020, 7:48:17 PM3/11/20
to Open PHD Guiding
Hi Bruce.
After some days of "no sense" weather with just few rain but always cloudy and high clouds rolling by, i was able to do another test this night.
Same issue about backlash test: phd2 failed to complete it. Same incosistent south moves..
I attach a screenshot.

I also did a 2 guiding sessions with new parameters: guiding rate set to 0,9X, ppec for RA (correct worm period set), resist switch for dec with aggresivness 100%. No evidence of improvements compared to my first tests (used same settings except guiding rate 0,75X). On the contrary, these 2 sessions were a little worse than those ones..
The second guiding session in the attached log file was done with backlash enabled 800ms: compared to the previous session (with backlash disabled) the final result was a little worse... 
What do you think about?

And what do you think about this idea i asked you some days ago? about my mount "pattern of smaller-larger-smaller-larger" and its sort of "bumpiness" movements.
Could it be the extremely precise balancing of all the stuff the reason of above behaviour? I usually spend hours and hours achieving precise balancing of all: nothing moves in every sort of position combinations, also in some unusual positions everything stay still (IOptron mounts are extremely susceptible to very very small imbalance, eg. something positioned 2 mm out of usual place can be found out easily). On the contrary, might a very little little very small DEC imbalance help to improve or remove  its"bumpiness"?

Thanks Bruce

Cheers
Annotazione 2020-03-12 004329.jpg
PHD2_GuideLog_2020-03-11_190848.txt

bw_msgboard

unread,
Mar 12, 2020, 2:36:06 PM3/12/20
to open-phd...@googlegroups.com

Hi Andrea, comments below…

 

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of andrea arbizzi
Sent: Wednesday, March 11, 2020 4:48 PM
To: Open PHD Guiding
Subject: Re: [open-phd-guiding] PHD2 backlash measure test failed: "mount never established consistent south moves:test failed"

 

Hi Bruce.

After some days of "no sense" weather with just few rain but always cloudy and high clouds rolling by, i was able to do another test this night.

Same issue about backlash test: phd2 failed to complete it. Same incosistent south moves..

I attach a screenshot.

 

I also did a 2 guiding sessions with new parameters: guiding rate set to 0,9X, ppec for RA (correct worm period set), resist switch for dec with aggresivness 100%. No evidence of improvements compared to my first tests (used same settings except guiding rate 0,75X). On the contrary, these 2 sessions were a little worse than those ones..

The second guiding session in the attached log file was done with backlash enabled 800ms: compared to the previous session (with backlash disabled) the final result was a little worse... 

What do you think about?

 

I’m not surprised but it was worth a try.  This also suggests that getting the GA to produce a backlash estimate is not likely to accomplish much.  

 

And what do you think about this idea i asked you some days ago? about my mount "pattern of smaller-larger-smaller-larger" and its sort of "bumpiness" movements.

Could it be the extremely precise balancing of all the stuff the reason of above behaviour? I usually spend hours and hours achieving precise balancing of all: nothing moves in every sort of position combinations, also in some unusual positions everything stay still (IOptron mounts are extremely susceptible to very very small imbalance, eg. something positioned 2 mm out of usual place can be found out easily). On the contrary, might a very little little very small DEC imbalance help to improve or remove  its"bumpiness"?

 

I can’t really picture how balance would do this because we’re not changing the direction of the guide pulses during the sequence.  We can see this behavior on your mount on both axes even during calibration:

 

 

With other mounts, we’ve heard a number of explanations for this sort of pattern ranging from too-tight gear mesh to firmware problems.  Have you asked iOptron about this?

 

Setting all that aside, your guiding results seem pretty reasonable to me – 0.86 arc-sec to 1.0 arc-sec total RMS over a 4 hour period.  In all cases, you have pretty good agreement between RA and Dec performance and should certainly be getting nice round stars in your main images.  So you may just be operating at the limit of the seeing and what this mount can do.  

 

I’m still willing to change the backlash test a bit to make it more likely you’ll get a measurement result, but I doubt it will help you much.  If things settle down a bit on the support forum, I will try to do that in the next few days.

 

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.

image002.jpg

andrea arbizzi

unread,
Mar 12, 2020, 4:17:11 PM3/12/20
to Open PHD Guiding
Thanks, Bruce.
I contacted IOptron, explaining them this behaviour. I'm waiting for your answer (i hope they will reply to me...)
"If things settle down a bit on the support forum, I will try to do that in the next few days.": thanks very much.

Just for me to understand better: in the log guiding file, you can see some Dec peaks corrections.
Could those dec corrections peaks be caused by this "bumpiness" Dec movements?

Thanks

Cheers

Andrea
Immagine.png

bw_msgboard

unread,
Mar 12, 2020, 4:43:04 PM3/12/20
to open-phd...@googlegroups.com

Hi Andrea.  It’s always a bit hard to tell, but it looks to me like you have some stiction (static friction) on the Dec axis.  Many of these points you highlighted are associated with direction reversals in Dec.  We often see a pattern where a number of guide pulses are issued and the axis doesn’t seem to move – that would be the backlash part – and then there is often (but not always) an over-shoot.  That can be caused by stiction.  The conceptual model is that the motor is being commanded to turn, which it does, but because the axis is slightly resistant to rotating, that energy is being directed elsewhere.  The “elsewhere” is often temporary flexure in other parts of the hardware – mounting blocks, bearings, whatever.  This process goes on for a time and then the axis breaks through the resistance and actually rotates.  But when that happens, most of that stored energy is also released and the axis over-shoots.  We have to keep in mind we’re talking about tiny movements and very small amounts of energy but it doesn’t take much to disrupt guiding. 

 

There’s a somewhat more elaborate explanation starting on page 12 of this document:

 

https://openphdguiding.org/tutorial-analyzing-phd2-guiding-results/

 

This is the major reason I don’t really expect a revised backlash test to do you much good.

--

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.

andrea arbizzi

unread,
Mar 12, 2020, 4:58:30 PM3/12/20
to open-phd...@googlegroups.com
Excellent explanation: chapeau...
I suspect it was a "stiction" behaviour because i had already read that pdf almost 4 times...
At this point, after your explanation, anyIOptron reply will be unnecessary: but i'm curious what they will say about...

Thanks Bruce 


You received this message because you are subscribed to a topic in the Google Groups "Open PHD Guiding" group.
To unsubscribe from this topic, visit https://groups.google.com/d/topic/open-phd-guiding/K75yP6ufYDQ/unsubscribe.
To unsubscribe from this group and all its topics, 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/185920934759437B9FBEBC39FABD80C7%40HomeDesktop.

bw_msgboard

unread,
Mar 12, 2020, 5:54:02 PM3/12/20
to open-phd...@googlegroups.com

 

 

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of andrea arbizzi
Sent: Thursday, March 12, 2020 1:58 PM
To: open-phd...@googlegroups.com
Subject: Re: [open-phd-guiding] PHD2 backlash measure test failed: "mount never established consistent south moves:test failed"

 

Excellent explanation: chapeau...

 

Merci! J

 

I suspect it was a "stiction" behaviour because i had already read that pdf almost 4 times...

At this point, after your explanation, anyIOptron reply will be unnecessary: but i'm curious what they will say about...

 

Actually, I think an explanation from them could be most useful if it comes from one of their engineers.  Everything I’ve said is very high-level and generic, just describing an observed behavior and speculating on what might cause it.  An engineer who knows that mount specifically would have much more insight and might identify a root cause quite different from stiction.  Here’s a historical example of a problem on another kind of mount, nothing to do with yours, but still interesting:

 

https://www.youtube.com/watch?v=bVqW8ZGi7_4

 

This sort of cogging behavior can look like stiction in a guiding graph because you see an unpredictable, ragged response to the guide commands.  But in this case, it had nothing to do with the mount mechanics or the Dec axis, it was a firmware problem in the motors.  So, I think you should look for some help from iOptron. J

 

Good luck,

Bruce

 

Thanks Bruce 

Reply all
Reply to author
Forward
0 new messages