Guiding the Sky-Watcher AZ-GTi

576 views
Skip to first unread message

Rafael Barberá Córdoba

unread,
Jan 3, 2020, 6:20:53 PM1/3/20
to Open PHD Guiding
I know that my setup is a little unconventional, but... it's the one that I've at this point in time. I'm using a Sky-Watcher AZ-GTi mount in equatorial mode. I'm using the EQMod INDI driver from Ekos on Mac. I know that the mount is a little "overweighted" because I'm mounting on it a TeleVue-85 with a Canon 40D and a generic 50mm / 190mm guidescope with an ASI120MM-S. Anyway I try to have the mount well balanced in both axis. With this setup, the resolution for my main telescope is around 2"/px and the guider is 4"/px.

Over the last month I've been experimenting with guiding with irregular results (using the internal module inside Ekos). As I need more insights in what was happening I've started to use PHD2 yesterday. It's a big improvement on the kind of information that I can retrieve about the performance / calibration and problems (A big thanks! to all the community around PHD2).

Well, after my first calibration it was clear that something wrong happens with the DEC axis (it seems a lot of backslash.. or worse). Anyway, I get a guiding almost valid for 180sec subs, but clearly something is wrong. I've uploaded the logs for this session (https://openphdguiding.org/logs/dl/PHD2_logs_3XD5.zip). As you can see the guiding behavior is very odd. It has periods in which all seems to woks as expected with the mount responding to all the commands and the a drift in DEC begins and any correction works.

If you look at the calibration graph you can see that it works more or less fine, except for the GuideSouth pulses. They are all clustered just outside the horizontal branch of the graph.

The, the problem appear while guiding. For example, the 10 minutes run that begins at 19:26:12 (2020-01-02), until frame 117 the mount more or less respond to the GuideSouth commands, but then, the drift in DEC begins to increase and the pulses are unable to recover the guiding.

In the next run (the 12 min one), you can see a continuous drift that can't be controlled by the guiding commands. The great "pulse" between frames 69 and 78 was me, touching the cables trying to easy the load. Yes I know, I can't use this rig under wind... or breeze :)

Then without any apparent cause, PHD2 begins to lose the reference star, low SNR. I goto other areas of the sky, refocus the finder, and nothing. I must restar PHD2 to make it works again. (But this is another problem, I'm not interested right now.)

Then I tried to increase the guiding rate from 0.5x to 1.0x. I changed it on the INIDI EQMod driver and the performed a new calibration. This time all the south commands where clustered on the center.

I tried also a Stat cross-test but the only moment visible was in AR.

So, I'm very perplexed, because the mount moves fine on slew movements (the goto, assisted by Astrometry, was dead center around the full sky), but it fails on guiding... sometimes.

I know that's a cheap mount, overweighted but I want to know if I can do anything to fix some of my problems, because when it works... I love this light and portable rig. I've attached the only valid result: an stack of 6x180sec showing the moment of comet C/2017T2 in about half hour. You can see the the stars are very round and small, just that I've expected from the TV-85.

Regards and clear skies!

2020.01.03_C2017T12_detalle.jpg


 


Rafael Barberá Córdoba

unread,
Jan 3, 2020, 6:24:21 PM1/3/20
to Open PHD Guiding
One thing that I've forget: the Guiding Analysis tool shows a big PA misalignment, but at the beginning of the session I've performed PA assisted by astrometry and they put my mount under the 1 arcsin from pole, so I'm very confused about how to interpret all this data

Captura de pantalla 2020-01-02 a las 18.46.21.png


Michael O'Brien

unread,
Jan 3, 2020, 6:31:47 PM1/3/20
to Open PHD Guiding
Your PA error looks to be 41 arc seconds not under 1 arc second.

Brian Valente

unread,
Jan 3, 2020, 6:44:17 PM1/3/20
to Open PHD Guiding
Hi Rafael

sorry to hear you're having troubles. Slewing and gotos are quite different beasts than accurate guiding, so would not think good slews = good guiding. Guiding requires a lot more precision, often at sub-pixel accuracy, whereas goto is generally considered good within 50 pixels, so we're talking many orders of magnitude difference in the demands on accuracy of the mount.

I think we need more data on this, but looking at your logs my initial reaction is your DEC axis is not responsive at all to guiding corrections. here's a typical example where you can see dozens of attempts to change direction and DEC never does (even up to the max 2000ms) so here this is either suggesting you have backlash in the many tens of seconds (unlikely but possible) or the DEC axis is simply not responding at all

image.png

this has happened in nearly all (or all) of your guiding runs. you might check your cables to ensure it's responding, and use PHD's manual guide tool to confirm your mount is responding in all four directions. 

the periods in which "it just seems to work" is really when DEC is within your limits and doesn't require any corrections. Once it gets above the limits, you can see the dec axis is unresponsive. 

in addition, there appear to be some mechanical issues in DEC, such as this big jump here that is nearly 15":
image.png
you can see the guiding corrections again not having any impact. 

(btw there are some similar but smaller jumps in RA, but maybe that's the next thing to focus on)

while it's possible (and probable) things like overweighting your mount, poor tripod etc. can contribute to these issues, there seems to be something fundamentally problematic with your DEC here. 

I suggest some additional information that would help confirm or isolate the issue:


I'd start from scratch - 
update to the latest 2.6.6 dev 3 version of PHD

create a new profile and make sure you enter all your information in correctly  (it seems like you've done a pretty good job on your settings, but again just to make sure)

make sure your EQMOd settings are set according to these recommendations: https://github.com/OpenPHDGuiding/phd2/wiki/EQASCOM-Settings 

then slew to the intersection of meridian and celestial equator, bump your mount north, and then do a calibration
after a successful calibration (i.e., no errors reported) 

immediately after that, run a guiding assistant run of 2-3 minutes, press stop, and accept the recommendations

then allow your mount to do 15-30 minutes of uninterrupted guiding to see how the mount performs based on the settings above (the runs of less than a few minutes aren't helpful to see what's going on)

this may not uncover anything new, but it's a good way for us to see more clearly what's going on

I'd definitely check the manual guide tool to see if DEC is responding and check those axis connections

there may also be a setting somewhere with an alt az mount about enabling dec corrections? just a guess here


Brian



--
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/5916a31a-b69d-401b-8ea5-bd87a0bfc806%40googlegroups.com.


--
Brian 



Brian Valente

Rafael Barberá Córdoba

unread,
Jan 3, 2020, 6:46:21 PM1/3/20
to Open PHD Guiding
Yep, I tried to write under 1 arc minute :)

Brian Valente

unread,
Jan 3, 2020, 6:49:18 PM1/3/20
to Open PHD Guiding
btw here's another reason why I think DEC Is being totally unresponsive vs. unresponsive in just one direction:

image.png

after the big bump/jump you can see phd attempting to adjust in the opposite direction, and while it does eventually come back, it continues on - the linear progression suggests to me the guide pulses didn't have any impact on this DEC drift

Brian

On Fri, Jan 3, 2020 at 3:20 PM Rafael Barberá Córdoba <rbar...@gmail.com> wrote:
--
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/5916a31a-b69d-401b-8ea5-bd87a0bfc806%40googlegroups.com.

Rafael Barberá Córdoba

unread,
Jan 3, 2020, 6:53:10 PM1/3/20
to Open PHD Guiding
Thank you Brian for your first analysis, as soon as I've a clear night I'll perform the test sequence you have proposed. But one thing is bothering me right now. On the first half of the first graph you have attached, I can see how the mount is responding to the guidesouth pulses, isn't it? Slow, but is responding or I'm misinterpreting the correlation between pulses and change in drift?

Regards

Brian Valente

unread,
Jan 3, 2020, 6:54:42 PM1/3/20
to Open PHD Guiding
one more thing - just poking around the internet there's apparently an alt-az mode and an eq mode for that mount. i'm guessing you'd want it in EQ mode? otherwise the dec would not be responsive, which seems to be the case in your situation. sounds like you'd need a wedge too

if you google 'guiding AZGTI' you'll see a lot of additional threads and info on this mount. 

On Fri, Jan 3, 2020 at 3:20 PM Rafael Barberá Córdoba <rbar...@gmail.com> wrote:
--
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/5916a31a-b69d-401b-8ea5-bd87a0bfc806%40googlegroups.com.

Brian Valente

unread,
Jan 3, 2020, 6:57:59 PM1/3/20
to Open PHD Guiding
I'm not sure which graph you are referring to, but if it's the one below, you can see it is not responsive in either direction - the only reason it eventually flops over is this huge and sudden correction (which is either something banging the mount or a mechanical problem). and then you can see the same slow linear progression in the same direction. so afaics none of the corrections are being responded to by the dec axis

image.png

On Fri, Jan 3, 2020 at 3:53 PM Rafael Barberá Córdoba <rbar...@gmail.com> wrote:
Thank you Brian for your first analysis, as soon as I've a clear night I'll perform the test sequence you have proposed. But one thing is bothering me right now. On the first half of the first graph you have attached, I can see how the mount is responding to the guidesouth pulses, isn't it? Slow, but is responding or I'm misinterpreting the correlation between pulses and change in drift?

Regards

--
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.

Rafael Barberá Córdoba

unread,
Jan 3, 2020, 7:10:06 PM1/3/20
to Open PHD Guiding
Hi Brian,

Of course, yes I'm in equatorial mode (EQMod drive only works in this setup) and with an equatorial wedge

IMG_7912.jpg

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

Rafael Barberá Córdoba

unread,
Jan 3, 2020, 7:18:46 PM1/3/20
to Open PHD Guiding
Yes, when I saw this graph I thought that something is completely wrong with the DEC (the big spike in both axes was me, touching the mount and moving the cables).

The graph I'm talking was this one:

dec_working.png


I've put a vertical arrow in all place where the mount seems responding to the guide commands.


Also from the calibration data I see response in DEC:


Captura de pantalla 2020-01-04 a las 1.17.00.png


Captura de pantalla 2020-01-04 a las 1.17.35.png



Or I'm misinterpreting those graphs?

Regards


El sábado, 4 de enero de 2020, 0:57:59 (UTC+1), Brian Valente escribió:
I'm not sure which graph you are referring to, but if it's the one below, you can see it is not responsive in either direction - the only reason it eventually flops over is this huge and sudden correction (which is either something banging the mount or a mechanical problem). and then you can see the same slow linear progression in the same direction. so afaics none of the corrections are being responded to by the dec axis

image.png

On Fri, Jan 3, 2020 at 3:53 PM Rafael Barberá Córdoba <rbar...@gmail.com> wrote:
Thank you Brian for your first analysis, as soon as I've a clear night I'll perform the test sequence you have proposed. But one thing is bothering me right now. On the first half of the first graph you have attached, I can see how the mount is responding to the guidesouth pulses, isn't it? Slow, but is responding or I'm misinterpreting the correlation between pulses and change in drift?

Regards

--
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 Valente

unread,
Jan 3, 2020, 9:46:49 PM1/3/20
to Open PHD Guiding
Hi Rafael

it's possible those guide pulses made some impact, but more likely it was random.

Look at your overall picture: assuming those guide pulses did move the mount, why didn't it continue to move until it was back to 0? 

compare your graph to one that is a typical guide log, where there are a few pulses in each direction, and it's mostly staying around 0:

image.png

so while it's possible, but there are literally hundreds maybe thousands of pulses before and after that on many other graphs that clearly show lack of responsiveness. the larger pic of your graph is that DEC never fully gets back to 0 and continues to drift. 

Even in the case that those did cause some movement, an axis needs to be repeatable and consistent when guide pulses are sent to it, and in your case that doesn't appear to be the case.

Brian

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/6949e3e8-6642-4a02-999d-9b87093fd232%40googlegroups.com.

steve

unread,
Jan 4, 2020, 4:10:13 AM1/4/20
to open-phd...@googlegroups.com
Hi

The easiest modification you could try is to slacken the DEC mesh: worm
to gear. It's sticking. Even better, dismantle, clean and grease the same.

Cheers, happy new 2020 and HTH.

Rafael Barberá Córdoba

unread,
Jan 7, 2020, 3:31:58 AM1/7/20
to Open PHD Guiding
Hi Brian,

I've been able to follow your advices about how to test the mount. I've been unable to upload the logs, because 2.6.6dev3 keep crashing on macOS each time I try to upload the logs. I've attached the zip file manually to this post.

As you can see the mount is in fact tracking. It has a very odd behavior that looks as if some elastic element is involved in the DEC gear train. But I'm open to other interpretation from you.

As a complementary analysis I've performed two star cross test: one failed and other semi successful. Also I've performed a manual guide test in the four directions and have recorded the screen on video, so you can watch how the mount is responding to the guiding commands. 

For the manual guiding test, I've been sending 2000ms pulses until I see some movement on each direction. This way I've discovered that I need 16sec (8x2000ms pulses) to reverse from guiding north to guiding south.

All this complementary data is available on my dropbox account: https://www.dropbox.com/sh/kdyxh9zzfaggar4/AACvClXK-srEQIF8dbk1me_Ea?dl=0
PHD2_Logs.zip

Rafael Barberá Córdoba

unread,
Jan 7, 2020, 6:25:06 AM1/7/20
to Open PHD Guiding
Sorry Brian, I've sent the wrong logs. The cleaner logs (only calibration, assistant run and two long guiding runs) are this ones, from January 4th
PHD2_GuideLog_2020-01-04_184707.txt.zip

Brian Valente

unread,
Jan 7, 2020, 10:54:58 AM1/7/20
to Open PHD Guiding
Hi Rafael

" I've discovered that I need 16sec (8x2000ms pulses) to reverse from guiding north to guiding south."

I think you confirmed my suggestion that your DEC is unresponsive for any reasonable guiding application. 

I will update my comment to this: you have SO much backlash (16 seconds apparently) that your DEC axis will not be effective for guiding. 

you need to hunt down the mechanical reasons for this before you can address your DEC guiding performance

your calibration graph demonstrates the problem. 

image.png

the right hand side of the red line should be roughly following the white line, but instead it is almost a straight line towards the right

that is your DEC not responding in any reasonable timeframe

Brian


--
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.

Rafael Barberá Córdoba

unread,
Jan 7, 2020, 11:32:28 AM1/7/20
to Open PHD Guiding
Thanks for your time Brian, I'll investigate the mechanical issue with this mount, and try to use unidirectional guiding on DEC in the meantime.

El martes, 7 de enero de 2020, 16:54:58 (UTC+1), Brian Valente escribió:
Hi Rafael

" I've discovered that I need 16sec (8x2000ms pulses) to reverse from guiding north to guiding south."

I think you confirmed my suggestion that your DEC is unresponsive for any reasonable guiding application. 

I will update my comment to this: you have SO much backlash (16 seconds apparently) that your DEC axis will not be effective for guiding. 

you need to hunt down the mechanical reasons for this before you can address your DEC guiding performance

your calibration graph demonstrates the problem. 

image.png

the right hand side of the red line should be roughly following the white line, but instead it is almost a straight line towards the right

that is your DEC not responding in any reasonable timeframe

Brian


On Tue, Jan 7, 2020 at 12:32 AM Rafael Barberá Córdoba <rbar...@gmail.com> wrote:
Hi Brian,

I've been able to follow your advices about how to test the mount. I've been unable to upload the logs, because 2.6.6dev3 keep crashing on macOS each time I try to upload the logs. I've attached the zip file manually to this post.

As you can see the mount is in fact tracking. It has a very odd behavior that looks as if some elastic element is involved in the DEC gear train. But I'm open to other interpretation from you.

As a complementary analysis I've performed two star cross test: one failed and other semi successful. Also I've performed a manual guide test in the four directions and have recorded the screen on video, so you can watch how the mount is responding to the guiding commands. 

For the manual guiding test, I've been sending 2000ms pulses until I see some movement on each direction. This way I've discovered that I need 16sec (8x2000ms pulses) to reverse from guiding north to guiding south.

All this complementary data is available on my dropbox account: https://www.dropbox.com/sh/kdyxh9zzfaggar4/AACvClXK-srEQIF8dbk1me_Ea?dl=0

--
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.

Tarun Kottary

unread,
Jan 7, 2020, 12:07:06 PM1/7/20
to open-phd...@googlegroups.com
Dec backlash is a common issue with this mount. There was a thread on cloudy nights along with youtube video on how to fix the sloppy dec gear. Better yet reach out to skywatcher first.

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/78084e2e-49fc-49b5-bfdc-46522370f4d3%40googlegroups.com.

Rafael Barberá Córdoba

unread,
Jan 7, 2020, 12:15:37 PM1/7/20
to Tarun Kottary, open-phd...@googlegroups.com
I was unable to fin updated information about the fix. All the thread are from 2018 and the image links are expired. If you have any working link I will be very glad if you share it with me.

Unfortunately SW customer service in my country (Spain) is very deficient :(.

Saludos,
Rafa Barberá

> On 7 Jan 2020, at 18:07, Tarun Kottary <tkot...@gmail.com> wrote:
>
Reply all
Reply to author
Forward
0 new messages