Star Lost During Calibration

[Blaine Hebert]

I am having trouble completing a calibration.  On many attempts I am getting the same Star Lost error.  My mount moves west, east and north, but it refuses to follow the star south and gives me a Star Lost error.  The star does move south, but the tracking box doesn't follow.

This is after a mount cleaning and backlash adjustment, but the mount seems to slew and track adequately.  I also had trouble getting Star Lost errors on large stars (Deneb), but if I use Auto Select Star I get a small one with this tracking issue.

I also have had several crashes; I am using 2.6.2.

[Blaine Hebert]

I am using a Meade LPI-G monochrome camera for guiding and an Orion Sirius mount connected directly to a Dell Windows 7 computer.  I can get PHD (the original version) to connect, calibrate and guide using a Windows WDM style web camera setting.  I do see to have Dec backlash, but until I can calibrate I can't measure it.

[Andy Galasso]

Blaine,

We'll need to see your guide log to be able to help with the calibration failure (see: Getting Help).

Also, could you please post your Debug Log covering the time period when you experienced the crash. Can you be more specific about what you mean by crash? Did phd2 stop responding? Close unexpectedly? Did windows display an error message? If so what did the error message say? Could you post a screen-shot of the error message?

Andy

[Blaine Hebert]

I use a separate computer for my camera and telescope control, I will find the log file.

I have had several crashes where I get a Windows error message "This program has stopped working..."  I will try to get you more information.

Last night I was finally able to actually get calibration; I reduced my DEC ms from over 2500 to 500 (I repeatedly changed this but it seems to get reset).  Apparently there is some residual backlash in my DEC, perhaps when calibration clears backlash from dec the star moves too fast to follow; I can clearly see the star in my image screen but the tracking box doesn't follow it.

With the one calibration I got I was able to get good guiding (before clouds and my work schedule shut me down).

BH

On Friday, November 11, 2016 at 10:04:02 PM UTC-8, Blaine Hebert wrote:

[bw_msgboard]

Hi Blaine.  As Andy said, we need to see your log files to figure out what’s going wrong.  But I wonder, do you have backlash compensation set in the mount?  If so, that isn’t going to work – those compensation values in the mount firmware should be set to zero.

 

Good luck,

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.
For more options, visit https://groups.google.com/d/optout.

[Blaine Hebert]

I find 9 debug logs and a similar number of guide logs.  How do I determine which are important?

Should I use a utility to examine them for problems or errors?

BH

[Blaine Hebert]

Attached is one guide debug log and one guide log.

BH

On Saturday, November 12, 2016 at 12:04:02 AM UTC-6, Blaine Hebert wrote:

[Andy Galasso]

Hi Blaine,

The Guide Log you attached is empty. The way to tell which log file to choose is to look at the date and time embedded in the filename.  That timestamp represents the time that PHD2 was launched.  You can use PHD2 Log Viewer to view the guide logs too.

The debug log shows the camera looping exposures for a few minutes then hanging (failing to download a 1-second the exposure after 16 seconds) and PHD2 disconnecting the camera, trying to re-connect, and failing again to download an image.  This type of error nearly always indicates a USB connection problem to the camera and the best advice we can provide is to look at each USB cable, USB hub, USB extension cable, and USB connection between the camera and the PC and try changing things one piece at a time until you find the culprit.

So far we seem to have 3 distinct issues we need to track down.

1. The calibration problem.  We'll need to see a non-empty guide log that contains an example of a failed calibration.

2. The PHD2 crash.  We'll need to see a debug log that corresponds to a time when phd2 crashed. In our experience crashes are almost always caused by the the drivers (camera, mount or other hardware) so you should make sure you are using the most up-to-date camera driver.

3. The camera timeout and disconnect. Check you USB cables etc. as mentioned above.

Andy

[Blaine Hebert]

Forgive me, but I have not worked with guide logs much.

Try this attached file.

BH


On Saturday, November 12, 2016 at 12:04:02 AM UTC-6, Blaine Hebert wrote:

[Blaine Hebert]

Here is another guide log.

BH

On Saturday, November 12, 2016 at 12:04:02 AM UTC-6, Blaine Hebert wrote:

[Andy Galasso]

BH,

Thanks for the guide logs.  In those logs we see several consistent and successful calibrations so I think your mount is behaving normally.

One thing that seems questionable is that your camera's ASCOM driver is reporting a pixel size (7.5 um) that does not seem right.  What type of guide camera are you using?

Andy

[Blaine Hebert]

Here is another guide log:

BH

On Saturday, November 12, 2016 at 12:04:02 AM UTC-6, Blaine Hebert wrote:

[Blaine Hebert]

Another guide Log.

BH

On Saturday, November 12, 2016 at 12:04:02 AM UTC-6, Blaine Hebert wrote:

[Andy Galasso]

BH,

The first log (PHD2_GuideLog_2016-11-11_235135.txt) was what you already sent earlier today. That was the one where we saw several successful calibrations, but led me to ask you what guide camera you are using because the pixel size being reported seems questionable.  What guide camera are you using?

The second guide log (PHD2_GuideLog_2016-11-11_232723.txt) shows more of the same ... numerous successful, normal looking calibrations. At this point I'm a bit confused about what problem we are trying to solve, since the original question was about failed calibrations, but the log shows 9 successful calibrations, with one failed calibration (#7 of 10) that failed because guiding was stopped during the RA calibration phase. Guiding could have stopped because you clicked Stop, or for some other reason (we would need to see the Debug log to say definitively why guiding stopped.)

Andy

[Shane Lloyd]

My guess would be toupcam as I recently updated the ascom driver for mine and noticed it reports that pixel size instead of 3.75 um, I have informed the site owner and are waiting to hear about a fix. 

[Blaine Hebert]

OK, now this has become a bigger mystery.

On every calibration except the last one I got an error message stating "star lost".  The tracking box failed to follow the star on the last, south correction and gave me an error message and I was unable to proceed.  Only after lowering my calibration step size did I apparently complete the calibration.

BH


On Friday, November 11, 2016 at 10:04:02 PM UTC-8, Blaine Hebert wrote:

[Andy Galasso]

BH,

We suspect the camera ASCOM driver is reporting a pixel scale that is 2x too large. It is reporting 7.5 um, but the pixels are actually 3.75 um.   Since PHD2 uses the reported pixel size to compute the recommended calibration step size, the step size is coming out too large.

This seems to be confirmed by your report that reducing your calibration step size helps your problem.  We could confirm the theory for certain if you could send us your debug log and guide log for the session where you experienced the calibration failure.

Assuming that is the problem, the immediate workaround is to reduce the calibration step size (as you have already done.)  In addition, you should report the problem to the ASCOM driver provider so they can fix the ASCOM driver to report the correct camera pixel size.

Andy

[Shane Lloyd]

Hi guys, I have recieved word from my supplier re the driver his reply is below. Basically at 640x480 it reads double pixel size and at full resolution of 1280 x 960 the driver will read correct pixel size. Hope this is of some use.

"Hi Shane

Apparently the pixel reading you got from PHD2 has to do with the capture window scale you are using. So it's 7.5 micron under small resolution used like 640x480. The development engineer said this reads correct since it reflects double (3.75 x 2) at this resolution. So if you set full resolution of 1280x960, it will be 3.75 mircron"

regards Shane

[Andy Galasso]

Shane,

Thanks for the information.

Andy

[Bruce Waddington]

Hi Blaine.  We don' t seem to be converging on an understanding of your problem.  I think we could make headway if you could do three things for us:

1.  Locate and post the debug log file with the name "PHD2_DebugLog_2016-11-11_235135.txt".  We already have the guide log, we need the matching debug log.
2. Tell us if you have any sort of backlash compensation or backlash correction value programmed into the mount.
3. Tell us if you have any settings programmed in the EQASCOM interface that attempt to alter the guide commands being sent to the mount in any way - things like "RA/DEC pulse width Override" or "RA/Dec width Gain."

If the answer to either (2) or (3) is a yes, that's a problem that has to be fixed.

Thanks,
Bruce

[Blaine Hebert]

Thanks for your patience!

Attached is the Debug Log.

I do not believe that I have ever changed anything in EQASCOM, I work with CdC and PHD2 only.

In a previous post I was having trouble with strange "saw-tooth" guiding and someone suggested I turn off backlash compensation and that helped with that problem.  Since then I have attempted to adjust my DEC backlash.  this was my first session after cleaning and adjusting my mount.

BH

[bw_msgboard]

Hi Blaine, thanks for supplying the debug log file.  I think we should focus on the calibration and “lost star” problem for now and try to get that figured out.  I think that problem is telling us something important about the mount, and we need to understand it.  So I’m going to suggest a sequence of things to do that will help us narrow down the problem.  Your debug log shows some historical calibrations that suggest to me you’ve got some kind of backlash compensation occurring either in the mount firmware or in EQMOD.   And my theory for now is that this is what’s causing the lost-star event.  I think the reversal in direction from north to south at the end of calibration is resulting in a very large move, much larger than anything that should happen and large enough to drive the guide star out of the search region.  Unfortunately, since the star is lost, we don’t get a measurement of how big this move was.  So here’s what I’d like you to do:

 

1. Once again, try to disable backlash compensation any place you know about it (see below).
2. Go into the PHD2 brain dialog, on the guiding tab, and disable the checkbox that says “Fast recenter after calibration or dither”
3. On the same tab, increase the search region size to 25 pixels

 

 

These two changes will make it less likely that PHD2 will lose the star during calibration.

 

Next, try to do a calibration as follows:

4. Slew to a field with good candidate guide stars – some place near Dec=0 and at least 45 degrees above the horizon
5. Once you’re in the area, first move the mount NORTH at guide speed using the hand controller or the PHD2 manual guide tool.  Move it NORTH for at least a total of 20 seconds.
6. Choose a guide star that is by itself in the larger search region, then start the calibration without any additional movement of the mount.

 

I think you might get a successful calibration, one that doesn’t lose the star.  If so, stay on the same guide star and run the Guiding Assistant.  Most importantly, let the Guiding Assistant measure the declination backlash.  

 

One this is all done, post both the guide and matching debug log files so we can analyze them.  

 

Getting back to the question of backlash compensation, it sounds like you’ve been struggling with backlash a lot and have probably been all over the place trying settings you thought might help.  I don’t know anything about your mount, but I know that EQMOD is chock-full of settings that can ruin calibration and guiding.   Here are some of the ones I know about from the documentation:

 

 

 

Set the RA and Dec rates in the upper-right corner to the same values, somewhere in the range of 0.7 to 1.  Use the same setting for both axes.  Be absolutely sure the DEC Backlash (msecs) slider is set to zero.  Then be sure the RA Width Gain and DEC Width Gain at the left are set to 100%.  When this is done, you need to be absolutely sure the changes “stick” and don’t somehow revert back to old values.

 

I think if you can work through all these steps, we’ll be able to see what’s going on and get you back in business.

 

Good luck,

Bruce

 

 

---

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Blaine Hebert
Sent: Wednesday, November 16, 2016 9:33 PM
To: Open PHD Guiding
Subject: [open-phd-guiding] Re: Star Lost During Calibration

 

Thanks for your patience!

[bw_msgboard]

Sorry, I forgot to mention a few other things that may be self-evident, but it’s better to reiterate them anyway:

 

1. When you re-do the calibration, please use 2-sec guide exposures.
2. Make sure the Dec guide mode is set back to ‘auto’ – otherwise Dec calibration won’t happen at all
3. Please don’t fiddle around with *any* settings while the tests are being done.

 

Good luck,

Bruce

 

---

[Blaine Hebert]

Bruce,

I attempted to follow most, if not all of your recommendations and got a successful night of guiding.
I found these settings in ASCOM:
Pulse guide settings; RA 0.90, DEC 0.80
Min pulse width ms 50
Dec backlash msec 2000

I am attaching my guide logs and debug files.

I did have trouble with DEC moving in the wrong direction after a meridian crossing (but that is another issue).

BH


My search window in PHD2 was about 10, I set it to 25.
I did find backlash compensation in several places.

[bw_msgboard]

Hi Blaine.  It sounds like we’ve resolved the mystery of the lost star during calibration.  Without knowing which recommendations you ignored (and why), I’m not how else to help you.  The backlash measurement in the Guiding Assistant shows a modest amount of residual backlash, so you must have done a good job with the mechanical adjustments.  With a residual backlash of 2 seconds at these guide speed settings, you should be able to guide in both directions for Dec.  You can certainly try the recommended Dec backlash compensation in PHD2, BUT LEAVE THE BACKLASH SETTING IN EQMOD AT ZERO.  I can’t emphasize this too much, I think this was the source of most of your calibration problems and all this back-and-forth trouble-shooting.

 

The Guiding Assistant shows you have quite a bad polar alignment, about 48 arc-min.  That probably won’t cause any guiding problems, but you will probably see field rotation when you work closer to the pole.  Try using the drift alignment tool to get this improved – it’s pretty quick and easy once you get the hang of it.  Also, I wonder if you’re clinging to the idea that you need to re-calibrate every time you slew to a different location.  That’s just a waste of time and a source of calibration errors – get one good calibration near Dec=0 and just keep using it.  

 

Good luck,

Bruce

 

---

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of Blaine Hebert
Sent: Friday, November 18, 2016 12:48 PM
To: Open PHD Guiding
Cc: blaine...@charter.net; bw_m...@earthlink.net
Subject: Re: FW: [open-phd-guiding] Re: Star Lost During Calibration

 

Bruce,

I attempted to follow most, if not all of your recommendations and got a successful night of guiding.
I found these settings in ASCOM:
Pulse guide settings; RA 0.90, DEC 0.80
Min pulse width ms 50
Dec backlash msec 2000

I am attaching my guide logs and debug files.

I did have trouble with DEC moving in the wrong direction after a meridian crossing (but that is another issue).

BH


My search window in PHD2 was about 10, I set it to 25.
I did find backlash compensation in several places.


On Thursday, November 17, 2016 at 8:20:33 PM UTC-8, Bruce Waddington wrote:

Sorry, I forgot to mention a few other things that may be self-evident, but it’s better to reiterate them anyway:

 

1.      When you re-do the calibration, please use 2-sec guide exposures.

2.      Make sure the Dec guide mode is set back to ‘auto’ – otherwise Dec calibration won’t happen at all

3.      Please don’t fiddle around with *any* settings while the tests are being done.

 

Good luck,

Bruce

 

---

From: open-phd...@googlegroups.com [mailto:open-phd...@googlegroups.com] On Behalf Of bw_msgboard
Sent: Thursday, November 17, 2016 7:42 PM
To: 'Blaine Hebert'
Cc: 'Open PHD Guiding'
Subject: RE: [open-phd-guiding] Re: Star Lost During Calibration

 

Hi Blaine, thanks for supplying the debug log file.  I think we should focus on the calibration and “lost star” problem for now and try to get that figured out.  I think that problem is telling us something important about the mount, and we need to understand it.  So I’m going to suggest a sequence of things to do that will help us narrow down the problem.  Your debug log shows some historical calibrations that suggest to me you’ve got some kind of backlash compensation occurring either in the mount firmware or in EQMOD.   And my theory for now is that this is what’s causing the lost-star event.  I think the reversal in direction from north to south at the end of calibration is resulting in a very large move, much larger than anything that should happen and large enough to drive the guide star out of the search region.  Unfortunately, since the star is lost, we don’t get a measurement of how big this move was.  So here’s what I’d like you to do:

 

1.      Once again, try to disable backlash compensation any place you know about it (see below).

2.      Go into the PHD2 brain dialog, on the guiding tab, and disable the checkbox that says “Fast recenter after calibration or dither”

3.      On the same tab, increase the search region size to 25 pixels

 

 

These two changes will make it less likely that PHD2 will lose the star during calibration.

 

Next, try to do a calibration as follows:

4.      Slew to a field with good candidate guide stars – some place near Dec=0 and at least 45 degrees above the horizon

5.      Once you’re in the area, first move the mount NORTH at guide speed using the hand controller or the PHD2 manual guide tool.  Move it NORTH for at least a total of 20 seconds.

6.      Choose a guide star that is by itself in the larger search region, then start the calibration without any additional movement of the mount.