Category Archives: Failures

Replaced observatory computer

My old observatory PC has started becoming unstable, with sudden unexplained restarts and failures.  Not bad – it’s had a good 5-year run, living permanently outdoors in the cold.  It was also time to do something about the demise of Windows 7, and I found I couldn’t upgrade it to Windows 10 easily (nor was I sure that would fix the failures problem).

So, I’ve replaced it with a new machine.  As before, I configured the machine for no moving parts – especially no rotating hard drives – to reduce the effect of cold weather. I found, a couple of PCs ago, that hard drives seize up in the cold. (In fact the new machine has one moving part – a fan – but, being thermostatically controlled, a computer fan is rarely a problem in the cold since it doesn’t try to come on anyway.) The new one is installed in the observatory now and seems to be fine, and stable.

I forgot a number of configuration things though.  It took a couple of weeks to get all the necessary software and drivers installed and working.  Then, last night, I tried a “real” observing run and instead generated a big to-do list of things I forgot to set up.  Things such as transferring the PEC and TPoint models for the mount.

I think I’ve done all the necessary things now, but we’re back in cloudy weather for several days, so next test is some time away.

Starting Mount Forensics and Maintenance

It’s time to admit I have a mount performance problem

I am so impressed with the build quality of the MX mount that I’ve been reluctant to admit that there seems to be a performance issue showing up in my data. I think I’m ready to admit that now.

  • There is pretty good evidence of a problem;
  • I’m pretty sure it’s a rather basic problem, likely easily correctible (but only once I admit I have a problem and work on it);
  • The SB support forums show excellent responses from both manufacturer and other users, and every problem someone reports and works through seems to be solved.

Symptoms of one or more problems

There are a couple of things that I’m sure are indications of problems, although it remains to be seen if there is a single underlying problem or several. The symptoms are

Periodic Error (PE) is out of spec.


uncorrected periodic error 12 arc-seconds peak-to-peak

I gathered data incorrectly several times (mainly by forgetting to record the camera rotation angle, or to ensure it was 0 or 180), so it took me a while to see this. However, now I have reliable data showing I’m getting uncorrected Period Error of almost 12 arcseconds, peak to peak.

The spec for the mount is “maximum 7 arcseconds, peak to peak”. And this isn’t just marketing hype – most Paramounts do much better, and the SB support forums are full of examples of SB being horrified in the rare case when someone’s mount exceeds this spec, and of SB taking it seriously to correct it.


Corrected periodic error, 3.9 arc-seconds peak-to-peak

Note that with Periodic Error Correction (PEC) active, the mount tracks at 3.9 arc seconds, which is quite good. However, I think I should work on getting the uncorrected performance into spec.

    • Surely if I get uncorrected performance as good as possible, then PEC should improve corrected performance even more; and
    • Surely being out-of-spec means something is wrong, and that problem could have other effects that show up in results, or that cause wear and tear on the mount’s drive mechanism.

So, I’ve resolved to work on this, with the minimum target of getting uncorrected PE into spec.

Strange Dec oscillation when guiding

Next, I’m getting very strange behaviour in the Declination axis when auto-guiding. At this point I don’t have a good theory for whether this is the same problem as the PE above, or a separate problem.

When autoguiding, there is no reason why there should be Dec motion or Dec corrections at all unless Polar Alignment is off, or unless ProTrack is enabled and making corrections that it thinks are needed.

Simulation of strange behaviour seen in guidingWhat I’m seeing in Declination is large-scale, slow-period oscillation in the Dec axis. Dec shows as being several arc-seconds off, but multiple successive autoguiding periods don’t seem to be able to correct it. Then it suddenly swings over to being several arc-seconds off on the other side of the axis. At no point does it converge to a “no or small error” state. (This image is a simulation since I forgot to take screen captures the last time I did this test.)

Aside from the obvious – that this indicates something is wrong – this also means that I can’t use dithering when collecting a series of images. After the first image, guider “settling” back within a specified error range never happens. RA settles quickly, but Dec just keeps oscillating from far-off on one side to far-off on the other side.

Things that look odd but I’m not sure are symptoms of problems

The above two issues are things that I’m sure are problems. There are also a couple of other strange things going on that I’m not sure are problems, but they look odd.

TPoint-Scatter-Graph-2First, my TPoint model isn’t very good. The point-spread graph is quite spread-out, and quite oblong. This is probably just another symptom of the same problem causing the above issues. It could also be giving ProTrack ammunition to make bad adjustment decisions, compounding the problem.

Second, I note, when gathering uncorrected data for doing a PE calibration, that I’m getting a large measured declination drift, in addition to the expected periodic error in Right Ascension. I haven’t thought through the mechanics enough to know if this is normal or an indicator of a problem.

Research in support group

I spent quite a bit of time searching through the SB support forums and documents (which are not well organized, and not easy to search, by the way). Here are some miscellaneous points I’ve noted – in no particular order.

  • Polar alignment is important, of course. If it’s off, drift will occur in both axes, compounding error. (I’m quite sure mine is quite good.)
  • Camera rotation is critical when gathering data to analyse PE. It must be close to 0 or 180 degrees. If it’s in-between, PE is masked by being distributed between the RA and Dec axes. (I made this mistake several times, gathering PE data with a poor camera angle.) I note that the software could be coded to correct for this (if given the rotation value), but it isn’t.
  • ProTrack affects tracking. That’s its job; but, when diagnosing problems, start by turning it off so we’re looking at “raw” performance, and to help determine whether ProTrack (or the TPoint model that drives it) are part of the problem.

The most common causes of problems seem to be:

  1. Flexure in system (mirror flop, loose physical mounting, etc.) Externally, I’m pretty sure I have this under control, but there could be something loose inside the mount.
  2. Problems with worm gears, any of:
    • Spring tension plungers needing adjustment;
    • Cam stop needing adjustment;
    • Pivot screws being loose;
    • Various other mounting screws being loose;
    • Rarely: gears or gear assemblies needing service or replacement.
  3. There are some reports of through-the-mount cabling contributing to problems, if the cables tangle, don’t have the necessary slack, bind against internal shafts or surfaces, or “wind and unwind”, contributing force to the expected movement of the shafts.
  4. Tension on the drive belts can also be checked and adjusted. It’s not clear how incorrect tension would cause the problems I’m seeing, but it’s easy to check and it would presumably contribute to longevity of the belts to ensure it’s correct.


To be confirmed by experimental corrections and measurements, my hypotheses for my problems are:

Higher-than-spec PE

I think the spring plungers on my RA worm gear are too tight. If they were too loose, it wouldn’t affect normal RA tracking (but would show up as backlash in slews, and in TPoint). Too tight, however, might be causing gear friction, stiction, or over-sensitivity to minor machining imperfections.

Through-mount cabling could also be contributing to uneven PE performance if cables are binding.

How to test this hypothesis:

  • Check TPoint model for RA backlash, which would suggest “too loose” and disprove this hypothesis;
  • Make a new PE data-gathering run after making the following checks and adjustments.
    • Temporarily remove all in-mount cabling, routing cables on the outside;
    • Adjust the RA spring tension plungers to spec;
    • While I’m in the RA gear box, adjust cam stop and confirm other mounting screws are tight.

Erratic Dec during guiding:

I think the spring plungers on my Dec gear are too loose. Backlash could be causing the non-response to guiding adjustments, and is probably also polluting the TPoint model and causing ProTrack to make adjustments. Occasional ProTrack “adjustments” could be responsible for periodically throwing the Dec error to the other side of the zero axis.

Through-mount cabling could also be contributing to uneven performance if cables are binding.

How to test this hypothesis:

  • Check TPoint model for Dec backlash, which would confirm “too loose” and support this hypothesis;
  • After fixing PE as planned above, do a new guiding test after making the following adjustments:
    • o Temporarily remove all in-mount cabling, routing cables on the outside;
    • o Adjust the Dec spring tension plungers to spec;
    • o While I’m in the Dec gear box, adjust cam stop and confirm other mounting screws are tight.

Plan to deal with it

So, here is the plan to work through this problem. The plan will need adjustments as data come in.

  1. Change mount-to-pier fastening (see post of 2018-4-14)
  2. Remove through-mount cabling.
    (Temporarily – I like it, and hope to be able to restore it once I understand its role, if any, in the problem.)
  3. Check and adjust Cam-Stop on both axes
  4. Check tightness of worm-end pivot screws, both axes
  5. Check that RA and Dec worm plungers are adjusted to spec.
    It wasn’t easy to find out what the spec is: the SB documentation isn’t well organized. The specs are:

    • Cam Stop 1/8 turn back from bottomed out (measured with clutch in “run mode”);
    • Spring tension plungers: 3 to 3.5 turns back from bottomed out. (measured with clutch in “balance mode”)
  6. Check belt tightness, both axes
  7. Do a new 20-30 point TPoint run to verify polar alignment is good
  8. Carefully record uncorrected Periodic Error
    • double-check camera angle is 0 or 180
    • double-check PEC, TPoint, ProTrack, and Guiding Relays are off
    • keep a plate-solved image to verify camera angle and image scale
    • keep the gathered guiding log, and a screen shot of the data gathering graph
    • hoping, with above adjustments, PE will fall in spec; get help if not
  9. Upload PE Correction and record results
    • double-check “west side of mount” checkbox is correct
    • do a PE recording with PEC on but TPoint and ProTrack still off
    • keep the gathered guiding log, and a screen shot of the data gathering graph
    • residual error should be on order of 1 arcsecond
  10. Carefully record an autoguiding run
    • compute autoguider correction parameters with CCDWare calculator
    • hoping strange Dec issues gone; get help if not
  11. Reset TPoint
    • do a new large run (150+ stars) and supermodel
    • verify if it still thinks I have good polar alignment
    • test unguided tracking, with PEC and ProTrack on
    • test guided tracking with PEC and ProTrack
  12. Restore through-mount cabling
    Replace through-mount cabling, one cable at a time, verifying PE and guiding after each. This will be a slow and painful process, that I’ll work at as other reasons to take OTA off mount arise.

Changed scopes for summer

It was clear and warm last night, but a full moon.  So no serious imaging but a good night to adjust things.  With winter skies basically done, I decided to take off my widefield refractor and go with a longer focal-length OTA for the summer season of galaxies and nebulae.  So, the SV80S is put away, and I mounted the AT8RC 8″ Ritchey-Chretién, and rebalanced the mount accordingly.

M67-3x3-10secs--15C-LuminanceThat creates new re-adjustments that are needed, of course.  I re-built a V-Curve model for FocusMax, and got nice round stars on a quick 10-second test image of M67.

I was going to do a new TPoint pointing model, and also use that to re-check the polar alignment, but I became distracted.  I tried a quick test of autoguiding, intending to confirm that my previous autoguiding problems were field curvature on the SV80S.  They were – the AT8 produces a nice flat field, and the guide starts weren’t too badly distorted to use.

The problem was that I tried doing a guider calibration, and it didn’t work.  That ended up distracting me for the rest of the night, as I experimented with the parameters that control the amount of movement the calibration routine injects before measuring direction and distance.  No matter what I specified, even absurdly long displacements, I got the same error:  “Error, insufficient motion in X-direction during calibration”.

Guider-calibration-motionWhich is clearly not the case.  As these two guider frames show (click for animated gif), there is good motion happening in both axes between calibration images.  I puzzled over this for an hour or so, then it clouded over so I shut down and went to bed.

This morning an idea hit me which I think is the solution.  The guide stars aren’t in focus. That, alone, shouldn’t be a problem, as the guiding software should calculate the centroid of each star.  However, there are bright hot pixels in the image.  I bet the guider software is selecting and guiding on a bright pixel, which doesn’t move, not on a star.

I’ll test this next clear night:

  • Adjust focus of guide camera to get nice bright pinpoint guide stars; and
  • Use a dark frame with the guide camera to calibrate out the hot pixels.

We’ll see.

Hmm. Guide Collimation doesn’t look like the problem.

Another clear night – that’s 3 in a row, possibly the only time I’ve had such a chance in a decade.

I spent it investigating my guider collimation problem.

Badly distorted, elongated stars in Lodestar guider in WSG guider port.Imaging with an SV80S (480mm f/6 triplet refractor) I’m getting very elongated stars in my guide camera, seeing through the off-axis guider port on the QSI583WSG camera. So elongated that I doubt guiding will be accurate. And considerably worse than I remember seeing a couple of years ago. What’s going on?

My assumption was that the pick-off prism or the guider camera mount was somehow out of collimation – that it wasn’t orthogonal to the incoming light path. That would seem consistent – the distortion in the guider stars is exactly vertical, as though the prism was out of alignment in the vertical direction, but fine horizontally.

However, before that it occurred to me that the entire camera/guider system might be out of alignment. Better check that first.

Using CCDInspector, I discovered a minor out-of-plane error in the main camera, which I was able to correct by snugging it better into the focuser drawtube.

Main camera field well centred, no collimation errorThen, a re-check shows a field with no skewing off orthogonal, although with significant field curvature (back to that in a moment). So, it’s not the overall alignment of the system.

Guide-Camera-CurvatureNext, I did a field analysis of an image from the guide camera, which confirms what I could see with my own eyes – there is major distortion, almost exactly vertical, toward the bottom of the image field.

Now we come to the problem. I think I need to tweak the alignment of my off-axis guider port’s “pick-off” prism. However, it doesn’t seem to have any such collimation adjustments. It’s solidly machined, as is the threaded guide camera mounting port. It’s not supposed to be out of alignment, and there doesn’t seem to be any way to adjust it if it is.

By now I have run out of evening time and patience, and I have to get up for work tomorrow, so I’m stopping, stumped, for now.

On the weekend I will use daylight to take the camera off and apart and look more closely for some way to adjust or shim the pick-off prism or guide camera mount.

My fall-back plan if I can’t adjust that will be not to worry about it. I’m pretty sure the problem is greatly exaggerated by the rather tight field curvature in this scope – evident in the curvature map shown above. The pick-off prism is very close to the edge of the field, where curvature is strongest. Maybe that’s why I don’t remember having this problem before – because I was using a different scope, with a famously flat field.

With this short focal-length scope I should be able to take 5- to 10- minute images unguided and, if guiding is needed, at this short focal length a separate guide scope should work well.

So, I’m going to have a look at the mechanics on the weekend. If I can find a way to adjust the prism alignment, I will try that. Otherwise I’m going to declare “don’t care” and use this short scope without the off-axis guiding. Its season is nearly over anyway – I’ll be returning to a longer focal-length, and flatter-field, scope for the summer. If I have a guide-star problem with that scope I’ll pick up this problem again, but I’m guessing I won’t.

So, assuming I don’t find an adjustment, next clear night will be spent refining my PEC training now that I have perfect focus, and then I’ll do some tests of unguided exposure length to see what the Paramount can do at a short focal length. I expect to be impressed.

Changed Internet Service Providers

In my experience, good ISPs seem to stay good for 5-7 years, then go sour. That just happened; after many years of good service, my previous preferred hosting company has become poor (probably not coincidentally, just after their shared-hosting division was sold to another firm). After a string of outages and dead-ends with several sites I used to host with them, enough was enough, so I’ve moved. So this site is now up on a new provider (, which seems pretty good so far. The site has been down and broken for a few weeks before and during this change, but I think it’s back to stable now.

-20 is quite cold, actually

Last night was the first clear night in over a month (that I was was available), so I spent it doing some imaging. The 3/4 moon was too bright for anything serious, but it was OK to work on adjusting autoguiding parameters.

It was about -20C. That’s cold. The PC wouldn’t start, so I think I am officially off the air in the cold until those enclosure heaters arrive. The LCD monitor also faded to near-invisibility, and I’m thinking I may need to apply a heater of some kind to it too.  I eventually got the PC going by leaving it powered on with the “can’t find hard drive” message on the screen, until the internal power consumption heated the interior somewhat.

Then I found I was getting strange effects with autoguiding, with the Declination guiding adjustments moving four times more slowly than the RA. I suspect a problem with backlash or stiction in the Declination gear. I did a bit of data gathering to better quantify some of the effects, but I wasn’t having fun in the cold, and called it a night after a couple of hours. I’ll have a look at the Declination gears and backlash during the day on a coming weekend (in a thoroughly-heated dome).

Once I get the heaters working and the declination adjusted I think it will be practical to do some basic image acquisition in extreme cold, but it’s not fun to be trying to do adjustments in those conditions.

Project: Heating Interior of PC cabinet

It’s been cloudy for a month now, so no observing or imaging.

I am continuing my long-term project of setting up the obervatory PC to work well in cold temperatures.

I had originally tried setting the PC up with a solid-state drive, thinking the lack of moving parts would make it an excellent cold-weather PC.  The cold-weather part worked fine, but the PC was unreliable booting from the solid-state drive, and I eventually gave up on that and went back to the original hard drive.

The problem with the hard drive is that when the observatory is cold (below about 0 Celius), the hard drive has trouble getting spinning.  Sometimes it fails completely, giving a BIOS error. Even when it boots, I have to believe it’s not good for it to be spinning up when things are tighter than usual because of the cold. Continue reading

SCT progress: collimation, visual, reducer

After a discouraging evening the other night, trying to do too many things at the same time, last night was more organized and more successful.

I took the SCT back to pure visual configuration and did a careful collimation using an out-of-focus star. More tuning later, using software help, but for now the collimation is pretty good.

Next, I tried a more methodical set of experiments with focal reducers. I couldn’t get the OpTec 0.5x reducer to work in any configuration that occured to me – only in a couple could I bring it to focus, and quality wasn’t good. I’ll try it more later but, for last night, I switched to the Celestron 0.63x reducer. It needs about 100mm distance to CCD so I had to mount the reducer outside the Crayford focuser, which meant a couple of threaded adaptors. That worked, and I was able to take test images and achieve focus. The unusual placement of the reducer means I’m not getting 0.63x reduction or f/6 results, but I haven’t done the math yet to figure out what reduction is happening. (Update: doing a plate solve on an image, this setup is yielding an effective reduction of 0.68x, for a resolution of 0.83 arc seconds per pixel.)

M51 Unguided

M51 Unguided

I took an unguided test image of M51 to get a sense of what this resolution will do. It’s quite a pleasing image scale – once guiding is working it should be good. I was surprised at the rather steep image vignetting that was happening in this setup – substantial flat correction was needed.

Continue reading

Good sky, calibration improving, guiding failed

NGC869-LRGB900It was a beautiful clear day and the sky remained clear through the night, we continue to have unseasonably warm weather, and the moon rose late, so last night was an ideal evening. Before dark, I mapped the horizon height all around the observatory, to configure Maxpoint and other software. Then, after dark, I carefully re-did the Gemini alignment routine on the mount, and overcame the pointing error problems I’ve been having. Finally, I put the 0.8x focal reducer into the imaging train to do some wide-field imaging, and recalibrated the focuser with this new setup.

One test image just for fun: a wide-field view of NGC 869/884, the Double Cluster. It seems to have worked well, including the automated filter changing, and alignment of the colours is good. As an image I’d like more colour depth and better colour calibration, but that will follow. Continue reading

Comedy of Errors imaging session

It was clear last night, after two weeks of cloud, and not as cold as the last clear evening. I opened the observatory at about 8:00 with a temp of -5, planning to acquire one or more colours of RGB for the B33 project I started a few weeks ago.

Not a very successful session – I ended up gathering a small amount of Red data, and several items for my to-do list. Continue reading