What Happened?

I recently acquired a new Losmandy G11 Mount to complement a new C11 ota. The primary aim was to take colour ccd images at focal lengths from 1700mm and higher. Having had great success with unguided imaging and my old Vixen GP-DX mount (capable of consistent perfect tracking over 3-minute exposures at 1260mm f.l. or perhaps more-importantly at 2 arcseconds per pixel), I fully expected the G-11 to way surpass the performance of the GP-DX but it didn't.

The first thing I noticed was a glitch in RA which saw stars jump by around 6 arcseconds, every 30 seconds or so. The end result of this was that any image of greater than 30 seconds duration had double stars in it. Sometimes triple stars. Most of the time, the shift was only in RA, but occasionally, there would be some dec movement as well. Sometimes mutliple stars would be in a triangular formation. This dec movement almost certainly has to be atmospheric as the mount is polar aligned and running in polar aligned mode, should not be tracking in dec at all.

What's going on?

Scouring the internet for information (as always), I found a host of other GM8 / G11 owners who's mounts suffered identical symptoms (same as when I started researching my C11 problems). Actually there were so many others, that I started getting that sinking feeling deep inside. (the first of several).

Suggested causes and fixes from various sources.

A lot of suggestions for the cause and fixes were forthcoming from posts to the newly-formed Losmandy egroup, emails with Scott Losmandy, Peter Ward of Advanced Telescope Supplies where I purchased the mount and Chris Houghton of Astrometric Instruments, creators of the Sky Walker II telescope controller.
Here is a summary and some notes relating to them. There is a summary of results provided for the ones I've tried. Each suggestion is listed in order of difficulty and there is usually an official recommendation which comes from either Peter Ward or Scott Losmandy. Chris Houghton was very forthcoming with suggestions and advice.

1. The Oldham Couplers (that connect the motors to the worms), are sticking during normal rotation.
Suggestion: Lube all connections to the nylon part of the coupler.
Astrometric: I found this one in the Sky Walker II manual.
Results: I lubed mine with 'Optimol White T'. Initially I thought the double-star problem was gone, but it soon returned with a vengeance. A re-lube didn't help, so I suspected we had lucked onto a good part of the worm, good balance point or something else.



The 3-part Oldham coupler joins the motor to the worm.

2. The RA worm and worm gear need aligning.
Suggestion: Adjust the RA worm / worm gear alignment ensuring minimal endplay by squeezing the bearing blocks together before tightening the block mounting bolts.
Officially: As Above, but also allow a liberal amount of backlash and ensure there is no binding of the worm and worm gear at any position on the worm gear.
Result : I have tried this several times and gotten better at doing it but it hasn't fixed the problem.



The cleaned worm and bearing blocks.

3. The RA worm is floating between the walls of the RA worm gear teeth.
Suggestion: Allow a slight imbalance on one side of the mount so that the worm and worm gear teeth are loaded together.
Officially: As above, recommended 1 to 2 lbs (~1 kg) of imbalance only on one side.
Astrometric: The need for imbalance is because the Losmandy worms are not spring-loaded onto the worm gear. PEC training is not interchangeable from one side of the mount to the other. Because of this, there is a need to re-train PEC after swapping sides because the worm gear is touching the opposite edge of the worm.
Result : Surprisingly, this had a dramatic effect. On my first night of trying an imbalance (weight not measured but probably about 500g), on the telescope side when pointing in the east. The scope performed better than I could have hoped for three hours. The unguided G11 tracked perfectly taking only 3 mistracked images out of 40. The M20 image below was the result and is made up entirely of unguided 4-minute images. So the question is 'Why aren't I happy?' Because this incredible night of perfection lasted just that. One night. After turning everything off and leaving until the following night, the mount had returned to it's old ways. And the only things I had changed were my clothes. Very mysterious stuff.



The RA worm and worm gear meshing.

4. The lube in the RA worm and/or worm gear has become dirty, hard and/or unevenly distributed.
Suggestion: Disassemble the worm assembly, clean with WD40 or similar, re-lube with a general purpose lithium-based grease.
Officially: Ok, but check warranty considerations before messing with the mount.
Result : I cleaned and relubed both the worm and gear, cleaning (but not lubing) the clutch plates in the process. The result was much smoother manual movement, and apparently less of the sudden large movements in RA. They still do occur though, even after accurate manual pec training at three times the imaging focal length.

5. One or both of the RA worm bearings is dicky.
Suggestion: Swap the worm bearings with the Dec worm bearings.
Officially: As Above.
Result: Swapped the DEC and RA bearing blocks and tested again. No difference. Bought new swiss ABEC-5 bearings and replaced the RA set. No difference. The bearings are R4-ZZ and available anywhere for anyone wanting to try it.
Miniature Bearings have a chart describing the ABEC rating system, and can mail-order high-quality bearings.



One of the RA worm bearings is visible in the bearing block that holds it.

6. The RA worm is rough.
Suggestion: Swap the RA and DEC worms (and worm bearings to incorporate 2 above) to see if the Dec worm is smoother.
Officially: As Above.
Results: Entirely different PE characterisitics as expected, but by no means better.

7. The RA worm gear is rough.
Suggestion: Haven't seen this mentioned anywhere but the Losmandy page.
Losmandy : Swap the RA and DEC worms and worm gears. This would be a big job.
Online Suggestion: Lap the worm and gear with carborundum paste to mate the two properly.
Officially: Don't do it! The worms are coated and lapping will remove the coating in an uneven way that matches the current configuration. If the worm/gears are later reoriented performance could be worse than ever.

Results

Here is the Trifid Nebula taken on that strange night where everything worked. Maybe all of my problems are atmospheric and this was the only still night this year? This image is made up of 4-minute unguided exposures at 1760mm f.l. Taken with a Starlight Xpress sxl-8 ccd camera.

Click on the picture to enlarge it

So how can this happen once only? You'd think that if it works perfectly for one rotation of the worm then that would prove that there is nothing wrong mechanically, therefore any other problems must be external.
It's this sort of inconsistency that I am trying to track down. PEC really only deals with errors on the worm. What about errors in the worm gear? Even though the worm turns once every 4 minutes, the worm gear only turns once in a day.

I had a close look at the RA worm gear and discovered a score mark on about one-third of it. I have no idea if this was caused by me, or happened before I received it. The mount was bought new, and even though I have adjusted the worms a dozen times, I have never detected or seen any hint of binding. You'd think when a score mark like that is being created there would have been some indication of it occuring. You can see it in the center of the picture. The score starts above center on the gear teeth and continues out of the picture to the right for about 90 degrees of the gear surface.


Scored RA worm gear.

Scott sent a new gear and worm over and I had high hopes that this would solve my problem once and for all. But it didn't!

Measuring Performance

On request from my Peter Ward from Advanced Telescope Supplies, I put together a graph of the G11 tracking accuracy. A star image was taken approx. once every 1.5 seconds, of a dim non-descript star in Scorpio, for one turn of the RA worm of 4-minutes. The result was 224 images covering one worm rotation.
Using the registration feature of Mira V5.0, a list of movements of the reference star in RA and Dec for each image, from the first image were logged and loaded into Excel to create the following graph.

Click on the chart for a bigger one
There is about +- 3 arcseconds of declination movement visible. This must be attributable to atmospheric effects because the mount was polar aligned and tracking only in right ascension.

The R.A movement shows periodic errors ranging from -13 to 7 arcseconds. If we accept +- 3 arcseconds due to seeing, then we could say that the the periodic error is around 15 arcseconds peak-to-peak, or approximately +-8 arcseconds. I would say that this is a quite acceptable figure. Except for the the fact that the total periodic error isn't the problem, it is the sharpness and magnitude of the individual errors that is the biggest problem, rather than the peak-to-peak error. For example, about 30 seconds after the run was started, there is a sharp spike of about 8 arcseconds over a 3 second period. This is typical of the kind of effect that produces double stars.


Click on the chart for a bigger one
Above is a composite of a 16-minute test run made up of 4 overlapping 4-minute runs. Each 4-minute turn of the worm has been seperated and aligned. The shape of the periodic error is obvious, but has expanded in magnitude to the first plot above. Since the first plot, I had relubed and readjusted the worm and worm gear. But there was a lot of endplay that I had neglected to adjust out when the worm was re-installed. This has since been corrected. Note the odd differences between Runs '1 and 2' and Runs '3 and 4' between 57 and 100 seconds. Not sure exactly what happened there, could have been a lumpy bit of lube, or an ant being crushed between the worm and gear....

The Next Step

It sure is lucky I happened to be single at this stage in my life because the amount of time I have spent messing with this would surely have tested the strongest relationship....I have dismantled, worked-on, assembled and re-tested so many times, I have lost count. Finally, after ten months I had decided that this mount was just not capable of taking more than a 10 second unguided exposure at image scales of less than 2 arcseconds per pixel. For my purposes this is totally unuseable.

I started researching other mounts and tracking methods. I had two options. Take the next step up in mount quality. Namely, an Astro Physics, Takahashi, Software Bisque Paramount or Vixen Atlux. The price difference in taking this next step is quite substantial and I decided I couldn't do it.

Adaptive Optics via the SBIG AO7 and self-guiding SBIG ccd camera was another alternative for dealing with fast tracking errors. After researching the cost of buying an ST7 with an AO7 and CFW8, I discovered that this alternative would cost as much as upgrading the mount. So alas, not a viable option.

Not 2 weeks after deciding to sell the G11 and save-up for a new mount, by some remarkable circumstances, I managed to pick up a very, very cheap, brand-new SBIG AO7/CFW8 combination at about 75% off the normal retail price! (no, they weren't hot!). This turn of events had forced my hand and I was obliged to get the ST7e which was required to control the new devices.

Dew Heaters and Tracking

Several more months had passed and with the new imaging system now up and running, I was absolutely amazed at how well the AO7 kept guidestars centered. The real-time guidestar display and error positions from Maxim DL were excellent, and immediately showed me a problem that was previously hidden by the resolution of my test data. A spike every 2 seconds, that was making it difficult even for the AO7 to correct.

After several days of elimination, it turned out to be the Kendrick Dew Heater. A little more experimentation showed the duration between spikes was determined by the power setting of the dew heater. The lower the power, the longer the duration between spikes and at full power, there was no spike!

A bit of a discussion started up on the sct-user mailing list about this phenomenon of RF being generated by the cable connecting the dew heater controller to the dew heater element. If you are a member of sct-user, read about it here. Otherwise you need to join yahoo groups and subscribe to sct-user before you can read from the list archives.

With this final hurdle out of the way, I could finally take well-guided images at focal lengths up to prime-focus.

Conclusions

The G11 has a long-standing reputation as the best available mass-produced mount for the serious amateur in it's price range. It has been used for many years by many people and has taken some excellent pictures. I don't know whether I got a dud, or whether my particular mount is working optimally, but I am still to this day, unable to take any images without the AO unit. At f6.3 the C11 yields an image scale of 1 arcsecond per pixel, and even with self-guiding and (and no AO), fast tracking glitches still exist that just cannot be handled by moving the mount alone. I know for a fact that there exists many others out there who have G11 mounts that track very well without AO, so my situation is neither isolated nor absolute. But still, the performance of the mount alone is inconsistent at best. My only real comparison is the last mount I owned. The old GP-DX mount consistently gave 3-minute unguided tracking at 2 arcminutes per pixel for over a year. Now thats what I call accurate tracking! Even at this image scale, the G11 cannot compare.

After a lot of time and effort, I am finally happy with the performance of my tracking system as a whole. Although I am convinced that the AO7 has tamed the beast, I have to concede that the cause of my mistracking has not been identified. But right now system is capable of consistently taking images at image scales of down to 0.6 arcsecond per pixel with guide times of up to 20 minutes. I couldn't hope for better than that !

I'm interested in your comments
If you have any comments, suggestions, links or stories relating to my particular issues or the G11 in general good, bad, past or present, please mail me!

Links to other G11 sites

The Losmandy G11 Page
The Online G11 user Manual
The unofficial G-11 Tuning page
A G11 Review
ATS G11 Overview