Wednesday, September 10, 2014

Full moon backyard testing

Preparing for Calstar 2014 

I've set up the gear in the backyard to test a few problems.  I set the goal to identify the issue with flexure, to collaborate with Carl and Marek to find and fix the problem experienced with DEC guiding, and to push the camera cooling to the next level.


First, the flexure issue was addressed.  I felt that an important step would be to swap the Gemini 2 computer with the Gemini 1 to see if the problem was system-based.  To do this, I had to get a baseline of performance under Gemini 2.  This required a night to establish.  Then swapped the Gemini 1 in minutes.  Ugh, the challenge of re-learning the single line interface and button chording was embarrassing.  In the end, I was able to get it working in one night. However, the clouds moved in before I could get any test subs.

Challenge that was immediately present was that guiding over serial line via the Gemini control panel was significantly slower than guiding over ethernet with Gemini 2.  So I swapped back to the ST4 jack on the SSAG.  Testing the serial port cost a night.  Testing the ST4 was another night.  So far, this test required 4 nights before clouds as the moon climbed higher towards full.  It was getting harder to see Eniph (Enif) as the sync star near my test target.

End result after all this testing?  It's not system-based.  The RA drift is independent of the computer.  Thus, with all the work for stiffening the setup, there's still movement.  Not much, but enough to smear stars with the long subs that I prefer to use.

Thus, I've thrown in the towel on flexure.  I'm no longer going to try to use 20 minute sub exposures.  Now, I'm planning to use 10 minute exposures and have started gathering a library of dark frames.  The shorter exposures make getting these files much faster and the test calibrations I've made show good promise.

I could revisit the issue and try replacing motors as it is the only remaining common element.  I think that this is unlikely to fix things as I've seen an interesting wander in these 10 minute subs.  Over 10 minutes (vs 20 minutes), there is a wander in DEC and RA that was masked by the longer subs.  With the shorter subs, I can see fine detail as the image center wobbles.  Some of this might be balance or cable drag, so I'm hoping to address this soon.

DEC Guiding

Second, I've been collaborating with Carl and Marek to identify and solve issues with DEC guiding.  Both of them had been working on the Sadr region earlier this summer and were having strange performance.  I noted that this might be due to tracking through the zenith as DEC imbalance is negated through this space.  The scope will hang mostly vertically and thus any front-back imbalance that would otherwise stabilize the axis is missing.  My setup has a side-by-side arrangement, so I can offset the balance there a bit more to provide imbalance through the zenith. 

During a night at Montebello, Marek and I looked at the gear slop in his DEC axis and felt that it was similar to mine.  There was a small amount of slack in the worm and ring gear mesh.  This is the kind of slack that is easily controlled with balance.  It could be tightened up at the expense of possible motor lag or stalls.  We pursued the idea of adding weights or shifting the scope on the axis to add more or less imbalance.  In the end, we decided that a significant amount of imbalance would be required for his setup, more than was easily attained given the size of the dovetail plate he uses.

At home, I tested the idea of playing with gear mesh.  I tried tightening the DEC axis to remove more of the slack that I saw on my system.  During this experiment, I noted that even though the motors did not stall, the movement was jerky and showed a lot of vibration.  This was not ideal and so I backed off on the mesh.  Instead, I concentrated on the idea of maintaining balance through the above flexure test. 

Once I'd committed to the shorter subs and switched back to the Gemini 2, I went back to checking the DEC balance.  Mallet and wooden stake in hand, I tapped on the dovetails to move them about.  The dovetail that the SV4 rides on is so tight that I have to use a mallet to get it to budge.  I recentered the balance as much as I could, then offset it to be camera-heavy.  The natural position of the axis is flat when pointing to the zenith.  I hoped that this would be similar to what Marek experienced.

Last night, I noticed real problems with DEC guiding.  No subs were usable, all had streaks.  Only by adding some weight to the camera end of the setup was I able to reduce the wild swing of guiding.  I did see interesting behavior when using Maxim.  When leaving the "stiction" option turned on, the guiding would mostly hold on one side of "0" with periodic spikes to the other side.  Eventually, these spikes would trigger a long slope as the motor would have to spin to take up slack.  This was the point I added the weight.

I also tried leaving the stiction option off, letting Maxim constantly correct as the axis wobbled.  This was a little better, but still was showing streaks 2 arc seconds long as the guiding went back and forth.  Pretty awful results, too.

This seems to be the product of three problems.
  1. Cable drag/stiffness: The umbilical cable has a reasonable amount of stiffness.  This is not surprising given the size of the wires passing through it.  There is a 10 gauge two conductor power cable, USB cable, power cable for the camera (ethernet cable), and power cable for the dew heater.  This is pushing against the mount in DEC, offsetting the imbalance.  I can note the difference when positioning the scope on either side of the meridian.  When the mount is pointing to the western sky, the cable bundle is pushing hard.  On the eastern sky, the cable bundle is not pushing.  I originally thought this was not a big deal, but now I see that this is likely the cause of the wobble.
  2. Slack in the gear mesh: I could see the lack of tightness in the 9x50 finder scope as there was a lag when centering stars.  This is usually never seen as I have the imbalance much stronger.  However, last night while the balance was more neutral, the slop was significant.  I tried turning on the TVC option and setting a value of 25 but this lead to horrible guiding as Maxim wasn't able to control the star.  This was even after recalibrating.  The TVC option is a way that the Losmandy system can spin the DEC axis faster than guide speed when changing directions.  The value is the number of milliseconds that the motor moves at 16x tracking rate to take up slack.  This is a great tool when manually guiding or visually centering, however computer guiding is unable to handle the aggressive moves.
  3. Polar alignment: There really should not be as much need for DEC axis movement.  I turned the controls off when it was pointing 1 hour past the meridian and saw significant drift within a short period of time.  The drift I saw was about 4 arc seconds in 2 minutes.  This should be reviewed with drift alignment to ensure that it's within the range of normal.  The PolarDrift Calculator suggests that I can allow up to 2.6 arc minutes of misalignment before rotation is visible over 60 minutes.  I have a sneaking suspicion that the mount has settled and thrown off the original alignment from a week ago.

I want to address these three items today for tonight's exercise.  While it's better to check one thing at a time, I feel that I have a good idea of the effect of each part.

I've reviewed the gear mesh.  I pulled the worm off the mount and noted that the grease was black and thin.  Also, the ring gear was mostly bare of grease!  I cleaned the gears, lubed the worm and ring gear with just enough grease to fit into the grooves.  Also, I checked the bearing blocks to ensure that they offered clearance for the rotating coupler with the motor shaft.  Once reassembled, I tightened the mesh more than before, leaving a barely felt slack.  When testing the slew of the axis, the motor sound did not strain, nor was the movement jerky.  I also tested slack by sighting a distant house on a hillside and moved the axis in DEC at centering speed.  I could not see any slack, however, this could be due to the slight camera-heavy balance or cable stiffness.

Regarding the cable routing, I'm considering a different path along the mount to put the loops of wire.  Currently it's passing to the north and then up between the scopes.  I think it might be better to approach from the south and have a loop there.  So long as there are no catches for the loops to snag on motors or other parts.

Camera cooling

Lastly, I've added more capacity for cooling on the camera.  I've added a water cooling system to better handle the waste heat and have moved to larger, stacked peltier devices.  The power budget has suffered but this is not a big concern yet it does require that I address the issue with voltage drop over the extension cords.  More on this in another post!  Suffice to say, the cold finger to the camera is very frosty and this is good.