If there is a DEC error when flipping the meridian, I would suspect mechanical flexure someplace in the optical axis.
From: ap-gto@... [mailto:ap-gto@...]
Sent: Wednesday, September 10, 2014 2:57 PM
Subject: Re: [ap-gto] Maximum SAFE shimming thickness under a DOVELM162 clamp set ?
Just a follow up on the mostly successful shimming process on my RCOS RC-14 truss.
Originally, the star was completely off the CCD frame after a meridian flip. I inserted a 1/4” wide x 0.032” thick brass shim between the truss and its attached male D-Plate, and managed to bring down what was a huge 27.7 arc-MINUTE shift of the star from the center of my STL-11000 frame, down to a nice 3 arc-minute. Looks like I only need just another 0.004” of shim to bring it dead center.
However, one thing still puzzles me.
Nothing I have tried with various thickness of that shim, did anything for the “north-south” frame error, on a meridian flip, for a target on prime meridian at the equator. With the RA error much improved, the star remained about the same DEC distance from centre frame, after each change of shim thickness. I still have an 8 arc-Minute frame error in DEC, every time it flips. The AP900 has a PemPro-adjusted polar alignment down to about 10 arc-seconds drift on both AZ and ALT axes.
Before anyone repeats the common suggestion - I am NOT interested in solving this by using a pointing model. I would prefer to optimize the mechanical fit, before generating an all-sky pointing model.
The only thing I can think of is that the “orthogonality error” is not only perpendicular to the DEC axis – as is typical in such cases - but also has an OTA axis “twist” component, producing a constant DEC shift – in addition to the ubiquitous RA shift. In effect, I hope be able to cancel both offsets, if the shim had a sloping thickness, raised slightly, at one edge of the D-Plate.
Does that make any sense – compensating two offsets? One might think that an RCAL would eliminate that DEC error, but it doesn’t seem to.
If so, taking the “standard Losmandy D-Plate” channel width as 2.963 inches wide, and using that as the “lever-arm distance” to raise one end of the shim,
it would have to be 0.003 inches lower at one edge of the brass shim, than the other to nullify my (+/-)4 arc-minute RA-axis orthogonality (twist), which is pushing the star a 8-arc-Min. (north/south) off centre frame.
This is calculated using the same equation and method as was in resolving the DEC axis orthogonality problem.
A shim this thin, CANNOT be machined with a “slope”. So, perhaps this could be done by first inserting the full D-Plate width shim #1, (to fix DEC axis orthogonality), then, adding a (0.003 inch) “dimple” to one edge of that shim (to fix the twist which results in the meridian flip’s, DEC shift).
This “additional twist correction shim” might just be a short piece of suitable gauge copper wire, possibly slightly hammered flat to achieve the required dimple height, and then inserted longitudinally, on top of shim #1, just to raise one corner of that shim.
The question remaining is determining which D-Plate edge - (the one toward or away from the RA axle) - should be raised by that additional “shim wire”.
Has anybody considered or tried this two-way orthogonality adjustment? Then again, are my assumptions valid, about a possible OTA attachment “twist”, as being the cause of the remaining DEC error?