Test series of the original Losmandy Gemini gearbox and the high precision McLennan gearbox using the Oldham coupler. The new block assembly uses a longer Ruland flexible disc coupler and due to a thicker mounting plate it's not possible to use the original Oldham coupler. I made a temporary coupler to use the Oldham with the stock gearbox for this test. PemPro V2 was used to measure PE, generate PEC and plot the performance results.

This series of tests include:

• Compare the stock Losmandy Gemini gearbox to the McLennan high precision gearbox using the Oldham coupler

A solid piece of aluminum stock is inserted into to the Oldham end and a cut down roll pin was used as a coupler clamp to attach the worm shaft to the Oldham coupler. This roll pin offers very high clamping force but is very difficult in install. This makeshift adapter was used to test the Oldham with the stock gearbox only because the new worm block spaces the gearbox too far back.	Six passes of data collection were taken using the stock Gemini gearbox with Oldham coupler.
A graph displaying the frequency spectrum. Note the .11 arc second peak gearbox noise at fundamental C which is the focal point of this test. Because it doesn't fall on a fundamental frequency this peak can not be successfully removed with PEC.	Drift fitting set to 'none' to show uncorrected PE

Performance with PEC enabled

   Before uploading, the PEC profile was modified with a -0.5 sec delay to offset file download and processing times. This setting was found by changing the delay and repeated testing to achieve the best star RMS profile as measured in CCDSoft as well as visual.
Afterwards 3 cycles of data were taken followed by a 4 minute (1 worm cycle) unguided exposure with PEC enabled.

With the new PEC profile uploaded and enabled, 3 worm cycles were measured using PemPro and the results displayed.	Frequency graph showing PE reduced. More cycles should have been taken to average out the seeing conditions and better show the fundamental at C if the weather permitted..
Drift fitting set to 'none' to show the uncorrected PE with PEC enabled.	Putting numbers aside I want to see the visual results. Shown here is an unguided 4 minute image with PEC enabled taken at 1856 mm F/L. Using the cursor information tab in CCDSoft a stars RMS profile between x & y can be measured. Ideally this should be the same for both axis.

Oldham coupler with McLennan high precision gear box tests
  The McLennan gearbox came with a longer shaft that allowed the Oldman coupler to be used instead of the makeshift rollpin. Alignment of the motor and coupler was done by rolling the worm, examining and adjusting until I could see the least amount of collinear movement of the nylon coupler disk.

Side view of gearboxes showing shaft lengths	McLennan gearbox with Oldham coupler. I had to install a washer under each mounting ear for a proper fit due to the longer shaft.
Six passes of data collection were taken and analyzed.	This spectrum plot shows the integer fundamental at 'C' has been completely reduced. .
A zoomed in view showing the gearbox noise at 'C' has been eliminated.	PEC curve with no fitting applied showing the periodic error values.
PEC curve with fitting applied showing the corrected periodic error values.

Performance with PEC enabled

Another 6 passes of data collection with PEC enabled.	Frequency graph
PE curve not fitted. This curve wasn't added to the mount's PEC table but is only a reference.	Shown here is an unguided 4 minute image with PEC enabled taken at 1856 mm F/L. Using the cursor information tab in CCDSoft a stars RMS profile between x & y can be measured. Ideally this should be the same for both axis.

Summary
  The uncorrected PE results from both gearboxes were similar but the McLennan unit clearly eliminated the gearbox noise at fundamental 'C'. Though this was at a low level originally with the stock gearbox (.11 - .16 arc seconds) every little bit of reduction counts when working with low numbers so eliminating outside error contributions from other extraneous sources other than the RA worm helps. This peak also didn't occur on a repeatable fundamental so PEC programming will not remove it.