dane I would be grateful for any help.
I had mentioned this in my posts in this forum regarding auto guiding in ASIAIR. Just search for them (hold mouse button down on my user name).
The RST-135 has a very large periodic error, but mostly sinusoidal, with a pretty small third harmonic. Its period is about 430 seconds. From a sine function that has that period, you can take the time derivative that is a cosine function with the same period, that has an amplitude with the original (peak to peak error up to 70 arc seconds, so amplitude of up to 35 arc seconds), divided by the period (just use the chain rule that you learned in calculus in high school). The worse case is at the the two zero crossings of the sine function, where the cosine is 1 or -1. The large errors occur at half the period or about 215 seconds apart.
This is the rate of change (derivative) that an autoguider needs to keep up with. Additionally, there is also a lag time for each pulse command.
As a result, you want to use fast feedback updates for mounts that have large periodic errors. Unfortunately, the ASIAIR's guide interface could not do faster updates that 1 frame per second, even when you set the guide camera exposure time to less than 1 second.
However, ever since ASIAIR v1.8 on, it has the capability of doing 2x binning with the guide scope. And with fewer pixels to process, it is capable of 2 FPS if you set the camera to 0.5 second exposures.
I don't know if the mini camera that you use can attain 2 FPS with 0.5 second exposure and 2x binning. If not, you will need to get the pancake version of the camera that has USB 3 interface.
In the past, using 0.5 second exposure was not practical since atmospheric turbulence ("seeing") causes more harm that the good from the faster feedback rates. However, with multi-star centroids, the variance of the atmospheric turbulence is reduced by a factor of 2 for each doubling of stars -- i.e., two stars will give an RMS turbulence that is 0.707 of one star; four stars will have turbulence of 0.5 of one star, etc. I.e., the effect of using 4 stars to reduce turbulence is equivalent to using 4 times the exposure time to reduce turbulence (assuming ergodicity).
So, try to reduce the sampling period of the autoguiding pulses by choosing 2x binning, set the guide exposure time to 0.5 seconds, and give it a shot.
With a 250mm focal length f/4.5 APO guide scope, and an ASI462 camera with 2x bin and 0.5 second guide exposures, I can get to about 0.4 to 0.45 arc second total RMS error with my RST-135 (that sits on top of a stable William Optics tri-pier), and if I add a 685nm IR-pass filter to the guide scope, I could get down to the 0.35" to 0.4" region.
A guide scope instead of an OAG will give you better star shapes to autoguide on. If you really cannot use a separate guide scope, at least try to focus the guide camera as best as you can (use a Bahtinov mask to adjust the guide camera focus after using the same mask to adjust the focus of the main camera).
Because of large periodic errors, while harmonic drives are ideal for traveling and portability, they are less than ideal as a backyard instrument. There are numerous mounts that outperform them at less than half the price, albeit with much worse weight ratios. I am sure people will also eventually discover that with the ZWO AM5 as well (which has worse periodic error shape -- i.e., the derivative of that mount's periodic error curve is quite large because the function is not a pure sinusoidal). At least the AM5 will force ZWO to develop better autoguiding rates in ASIAIR.
Chen
EDIT: Since the periodic error of the RST-135 is almost sinusoidal, with a peak-to-peak amplitude of 70 arc seconds, it therefore has an RMS (root mean square) error of 0.707*peaktopeak/2 = 25 arc seconds RMS. The autoguiding has to reduce this 25 arc seconds RMS down to the 0.5 arc second region to get sufficiently compact stars. That is a factor of 50.
EDIT 2: in case you are curious about the period of the "periodic error," the RST-135's period is one sidereal day divided by 200. This comes to precisely 430.82 seconds. I think the ZWO AM5 copied the same gear ratio, and if so it will have the same period.