Kevin_A all i got was the dreaded comb looking pulse pattern
Yep; a constant amplitude comb indicates that the pulse duration is not long enough to counter the slope of the PE curve at that moment.
The problem with that ZWO sheet of paper is that the expanded view shows the region for the worst-PE amplitude, not the region of worst-PE-slope. I.e., it is useless for guided operation, and only useful to gauge unguided behavior (keeping a target within an eyepiece FOV). There is probably some region of your mount whose slope is way higher than the slope in that expanded graph on the piece of paper.
You can try to empirically arrive at the worse case slope by slowing increasing the max pulse until that comb goes away at all hour angles. Remember that the worse case slope will occur at the same hour angle each night where the gear reaches the same angle relative to Meridian, and not directly to the RA of the target. I.e., it is the angle of the gears.
This new (real) worse case slope will then also tell you how well you can really guide. In my case, with 0.25x sideral guiding, I have no comb when the max pulse is set to the 150ms to 180ms range. Since 180+180ms is still smaller than 0.5 second exposure time, I am safely able to profit from using 0.25x sidereal rate to reduce the sawtooth amplitude.
So far, the best bang for the buck I get is to use a slow 0.25x sideral rate. That then determines the max pulse duration. Next, and just as important, is to be able to get undistorted 12-star guiding. Near-IR gives only 20% or so better, and is not really better when the 12-star distribution is good -- so I have stopped using near-IR since I often can find fewer stars. But with the QHY678 camera, I am going to eventually try near-IR again.
I have no idea if the Chinese strain wave gears have as well controlled slopes as the Harmonic Drive LLC gears, since ZWO does not reveal them either. You can't just willy nilly pick a gear for guiding a mount without understanding autoguiding.
I have been getting good enough guiding that I don't even look at guide graphs anymore. I will eventually test some of my guiding ideas in INDIGO (and thus, the QHY678 camera).
Right now, I am working on a different software project with INDIGO:
Naming the program Nut (pronounced "Nuit") after the Egyptian goddess of the day and night sky. My second generation All-Sky camera. The first generation was just an ASIAIR with a 178MC and a CCTV lens. This one is a Raspberry Pi 4 running INDIGO Sky, with an ASI294MC (uncooled) and a Samyang 8mm fisheye for the Sony E-Mount. A single PoE ethernet cable provides both data and power link.
The large window (enlargable) is the main GUI of the all-sky camera. The lower right window is the setup window. In the setup, I can set the camera parameters (exposure time, Gain, Offset, and a two-parameter Histogram control) for Day, Twilight and Night. The main window has a single slider as GUI that interpolates between these three parameter sets so that I don't have to play with changing gain, exposure and histogram as the sky condition changes. Just move a single slider. (By the way, a commercial product will not be able to use this interpolation scheme for a display/camera since it is patented by Apple -- I am one of the authors of a couple of the patents associated with this -- think "True Tone" in an iPad :-).
Future experiments on autoguiding can be done on top of this Nut code base. I had earlier written a Cocoa-like framework for INDIGO that I called Horus that Nut uses (Horus another Egyptian god of the Sun and the sky), so writing code for INDIGO is sort of trivial for me.
Chen