If I'm calculating this correctly, the theoretical best performance for my particular ZWO AM5 mount with a maximum period error of 23.6" is 0.136 arc-second error for a 1 second guide duration. So, 0.136" error is what's possible with perfect seeing, no atmosphere, in Earth's orbit somewhere. Here on the ground, ZWO has specified that all AM5 mounts should perform with 0.5" to 0.8" total error or better, given average seeing and atmospheric disturbance. And that seems to be exactly what I am experiencing.
For a detailed walk-through of the periodic error report values and calculations see 10 things you need to know about the custom AM5’s PE Test Report provided by ZWO

The weather doesn't look like it's going to cooperate in the next four to five days, and at the same time, the moon is seriously imposing its reflective properties on the night sky. So, not the best time to try out a new filter—even a narrowband one. I recently bought the Optolong L-eXtreme dual-band filter, specifically for my ZWO ASI071 color camera, which has an APS-C sized sensor. The L-eXtreme filters out everything except two narrow (7nm) bands along the Oxygen III line and Hydrogen-alpha line. I am anxious to try this out with the 8" f/4 Newtonian on the Veil Nebula.

I have been imaging the night sky for eight years with five different EQ mounts, most of it from my backyard in New Hampshire, and I rarely experience guiding this precise. I have been out imaging three nights with the AM5 so far, with reasonably clear skies and average seeing—nothing spectacular. These are some of the lowest total RMS error numbers I have seen from my Bortle 4 to 5 backyard. We're going through a stretch of clouds and rain, but hoping for better weather toward the end of the week. I'll post more guiding results along with some PHD2 logs.

I nearly found out the hard way that the new ZWO AM5 mount is so compact that power and USB cables from the primary camera can easily loop over the altitude and azimuth adjustment knobs at the base and snag, something I never really thought about with a larger mount. This happened the other night, and I just managed to unhook them and continue imaging.
So, a little cable management was in order for the William Optics SpaceCat 51 on the ZWO AM5. The guide camera (silver camera on top) and the filter wheel plug into USB ports on the primary camera (ZWO ASI1600MM-Pro) at the back. Another USB cable runs from the autofocus (ZWO EAF), the red box under the scope. I end up with five cables leading down to the Pegasus Astro Powerbox and the fanless Windows 10 machine running the control software (NINA). What's new in these pics is the SmallRig Switching Plate (1598) and two of the SmallRig Spring Cable Clamps (MD2418). These are designed for cable management on cinema cameras, but work just as well for our gear.
https://www.amazon.com/dp/B08NT68D3Z
https://www.amazon.com/dp/B00X5NXSRQ


Out of the box the DEC axis of my AM5 was a few degrees east, not lined up with the RA axis. There may be several ways to reset the home position, but I used the ASI Mount app on my phone. Plugged in the AM5 hand controller for wifi, connected to AMH_..., and when you tap the AM5 icon at the lower left, there's a "Home Position" in AM5 Settings. Tap "Calibrate" (it will move to what it thinks is home first), then you can adjust the DEC/RA with the buttons in the app (or maybe with the hand controller) and tap "Set" to save that position in the AM5.





The ZWO AM5 is here!
https://youtu.be/DE0lUUPakGE
First Light with the new ZWO AM5 mount
https://youtu.be/ckJPn3W-4Kw
One of the first things I did after powering up the new mount, was to check the ZWO ASI Mount iPhone app, which instructs you to connect to the WIFI hotspot provided by the AM5 (built into the hand controller). I then upgraded to ZWO AM5 firmware version 1.1.1 and Hand Controller version 2.1. This took the mount from version 1.1.0 and the hand controller went from 2.0 I believe--so point release and minor release.

First night out with the ZWO AM5 strain-wave gear/harmonic drive mount, and I am just stunned at how easy this mount is to setup, polar align, and run.
I swung the scope south to IC 4665—I just picked a nice star cluster in Ophiuchus, near the meridian in the south. I cleared calibration in PHD2, ran a new calibration, and then guided for an hour. I ran with defaults, 1 second exposures. Total RMS error through the session was between 0.4 and 0.7 arcseconds. And then I jumped into a bunch of 180 second subs of the Veil Nebula in Ha, watching the guide graph hovering in the high .4s and low .5s.
This was my first time setting up the AM5 for a night of imaging, and I think I was expecting something weird or some tech idiosyncrasy to appear, because that's the way it always is. New devices are just going to behave in ways you might not anticipate. This was my first experience positioning the mount, scope, and tripod, which I easily carried out to the deck with everything connected and ready to go—tripod, mount, scope, everything. Polar alignment in NINA is simple and automated, and the AM5 azimuth and altitude adjustments are smooth and easy to dial in. I had the mount aligned and ready to capture data in two or three minutes. Again, I'm a bit stunned at how uncomplicated this was—I was expecting something to go wrong, but the AM5 was just doing everything right, with accurate slewing, guiding, tracking, and I was sitting on the couch in the living room taking 180 second subs on the back deck.
Some surprises: I didn't expect the AM5 to be so compact and transportable—and I say this knowing my primary purpose for buying the mount was portability. It's crazy how easy this is to move around, and I think it comes down to the counterweights. They are a such a fundamental part of the system that it felt like I had missed something in setup, but after one night under the stars with the AM5, I'm thinking—unless you have a massive scope—counterweights should be a thing of the past. We have the technology. Why do we need them?
I did try out the new mount with Ekos/INDI and ran into an error connecting the AM5—an error setting the UTC Offset. I didn't dig into it, but it looks like a similar error—an old error—with the LX200 instruction set where Ekos is using a float and the mount is expecting an integer for the offset? Seems pretty minor to me, and easily fixable, so I will come back to Ekos soon as I have some time.
NINA/ASCOM worked flawlessly with the ZWO AM5, at least with my standard workflow: polar alignment, slewing, plate solving, guiding, focus, capture. I'm running Version 2.0 HF1 BETA015 on a fanless Windows 10 machine.
Finally, the skies weren't perfect, seeing was maybe a little above average, and the mount performed with those total RMS error numbers, .4, .5, .6 arcseconds. Amazing.


Here's 36 x 180-second subs, no calibration frames, stacked in DSS, stretched in Photoshop CC2022:
