Oh, you guys flatter me, you really do. I appreciate the unstated assumption here that I know what the <bleep> I'm doing. Thank you for that!
No, meridian flip is not at issue here, but thanks for suggesting that, I'll be sure to check that setting. Since my best-available dark site is a state park that officially closes at 2300, imaging for that long is rarely an issue! What's happening is that when I use Ekos to GOTO an object, it often sends the mount to seemingly random, wrong places. The hand controller usually does a very good job; Ekos, only sometimes. No idea why. Doesn't seem to correlate with whether or not I do a plate-solve-and-sync, either. I would rather just run everything from my iPad, but maybe I should use the hand controller to get in the neighborhood, plate-solve, and then tell Ekos to fix the result.
And some of those journeys are apparently trying to go through upside-down (i.e., RA counterweight shaft pointing up) on the way. Hence the noises: The clacking/grinding from the mount, and the "RACKUM FRACKUM FRAGGING BRACKUM" sounds from the astronomer. Also jumping up and down, and arm-waving.
As for my guiding -- let me see if I misplaced a decimal point in my original post -- nope, I didn't mention the values. On this most recent evening I was initially hard-pressed to get RMS values under 15, I say again, one-five seconds of arc. I had the most beautiful little star-trails on my images. Well, the two I managed before my camera battery died, that is. (It was a beautiful and incredibly frustrating night.)
So yeah, that wasn't something needing tweaking, that was something fundamentally dinked. (Though "dinked" was not the word I used. Aloud. Several times. Curiously, on these little soirees I usually wind up with the whole park to myself after awhile. Darndest thing, but I digress.)
As it turned out, by powering everything down including the Raspberry Pi, firing back up, running PHD2 instead of Ekos's internal guider, and forcing a recalibration near the celestial equator, I was getting more like 1.5". For me that's a really good night, but I doubt it's the CEM25P that's at fault there.
For reference, meteoblue forecast 1.45" seeing at my location for that time, and the PHD2 Guiding Assistant's high-frequency star motion values were a little under that. So I eventually managed to get RMS values within shouting distance of what the seeing was probably running.
I also pinged iOptron about it, assuming that I'd hear from y'all long before they replied. As it happens, they responded immediately, and reassuringly, agreeing with Bob: The noise is just the stepper motor skipping steps, and it doesn't damage anything, at least for short periods. Whew!
Thanks again folks.
Edited by fewayne, 18 May 2018 - 12:36 PM.