And, after all of that, I think my scope may have an actual hardware problem that might make me have to return it, if it doesn't straighten itself out.
There is definitely something wrong with CH4, and I'd like other owner/users to check their scopes to see if they have these problems too.
When any two of the other channels are turned on, even if no signal is being applied to them, CH4 gets a 1 volt or greater positive offset when in 5v/div or greater vertical scale. There's a relay that clicks when going from 2v/div to higher values, and when this relay clicks the offset hits. But only if two of the other channels are on. This happens whether CH4 is in AC or DC input coupling. But sometimes (rarely) it doesn't happen, seems almost OK but is noisier than the other channels, then after several seconds it will "jump" to the +1V offset. This is a real drag because the 5V/div setting is used a lot (by me anyhow) and a 20 percent positive offset error is unacceptable. It's fine at settings of 2V/div or less -- except see below.
Also, when no inputs to anything are plugged in, and CH4 is set to 10mV/div (the most sensitive setting with 10x probe attenuation selected) and AC-coupled input is selected, the channel jumps to _one full division__ of negative offset. Sometimes even more. Over 10mV of negative offset in AC coupled, in the 10mV/div range on CH4 !! This is regardless of whether other channels are on or not, it's a steady thing and doesn't matter what other settings are in use. And of course it is still there when an actual input is fed to the channel.
These problems are not due to probes or signals, they can be demonstrated with nothing connected to the scope at all.
The other three channels are fairly well behaved, there are slight offsets in certain conditions but acceptably small. The CH4 behaviour seems to indicate some kind of hardware problem, _I think_. I've tried self-calibration several times, am just about to test after Yet Another self-calibrate (it takes about 18 minutes for the self-cal routine to run).
I hate returning stuff. Especially after the long wait and the tension of UPS "delivery". I wonder if I should just "suck it up" and consider CH4 unreliable in certain conditions. One (even when the one is ME) hardly ever needs 4 channels anyway ... but GRRRRRRR.....
ETA: OK, after the latest self-calibrate run, the CH4 set at 5V/div now has about a 3/4 volt _negative_ offset when DC coupled, and a 2 1/2 volt negative error when AC coupled, and is noisier than the other channels.
Scopeshot below shows all probes connected to the scope's Calibrator output. CH1 and CH2 are DC-coupled, CH3 and CH4 are AC-coupled. Note the negative offset of the CH4 trace. CH3 is about right. Acquire mode "Average" with 2 averages selected, auto memory depth.
ETA2: And now, with no inputs (probes unplugged) and channels set to 10 mV/div, CH1 and 2 DC coupled, CH3 and 4 AC coupled, I get _30 mV_ of negative offset and lots of noise on CH4. See the second screenshot below. All four channels are set to have baseline at the center of the screen. There is definitely something wrong somewhere.