Btw these are all with downstream target SNRM set by AA to 3dB and actual downstream SNRM is about that too. Upstream target SNRM is 6dB and actual value matches that.
A few other points. The data doesn't point this out but it is clear that with this modem the sync rates (even the downstream) doesn't just go up and down for the hell of it, because of DLM mucking about or something. If you just reboot the modem and don't change any parameters at all then it tends to come up with the same sync rate it had before, so it's very consistent and this has not always been the case.
In the past there has been a lot of statistical ‘noise’ in these kind of sync rate stats, for reasons unknown. It could have been somehow linked to the downstream sync rate drooping, starting well below the downstream target SNRM and then sometimes going down further then yo-yoing. That doesn't happen any more either. Even on a 3dB downstream target SNRM the number seem a lot more consistent and healthy. The instantaneous downstream sync rates quoted for the DLink DSL-320B-Z1 modems in the past were at anything from ~0.6dB - 2.5 dB and rarely as high as an actual 3dB.
Coming back to the present, there does not seem to be much variation between sync rate achieved after a resynch performed at night time vs in the day time.
The bits per bin spectrum of the ZyXEL has some of the holes missing compared with the measurements taken with DLinks on two lines (although the line used in the ZyXel measurement is neither of those lines used with the DLinks, so not quite a completely fair comparison). It has a much longer tail at the top end, with a lot of values being 1 bit higher and there are a lot of 1-bit bins at the very top whereas there are no 1-bit bins with the DLinks almost as if they were not doing them 1-bit to a lack of such capabilities, or else a policy, almost like it thinks it is G.992.1. But who knows it could just be something to do with noise levels on those bins and bitswap decisions especially if the DLinks do not have monitored tones? Anyway, some of this might be related to why the thing seems to differ less between daytime and nighttime. I seem to remember reading something or other that mentioned an (analog?) filter, at the front end, maybe a higher-quality more expensive AFE on the B10A compared to some of the other competitor devices. Maybe better noise filtering means more resilience to RF ingress and that could help at night?