JB, Robin,
good news, it appears. I am not familiar with "zoneshare" but had a look - are you wired up like their graphic ?
(
http://www.cvpusa.com/doc_center/web_r6_zoneshare_manual.pdf)
The suspicion is that you have a "red to blue" situation at one or other zone boundaries.
It "should" be simple, and I don't imagine that you can reverse the connections except between the outputs and the track (unless you have another booster that is connected to a section of the track in parallel with, but not through the zoneshare - see below). It would appear that the zoneshare is all PCB based, so it's not going to have a reversed connection internally.
I presume you have it so that the whole layout does not shut down if one loco on one district has an issue. At least it suggests that all your track supplies radiate from a single point, and that makes a test quite easy. Looks like a useful unit, but not sure if extra boosters would not be a cheaper and more robust option.
Anyway, I would suggest temporarily connecting all the red track feeds together, ditto the blue feeds, and connect both direct to the main booster, without a zoneshare, or anything else, and see if your boundary problem still exists. If it does, you have a reversed track feed somewhere, but you would be able to find that with a meter, if all the track supplies are connected to the same terminals on the same booster.
You can then check the polarity across the boundaries using a multimeter set on 20V Ac or similar - you should have zero volts between the adjacent rails (A-B & C-D) and track voltage between opposite rails (A-C, A-D, B-C, B-D), of course.
If your multimeter is playing the goat, turn off the power but leave it connected as suggested above, and use the multimeter on resistance. Connecting the probes together should show zero, and disconnecting them should show "high" or "1" or infinity or something similar, so you can verify that it's working. You could also use a battery and bulb, and a pair of probes.
The resistance between A-B and C-D should be very low, and between opposite rails (A-C, A-D, B-C, B-D) an obviously higher value, though I do not know how high - "very" I expect.
Assuming that works as expected, and you can drive a loco over the boundary with no sparks and no misbehavior, then you can reconnect your zoneshare and see whether the problem recurs. If so, you know where to look.
You should not get a difference in speed when driving across a boundary where both sections are supplied by the zoneshare, unless it is damaged. You may get a difference when driving onto a zoneshare section from a section supplied by another booster. If this is the case, I suggest that you disconnect the zoneshare (connect all supplied sections together as above, to a single booster) whilst you sort out the boundary issue between boosters. You can then adjust all the boosters to give the same voltage, (and polarity) and the speed change issue should go away. It will return when you reconnect the zoneshare, because there are semiconductors in the track current pathway. You will need to turn the output voltage of the booster supplying the zoneshare up to compensate for this.
Hope this helps finally sort out the issue
best
Simon