I wouldn't have a problem with comms because I wouldn't put a transformer in the way

Again my posts were why the RJ45 LAN isolation transformer you linked to would be a bad idea for SpeakerLink - the whole thing, not just a cherry picked signal type out of the connector (without someone suitably qualified giving an OK)

RS232 relies on DC levels which would be taken out by the transformer.

RS232 relies on separate recieve and transmit datalines which would need separate transformers - unless I've missed the mention of a MAX250/MAX251 chipset in there (although that would need optoisolators in there as well)

An isolation transformer is not a pin-pin connector - the whole point is that they don't physically connect.

Is there an explanation for how the five different signals makes it through the (4 pair) isolator intact?

You've now mentioned comms - you still haven't addressed RS232.

This is no longer about wiring or 'adapters' - this is about modification of the signal - you are stripping DC levels and trying to run the equivalent of 5 pairs of signals through something designed for 4 - so it is significantly more complex than just wiring with wire. (Not sure why you'd mention measuring voltage levels - you'd need the transformer correctly loaded for that to be any use, and I've not yet seen a SpeakerLink simulator.)

Again (for the last time) this is referring to an isolation transformer that you linked to which takes 4 pairs on one side and giving 4 pairs on the other side being used on something that has effectively more than 4 signals.

Hopefully there's enough here to make someone think before risking an isolation transformer for SpeakerLink - I don't think I can contribute anymore without repeating the same statements over and over again, so I'm out of the 'but a LAN cable works so it must be the same' discussion. I've used this site extensively for research and I know how I'd feel if I'd wrecked my own kit following unconfirmed advice.

Matt