As for your edit: very heavy satellites cost (including launch) ~$250M whereas a manned Mars mission is estimated to cost $6B.
I really don't see how you could do 'several manned missions to Mars or another planet of your choosing' for the same budget as a single relay satellite, even a large one.
The problems I see with such a design are simply that it does not give you any advantage for spending all that money and actually stands a fair chance of making things worse. More stuff, so more stuff that can go wrong and if it does it is in a place where you can't fix it. Limited life-span as well compared to the Voyager itself because of the larger complexity (must keep two antenna's aimed at the same time while moving itself).
Uplink path of DSN is capable of significantly higher power levels than 20kW, several orders of magnitude more. Also the whole transceiver electronics are especially fiddly, with various cryogenically cooled and/or high-power microwave valves without meaningful solid-state replacements, which is not something you want to have in GSO without any chance of mainteance.
Yes, that's exactly what I wrote upthread, that's the hard requirement. And that is what will fail and then you've got a very expensive doorstop in an orbit outside of any repair capability. Inability do do maintenance / repairs is the killer.
On the other hand it is certainly true that RCS of such satellite would require exceedingly large stores of RCS supplies which will invariably run out and have to be somehow replenished.
Edit: even hall thrusters require stores of xenon and with the precision required the amount consumed is far from practical.
Edit: TLDR: for the amount of money required to design, build and deploy such relay satellite in geosynchronous or such orbit you can do several manned missions to mars or some other planet of your choosing