It just occurs to me that one quick way of confirming whether it's indeed the beltpack (hence not the unbalanced mic connection that I'm suspecting) would be if you happened to have another "trouble free" DPA headset mic and adapter, or (even better) a Line 6 lavalier mic to substitute for the combination that's giving you the problem.

Should the beltpack prove to be OK with another mic, you will be well on the way to localising the fault.  

As an aside, I'm always amazed by the harmonic variety and shattering loudness of the digital "storm" that I can induce from my TA4 to XLF mic cables when they're left un-terminated.

However correct location of the cable once the mic's connected invariably yields the pristine audio that I've come to expect from my XD-Vs.

As you've tried other combinations, it looks like the amazing Don Boomer's right "on the money" yet again, so your suggestion of a Sennheiser replacement would seem to be the only practical way to get on the air, unless these guys can either lend you a replacement or know of someone who hires them out:

http://www.studiospares.com/mics-wireless/line-6-xd-v75l-lavalier-wireless-system/invt/301160/?VBMST=line%206%20XD-V 

"Goes away when you completely envelope the transmitter in your hand ?"

That's still "bugging" me as I would not expect un-clocked or otherwise corrupt data to respond like that, whereas attenuation (by the salt-water filled human anatomy) of RF-induced "stray" digital noise certainly does.

I look forward to hearing the results of whatever further experiments you're able to try in the little time you have before "curtain up".

I reckon you needn't bother with a cap instead of a zener, ...as the actual method of blocking DC back to the input would seem to be irrelevant and everything else in your wiring would seem to be exactly as you and I would expect.

In the event that you're listening at a very high "critical evaluation" PFL level on headsets, then maybe you are right in thinking that your perception of the unbalanced mic-induced noise floor is (dare I say it) "too critical".

However, if you feel you really must have extra gain at that input, (and if you have the time to do so) I think you'll need to at least double the value of that "pull-up" resistor to about 20 or 30kΩ.

An earlier discussion related to such a "gain-maker" (albeit for a Shure headset) can be found here: http://line6.com/support/thread/70314?tstart=180

If it was me, I'd prefer the extra headroom and so would leave the values as they are, ...UNLESS there was a need for absolutely pristine individual multi-track recordings for later album-related re-mixing and/or overdubs.

Even then, it's my experience that the "Norm" setting of the superb adaptive processing that Line 6 calls their "Dynamic Filter" performs a "clean-up" that's more than adequate for most applications.

I just hope that you've enough rehearsal time remaining to try it, although enabling just the "Norm" setting ("live", at the receiver/s) would almost certainly pass un-noticed in "the house". 

I see a couple of points in your post that concern me.

You are using the mics even though the scan shows  full interference bars on the channels you are operating on?  That is risky, especially in RF2 mode.  I would suggest switching to RF1 mode.

Interference ... even if the performer's costume is causing interference, that interference cannot become audio in the radio.  We send a code in the header of every digital packet and when the receiver receives anything that does not have the code then it does not make sound.  And if that reception is "audio" that doesn't pass either as your are still in the digital world.  I can't say for certain, but the most likely cause is a mechanical one.  Did you also try swapping transmitters?