Unexplained Revlimit

[leed]

The short list:
91 CRX Si
B16a ran with a stage 2bd (P28)
Turbo w/ DSM 450's
8psi

The car has been dyno tuned and has not had any problems until today.
I am new to Hondata, but feel very comfortable in tuning it in every aspect.
I have read all the manuals (twice over, including this forum) and am familiar with most all the common problems.
I currently have an O2 CEL. (have not wired in the OBD1 5wire yet)
No other CEL's
Static timing matches Hondata setting.
Rev limit set to 8400

The problem:
I installed a GM 3bar 2days ago; enabled it and had a friend with 4b burn new rom. Car starts, runs, drives perfect.
Today I started to up the boost in small amounts and datalogged test runs.
Ran into problems at 10-11psi (stumbling), we got it to run fine but had to back the timing under boost down farther than what seemed normal and decided to wait until I got back on the dyno.
We switched back to the last ROM used. When under full throttle/boost, the car would hit what seems like the boost cut even though both my gauge and Hondalogger showed only 8psi max. I have hit my boost cut B4, so Im pretty sure that was it. (engine does not come back on until the revs drop under 3500ish).
We temp. disabled the boost cut, and it still does the same thing. After reviewing the datalogs, it is happening at different boost levels (7-8ish) and at different RPM's (4500-8000)
When the cut hits, the datalog shows (regardless of throttle input): MAP goes straight to 0mb, Throttle goes straight to 0) MAP and Throttle reading recover when engine does.
We reinstalled the stock MAP, and loaded a known good ROM. (from the dyno tuning)
Still hits the cut randomly.
Interesting note: When the cut hits, the CEL (from O2) goes out, then comes back on when the engine passes under 3500ish.

A bit long winded, but looking for any helpful tips, or things to check.
TIA,
Dave