The AAIB report states (para 7.2.16);
"The ON/OFF switch was found at OFF, and an absence of substantial damage to the lift-toggle type switch, including the ramp mechanism, together with ground fire sooting patterns suggested this had been the setting at impact".
That is as near a "source" document as you will find; it is certainly authoritative in the sense the AAIB had vast experience of determining switch positions from sooting patterns.
The AAIB report covers this crucial device in a single paragraph, their only statement being about the switch. They simply point the reader to the Racal report which the AAIB have obviously been told is authoritative. But, clearly, the author felt the need to make the point that there was an unexplained conflict - he simply stated a fact and if someone else implied otherwise it was up to the BoI to de-conflict.
Nowhere in any of the papers I've seen is there an attempt to ask what the effect of switching OFF is - i.e. the question of whether the device continues to process in the background, but not display.
Yet, Trimble 8000 GPS is covered by a detailed and independent AAIB report.
MoD's case (waypoint change, timings etc) relies entirely on Racal's interpretation of their own device - a device which was not cleared for use and which the pilots were deeply suspicious of. So, MoD didn't have sufficient confidence to clear it, but after the event regarded it as infallible. And all the above was considered in isolation, lacking a systems overview.