AerocatS2A
The reference I was using was from my engineering training notes, which were a direct copy of the Bombardier training notes. Further to that, after blowing the dust off more notes, I have found a copy from the wiring manuals that show that in an autofeather situation, relays 6121-K15 or K16 (depending on which engine fails) energise which shut off the opposite engine bleed air. The only additional note I have, is that Mod 8/1747 disabled the latching function of these relays to prevent a sustained uptrim signal being caused by a momentary loss of torque signal. ie a transient dip. I seem to remember that there were occasions when the torque on an engine would 'hunt' slightly, and this could cause an uncommanded autofeather, even though the engine was nowhere near the trigger point. Over the years we suffered a lot of autofeather problems which tended to be caused by electrical connector problems because of the very short length of the pins. It became the practice to check tighten all engine connections with connector pliers, as 1 click on the connector could make all the difference. Improved engine harnesses were introduced, and we bought the Delastek test set to carry out fault finding especially on the TSCUs, EECs and the harnesses. Best bit of kit we ever bought. We also wrapped the engine connectors in self amalgamating tape to stop moisture ingress. We also discovered that the TSCUs needed to be matched part numbers, even though P and W said that certain unmatched numbers were compatible with each other.
Hope this helps
Apia Pipe