I too was a prop holder for the "Protesting" engines. With a wind up the kilt you would hold it during startup until the engine pulled it out of your hands. We used to horrify the SLF by giving the prop a mighty push as the engine started. Made it look like we were "hand propping" it to start.
The engine icing problems caused "bump stalls" where the ice buildup in the duct came loose and slid forward into the compressor, got ground finely, turned to steam and put out the fires. The glow plugs that were fitted re-ignited the engine giving a characteristic "bump".
Reports of "bump stalls" meant opening the massive cowling and inspecting the compressor face. It meant looking around a curve in the inlet as the Proteus had a reverse flow configuration so we used an intrascope. Sometimes the ice was too much for the very robust compressor and I have looked in and seen several rows of compressor blades cleaned right off. We called it "corncobbing". Bent blades were acceptable provided there were no rips or tears or missing parts and the blade was not bent into a reverse pitch. Not at all unusual to see bent blades.
Aeronaves de Mexico operated two model 302 Brits that I worked and saw a lot of "bump stall" activity over the Gulf of Mexico. Aircraft had only four tanks and for Mexico City - New York had to fly high and pretty straight-line to get the needed non-stop performance. This made it difficult to avoid the wrong kind of clouds that caused the icing. I seem to remember those were alto-cirrus and were composed of ice crystals.
Proteus could be hard starting too and after two start tries required draining the fuel can before attemping another go.