But, after initiating the left turn, could the flightcrew have been cycling through the XYZ checklists to eliminate the causes and switched it off themselves, and been overcome/forgetting before they were able to switch it back on? Then the fire goes out
Sure, anything is possible. It also possible that the disconnection of the transponder and ACARs is a red herring. After all, it exists within the realm of imagination that both the transponder and ACARs could fail
independently of each other for mechanical reasons. The odds that two robust systems failing independently of each other within minutes is miniscule but it is non-zero.
The problem is that one begins to pile one miniscule probability on top of another miniscule probability in an additive fashion hoping to get to one (certainty). But in fact statistics multiplies probabilities. So the odds of one rare event followed by another rare event is more rare, not less rare. So the odds of the transponder failing followed by ACARs failing, then a fire breaks out, then the crew is overcome by smoke just as the fire burns itself out is to be so wildly improbable that it beggars belief. That doesn't mean it didn't happen. Even one in a trillion odds will come to pass sooner or later over an infinite number of trials. But speaking only for myself I would not be dedicating scarce resources on such scenarios.
The other poster is correct when they said I was being "too dogmatic" by saying a fire is "not possible" because anything at this stage is possible. But it's way down of my list of likelihoods.