What I can't find is some past event where a MAHF occurred, it's possible?
I doubt there has been one or will be one since asics and nodes are separate entities.
A MAHF/SF would likely only occur if a critical vulnerability was found that needed to be patched quickly imo as it has a much higher risk of a chain split than the current model of what seems to be in its simplest form:
1. Proposal generated
2. Code changes made to show what the proposal will do
3. Miners asked to signal their cooperation in the upgrade
4. UAF only goes ahead once a majority of the hashing power is dedicated to the change (only needs to be 50% but 90-95% is normally reached pretty fast iirc after the first few signal to accept it).
Because it's being discussed, segwit was a soft fork and was activated as one at the time because there was a consensus of miners.
The chain split to bcash was probably both a miner activated hard fork of bitcoin and a user activated one as bitcoin became incompatible with it immediately (at least because they had replay protection) but needed both miners (verifiers) and nodes (propagators) to make the new chain.