Quite a few Conway failures in the VC10K during the 5000 or so hours I had on it. One at MTOW / ISA+ quite a bit at Bahrain just after VR, well handled by the co-pilot. Perhaps the most entertaining was during my IRET; on a simulated 3-eng go-around one of the others let go quite spectacularly! But we were quite light and the '10 had bags of thrust, so after a bit of throttle readjustment a routine 3-eng circuit and landing was uneventful. Once did a 3-eng ferry back from Keflavik; during the climb one of the remaining engines surged rather impressively; fortunately it recovered OK; some component had been wrongly adjusted and we were well outside the temp/alt/rpm range which would normally cause a surge.
A few rough running engines on light aircraft too, but no serious events. Seeing oil spreading itself across the windscreen during on of my student's slow rolls was a bit interesting though. We left the power set as it was and climbed for altitude as she flew us back to Abingdon; after the resulting PFL pattern, we shut down and got out to find the aircraft covered in oil because one of the groundcrew hadn't secured the oil filler correctly. Unlike civilian practice, in the Bulldog the groundcrew checked the oil, secured the cowling flap and signed for it in the F700 - we didn't do any routine checking ourselves. It was the last in a series of poor servicing events by the contract ground crew person in question and he was invited to find another job.
At Brawdy I had a lot of thumping and banging in a Hunter just after take-off, so advised the leader that I might have a problem. Off with 2 x 230 tanks and a fixed power GCA followed. Nothing wrong with the engine although some settings were tweaked during the subsequent engine ground run. One of the staff told me that if the nosewheel rubbing strip didn't stop the wheel rotating after take-off as it should, it was likely that the tyre could rotate for quite a while and any manoeuvre could cause vibration - we found that the nosewheel was out of balance and the rubbing strip looked clean, so I guess that's what the problem had been...