The RB-211-22B had a overly simple fuel nozzle that used the same nozzle ports for starting as for takeoff unlike other engines with duplex nozzles. At starting flow it basically dribbled. Coupled with the fluidic amplifier of the fuel control and a starting regulator it was very, very marginal. The complaints from the northern region of Eastern about hard starting in winter were ignored by the engineering folk sitting under a palm tree in Miami. Rolls told them that Air Canada was having little trouble but omitted mentioning they were using JP-4 fuel which is wide cut kerosene/gasoline.
The droning noise during startup was an infallible indicator that the start was going well. If you heard a "pssst" and the droning stopped, reach for the HP cock and abort the start for the rpm's going down and the temp is going up.
I was called to LGA one day for a #1 that would not start. It was borescoped, FFR and start regulator replaced, inlet guide vanes checked and still no start. The engine would do what I mentioned above. I noticed the HP pneumatic bleed light flickered at the "pssst". In desperation I had the LGA mechs mechanically lock it out. Engine started. Onto the gate, fuel it, load it and it started fine and off it went to FTL. Was to ferry to MIA then but the engine wouldn't start. They played with it, ferried it to MIA and finally removed the engine. It was put in a test cell. They could not start it. Final determination that the IP and HP compressor were incompatible and should never be put together again.
All is forgiven with the later -524 and most especially the -535E4 models. Sort of like comparing the Wright R3350BA (B-29/L-049) to the BD (L-749) model if you are old enough to remember them.