From what all my driver friends have said, that must be an anomaly...
I emailed Hewland (the manufacturer of these gearboxes during the 1991 season) with that very question back in 2011. He said using the clutch on downshifts in 1991 in F1 was common.
This was the question asked:
We are told by a racing gearbox engineer that a gear shift in 1991 attempted with a difference in rpm between the input synchro components and the output components of 200-500 rpm or above would almost certainly cause damage to the gearbox and would definitely upset the cars stability. Would you agree with this delta figure between mating halves of a dog clutch of 200-500rpm in 1991? Also do you know if the accuracy of “blip” during clutchless downshifts to the desired engine rpm of the lower gear had to be accurate to within a certain rpm range?
This was the reply from Mr. Hewland (jr) himself:
"The rpm difference between gear and dog ring is just part of the equation. More importantly is the status of the torque (either engine or braking) on the gear as the engagement is attempted. If the rpm difference is 1000 rpm, the shift will be fine and undamaging if little torque / inertia is encountered (due to clutch being dipped or zero throttle being applied). Equally a 300rpm difference will cause damage if full throttle (or heavy braking with no clutch) is applied.
Also, if the shift is made very cleanly (quickly), the engagement will likely be very succesful and the dogs will spend little time at different speeds. There are masses of factors; so it is much more complex than simple rpm differences.
To understand the principle; imagine that you are trying to put a wooden stick into rotating fan blades! If the fan is rotating at 1000 rpm and you slowly push the stick in, you will just shave the end of the stick off! If you push the stick in quite fast, it will get a bit of damage, but then rotate with the fan. If you put the stick in incredibly fast, there may be no damage, as it does not touch any passing blades. THEN, imagine the difference of these three examples if you cut the power to the fan during entry. I hope you can see that the variations are great.
Different drivers reduce the torque in different ways. The most common upshift was a partial lift of the throttle. The most common downshift was either a touch of the clutch or a touch of the throttle pedal. Drivers would experiment and have their own style; but pretty much it all worked. Some would have to improve their styles in F1 to make the gearbox last the whole race.
I drove a few laps in a 1990 Tyrrell. The gearshift was very light and easy (given that I was used to other racing gearboxes)."
Edited by gold333, 07 April 2018 - 21:28.