F1 Gearbox information via patent searches

I spent a day looking at Ferrari and Magneti Marelli patents to try to learn about F1 and his buddy E-Diff. Could have been a dead end if they used a philosophy of holding trade secrets vs patenting. Turns out there are meaningful patents that likely describe our systems. (To find a patent go to USPTO.GOV and search on the numbers I have listed here. Or you can use the Boolean search and put in “Ferrari” “Assignee name” AND “Modena” “Assignee City”. For Magneti Marelli use the Boolean search for the two names, both times as “Assignee name”.) Here we go. In Magneti Marelli patent 6,234,933 they describe what happens with the 2 gear engagement solenoids (Gear1 and Gear2) as a gearshift takes place. It’s a tough read but here is what I learned. This patent is all about going straight up and down in the H pattern. The only time they refer to going across in the H pattern is when they refer to SEL, (selection) which is where the other axis moves horizontally in the H. For this up and down movement two solenoids act on one shaft in the actuator. Assume we are driving in first and going to change to second gear. As we click the upshift paddle both Solenoids get some small sub critical current to get them primed for their next task. The system has gone into a pre disengagement state. See Fig 6 Ipredis and note that Xpredis doesn’t change so the actuator shaft at the gearbox doesn’t move yet. (X with subscripts here are the various shaft positions measured at the actuator by one of those sensors hanging off the back of the actuator.) Next both solenoids get full current momentarily to put full pressure on both sides of the piston on the shaft. Then as soon as the shaft moves towards neutral and reaches X0 those currents taper linearly down till Xdis is measured by the sensor. Xdis means the shaft and thereby gearbox is close to its Neutral Detent. The shaft ends up centered with the mechanical detent engaged. Just like it would if you tugged the car out of first gear and took your hand off the shifter as it got near to neutral in a hand shifted car. Both solenoids then drop to a small holding current (and therefore small actuator pressure) while Selection takes place (Selection is the movement by the other 3 solenoids and their related sensor to get into the desired position in the H horizontally). When Selection is completed (I.E the other shaft is measured to be in the right position) one solenoid closes (it’s current goes to zero) and the other solenoid starts to ramp up at only a modest rate to move the shaft towards in our case second gear to start the synchromesh action. This gets the shafts spinning at the same speed so that the selector dogs can go in without gnashing. Remember all this time, the clutch is open. In our hand shifted car this would be the movement from the neutral position towards the bottom of the second gear gate. If you do this part of the shift too fast you feel some resistance as the synchro takes a bit of time to spin up the slower shaft and then it goes in and the detent engages and holds the car in that gear. In the automated shift the computer knows that the synchro phase has ended just like our hand does when the gear goes in. It is measuring X all the time and notices the end of synchronization and the satisfying end phase of the shift as we get near to the detent. Shift is done but clutch is still open, so now the computer drops the active solenoids current to I1. See the diagram. This holds while the clutch starts to close. About halfway through the clutch closing, the solenoid turns off and its job is also done.

Image Unavailable, Please Login


If I confused you, I will gladly answer questions and even correct mistakes in my interpretation.

If you know for a fact that the cars don’t use this sequence, let us all know. Patents don’t always represent the final implementation of an idea.

Steve

 

US 8,041,486, is an interesting read. Exciting the torsional resonance of the geartrain via an abrupt opening of the clutch makes the relative angular velocities of the input and output shafts a combination of a linear deceleration term plus a sinusoidal term. So with the right timing the system sees a brief moment to grab the next gear ahead of schedule, hence speeding up the shifts. Clever. Not sure if our single clutch F1 cars do this, but I like the non-linear thinking

 

5,609,545 does a nice job of describing the synchro phase of shifting a manual synchro box robotically. Having raced Formula Fords with Hewland gearboxes for more than 25 years I wonder how much better an F430 fitted with a dogbox would be, all the while shifting without the clutch.

 

E-Diff enlightenment might come from US patent 6,918,478 as well as US patent 7,641,585.

The E-Diff and Clutch Solenoid Valves both work on the same principal (sort of) . The valve is used to progressively expose the actuation chamber to either pressure from the accumulator or conversely to the little fill tank which is at atmosphere or neither. (The actuation chamber is the cavity behind the piston supplying force to open the clutch or lock the diff). The main difference I see is that the Clutch servo has clutch release bearing position as its feedback, while the E-Diff servo control uses pressure in the cavity behind the E-Diff thrust plate as feedback.