Honda F1 project leader Yusuke Hasegawa has outlined a number of reasons why Honda has been struggling so badly in the beginning of the 2017 Formula One season. He confirmed that lots of problems were not discovered while running on the dynamo meter.
Have you considered that even though it is not traction limited under full PU output, the car may become momentarily traction limited during the shift? As shown in my previous post, there could be as much as 1000 kW momentarily available while the rotating assembly is slowing.
Yeah, I considered that, but I was thinking this absorption of energy would need to be built into the drivetrain. But not necessarily 100% of the energy.
I would think that traction limit lies in the area of 140kph which would suggest 2-3 is traction limited and possibly 3-4, based on the gear ratios from @Hollus for 2017.
Presumably if the PU output isn’t trimmed in some way the combination of inertia and PU is more than double PU alone and would lead to braking traction at higher speeds if not managed in some way.
Fortune favours the prepared; she has no favourites and takes no sides.
Truth is confirmed by inspection and delay; falsehood by haste and uncertainty : Tacitus
Have you considered that even though it is not traction limited under full PU output, the car may become momentarily traction limited during the shift? As shown in my previous post, there could be as much as 1000 kW momentarily available while the rotating assembly is slowing.
Yeah, I considered that, but I was thinking this absorption of energy would need to be built into the drivetrain. But not necessarily 100% of the energy.
If the energy is absorbed as strain energy in the drive train, at some point it will be released. Also something has to anchor the road end of the drivetrain to build up the strain energy. This would be a combination of wheel inertia and tyre traction. Would they be sufficient? Or do you mean by absorption slip?
Fortune favours the prepared; she has no favourites and takes no sides.
Truth is confirmed by inspection and delay; falsehood by haste and uncertainty : Tacitus
Have you considered that even though it is not traction limited under full PU output, the car may become momentarily traction limited during the shift? As shown in my previous post, there could be as much as 1000 kW momentarily available while the rotating assembly is slowing.
Yeah, I considered that, but I was thinking this absorption of energy would need to be built into the drivetrain. But not necessarily 100% of the energy.
If the energy is absorbed as strain energy in the drive train, at some point it will be released. Also something has to anchor the road end of the drivetrain to build up the strain energy. This would be a combination of wheel inertia and tyre traction. Would they be sufficient? Or do you mean by absorption slip?
In the power limited region I would guesstimate that there would be minimal slip at the wheels. Could the drivetrain be made to be more elastic based on the needed energy absorption? I realize we are talking short distances for drivetrain components but quill shafts could be used to double the effective length(?)
Consider that the drivetrain is already a collection of several springs with only a couple of dampers: the friction of the clutch and the friction of the tires. The springs are primarily the three long driveshafts and the tires.
Also keep in mind gear ratios. For example a ~1 rev duration engine torque fluctuation is experienced through only 30 - 90 degrees of rotation by the half-shaft, wheel, and tyre. Assuming a gear ratio spread of 12:1 - 4:1.
On the engine side, the crankshaft loses something like 12% of its angular velocity with each upshift. This braking arrives mostly from the ratio change itself. The energy from that deceleration is necessarily absorbed mainly by the driveshafts and tires. Conjecture: when this shock absorbing system isn't damped well enough it can be heard in recordings as low frequency shudder. Conjecture: this undamped spring oscillation sometimes makes its way through the tire and a similar low frequency pattern can be seen as dotted tire marks on the track surface.
Regarding seamless shift duration: I see mention of times in the .005 - .010 s range, or sub 1 engine rev. I don't know what this is a reference to. The actual figure should be closer to zero. There is no break in torque delivery. I'm fact, it should increase during the upshift as the engine slows 12% within some milliseconds as the SST dogs step, or slam, into the next gear.
I think the milliseconds that people are refering to when they speak of a shift, or a gearshift, in this context, may actually be the *effect* of a gearshift. The bending of the springs, as it were. The gearchange proper should be instantaneous with these seamless mechanisms. The bending and deflection of all the parts connected to them is what lasts a few, perhaps many, milliseconds.
That exact duration is something I'm unaware of. I suspect there can be a lot of variance. It does not need to occur within one engine revolution, but it might. It might also be distributed across several revolutions, since gearing combined with drivetrain elasticity permits this.
Regarding clutch slip: is the hydraulic clutch actuator expected to operate at these timescales? If it is too slow to directly interact with the gearchange, it may be acting more slowly across the aftershocks and harmonics of the gearchange; effectively as the damper for the spring that is the driveshafts and tires i.e. acting as a friction damper that can transmit full engine torque while breaking free to truncate peak torque fluctuation resulting from rapid engine speed changes. Therein may be a point worth clarifying. Pressure upon the clutch plates need only be reduced to absorb torque spikes while still providing enough pressure to drive the gearbox slip-free. Perhaps that is the essence of the Honda paper.
Regarding the clutch slip issue. I do believe there is scope to develop the gearbox to handle small slippages during the race as has been pointed out the clutch handles multiple start scenarios as well as pit lane starts. The clutch is small but probably has more surface area to it then most cars (most of the ones I've seen appear to be triple plate designs). I also believe that a team is going to want the most power possible delivered from the PU to the wheels so slipping the clutch would be the last thing they would want, no matter how small the loss.
So I see two possible scenarios..
1) Traction limited - basically the 1-2 shift. I can see using a certain amount of slip during the shift so as not to ruin the drivers feel on the pedal during accelerations from standstill. He wouldn't want that sudden burst of energy when at the extreme limits of traction already. With maybe the exception of Monaco I don't think I've ever seen a driver downshift into 1st after the race start (pit stops excluded). So this would mostly be used only during starts from a full stop type scenario.
2) Power limited - basically the shifts from 2 to 8. The 2-3 shift is almost always done at WOT. I would think that in this region it would be advantageous to not slip the clutch and keep the rotational energy in the drivetrain to give that extra boost while WOT.
The formula one clutch is not made use off when up or down shifting because the use of present gearboxes permits/allows the change in gear to take place instantly by letting the next gear be engaged while still driving at the previous gear (seamless), all this results in eliminated interruption of torque delivery while shifting gear. If a clutch is not defective it does not slip while transmitting power. For a healthy clutch to slip or produce slip, its clamping force exerted on the friction and pressure plates will have to be relieved, That is under controlled conditions or not. Re the possibility of making use of clutch slip while car is on the move and shifting gear. Reference is made to:- 5.6.3 power unit control must not be influenced by clutch position, movement or operation. 9.2.5 the amount by which clutch is engaged must be controlled solely and directly by the driver, with the exception of:- A, B, C, D, E and F. Does anybody believe the driver is capable finely controlling clutch slip while shifting gear while car is on the move?.
Consider that the drivetrain is already a collection of several springs with only a couple of dampers: the friction of the clutch and the friction of the tires. The springs are primarily the three long driveshafts and the tires.
Also keep in mind gear ratios. For example a ~1 rev duration engine torque fluctuation is experienced through only 30 - 90 degrees of rotation by the half-shaft, wheel, and tyre. Assuming a gear ratio spread of 12:1 - 4:1.
On the engine side, the crankshaft loses something like 12% of its angular velocity with each upshift. This braking arrives mostly from the ratio change itself. The energy from that deceleration is necessarily absorbed mainly by the driveshafts and tires. Conjecture: when this shock absorbing system isn't damped well enough it can be heard in recordings as low frequency shudder. Conjecture: this undamped spring oscillation sometimes makes its way through the tire and a similar low frequency pattern can be seen as dotted tire marks on the track surface.
Regarding seamless shift duration: I see mention of times in the .005 - .010 s range, or sub 1 engine rev. I don't know what this is a reference to. The actual figure should be closer to zero. There is no break in torque delivery. I'm fact, it should increase during the upshift as the engine slows 12% within some milliseconds as the SST dogs step, or slam, into the next gear.
I think the milliseconds that people are refering to when they speak of a shift, or a gearshift, in this context, may actually be the *effect* of a gearshift. The bending of the springs, as it were. The gearchange proper should be instantaneous with these seamless mechanisms. The bending and deflection of all the parts connected to them is what lasts a few, perhaps many, milliseconds.
That exact duration is something I'm unaware of. I suspect there can be a lot of variance. It does not need to occur within one engine revolution, but it might. It might also be distributed across several revolutions, since gearing combined with drivetrain elasticity permits this.
Regarding clutch slip: is the hydraulic clutch actuator expected to operate at these timescales? If it is too slow to directly interact with the gearchange, it may be acting more slowly across the aftershocks and harmonics of the gearchange; effectively as the damper for the spring that is the driveshafts and tires i.e. acting as a friction damper that can transmit full engine torque while breaking free to truncate peak torque fluctuation resulting from rapid engine speed changes. Therein may be a point worth clarifying. Pressure upon the clutch plates need only be reduced to absorb torque spikes while still providing enough pressure to drive the gearbox slip-free. Perhaps that is the essence of the Honda paper.
In 2009 Honda went one better than the ‘ZERO-SHIFT’ system and many other types of a controlled one-way clutch by designing a shift system that eliminated the need to use dog-rings and dog-rings clutch collars and selector forks. Their gear selection system and style of controlled one-way clutch was all inside the hollow gears carrying shaft.
....
this results in eliminated interruption of torque delivery while shifting gear. If a clutch is not defective it does not slip while transmitting power.
...
The connection is seamless. However the torque transmitted varies, in particular the input torque to the transmission, including the clutch, has a dramatic, though short, increase. @gruntguru gave some ballpark figures on this.
If the change in torque was instantaneous the the torque would be infinite, and obviously something would have to give. Tyre slip is one, transmission wind up another and it is also possible to arrange the clutch pressure such that it doesn’t slip under normal torque but does if there is a significant overtorque. No driver intervention, no electronics, just straightforward mechanical engineering.
So no need for a defective clutch. Every friction clutch will slip if the torque is high enough.
Fortune favours the prepared; she has no favourites and takes no sides.
Truth is confirmed by inspection and delay; falsehood by haste and uncertainty : Tacitus
Henry. Please refer to 5.6.3 and 9.2.5 about the possibility of being allowed to arrange for the clutch to slip.
But for a moment let’s forget about 5.6.3 and 9.2.5. In my opinion if a formula one friction type clutch is designed to a physical size and with a clamping force that will let it slip at a certain torque, it will destroy itself in a very short time during the race. What we talking about is different from the clutch being designed physically size wise and with a clamping force to handle a one-off race start of short duration and a known torque output level.
Do we have info on today's diff's? Pure speculation, but I can imagine some energy can be stored in a smart crownwheel to diff mounting spring plate thingy. Perhaps even combined with the LSD internals. What if the LSD internals can bleed or store some energy
Henry. Please refer to 5.6.3 and 9.2.5 about the possibility of being allowed to arrange for the clutch to slip.
But for a moment let’s forget about 5.6.3 and 9.2.5. In my opinion if a formula one friction type clutch is designed to a physical size and with a clamping force that will let it slip at a certain torque, it will destroy itself in a very short time during the race. What we talking about is different from the clutch being designed physically size wise and with a clamping force to handle a one-off race start of short duration and a known torque output level.
Neither 5.6.3 or 9.2.5 are relevant to the point we’re discussing.
If the clutch were designed with a clamping force to only have the ability to transmit the torque of a race start it would inevitably slip during a gear change. The torque at the start is purely that of the PU (ICE + MGU-K). During a seamless change the torque is that plus the torque generated by the inertia change of the PU as described by @gruntguru. That torque depends on the time it takes for the revs to drop which in turn depends on the mechanisms that consume the energy that change implies.
In your model the PU will continue to drive at the torque demanded by the driver. In my model the PU can modify the torque to help manage the consumption of the inertia energy.
In your model the clutch is rigid. In my model the clutch might slip to absorb some energy but also increase the time taken over which the inertial energy can be consumed, effectively reducing the power.
In your model the inertia spike arrives at the gearbox/wheels unmodified. How it is handled depends entirely on the siffnesses, inertias and frictions of those components.
I’m not saying that I know which of the elements in my model are used and to what extent. I am saying that I believe them to be part of a toolkit that the engineers have to modify change quality at the circuit. We know they do that.
As for a clutch not surviving the 8 millisec events suggested by @Tommy Cookers. That’s about 60° movement, that’s about 400 revs in a race of 90 minutes. That’s equivalent to the first couple of tenths of a second of a race start. I don’t know how long they slip the clutch at a race start but I doubt the 400 revs is more than one additional start.
Fortune favours the prepared; she has no favourites and takes no sides.
Truth is confirmed by inspection and delay; falsehood by haste and uncertainty : Tacitus
@ROON. Just in case you didn’t know or didn’t notice. Re the splendid “zero shift” picture marked (2) indicating ring-1, dog-clutch hub splined to shaft, and ring-2. Notice the preannounced “rounding” on one side of each engaging lug of the rings. This rounding of the lugs on one side is what makes the “zero shift” of one of the two gears act as a one-way-clutch and prevents the engagement of two gears at the same time locking-up the gearbox. The difference in speed of the gear engaged but not being yet used will slip when pushing against those rounded sides of the lugs.
The formula one clutch is not made use off when up or down shifting because the use of present gearboxes permits/allows the change in gear to take place instantly by letting the next gear be engaged while still driving at the previous gear (seamless), all this results in eliminated interruption of torque delivery while shifting gear. If a clutch is not defective it does not slip while transmitting power. For a healthy clutch to slip or produce slip, its clamping force exerted on the friction and pressure plates will have to be relieved, That is under controlled conditions or not. Re the possibility of making use of clutch slip while car is on the move and shifting gear. Reference is made to:- 5.6.3 power unit control must not be influenced by clutch position, movement or operation. 9.2.5 the amount by which clutch is engaged must be controlled solely and directly by the driver, with the exception of:- A, B, C, D, E and F. Does anybody believe the driver is capable finely controlling clutch slip while shifting gear while car is on the move?.
I think you are inferring too much from my post. I did not mention any change to PU torque delivery, as measured by the output shaft of the PU, although that is allowed by the rules. I also did not imply that the driver was controlling the clutch during gearshift.
Further, 9.2.5.b) is gearshifts, so by the rules you referenced the driver does not need to be in control of the clutch during gearshifts as it is an exception.
9.2.5 The amount by which the clutch is engaged must be controlled solely and directly by the driver with the exception of :
a) Stall prevention. b) Gearshifts.
c) Bite point finder where brake pressure, wheel speed and driver clutch demand safeguards are used.
d) De-clutch protections.
e) Drivetrain protection on the track outside of any start lockout period or immediately following stall prevention activation only.
f) Test signals enabled only when the car is connected to the garage system.
The relationship between the clutch operating device in the cockpit and the amount of clutch engagement may be non-linear but must remain fixed.
Additionally, PU torque and clutch control are specifically released from driver control during gear changes per 9.8.
9.8 Gear changing :
9.8.1 Automatic gear changes are considered a driver aid and are therefore not permitted.
For the purposes of gear changing, the clutch and power unit torque need not be under the
control of the driver.
The use of a sequential type seamless shift gearbox will result in no loss of power or interruption of torque delivery when shifting or requirement of the use of clutch after first gear.
Wow, amazing stuff, SS. Very interesting. Also, the sky is blue. Now define 'use of clutch.'
Slip may be desireable for friction damping. The intervals are small enough to not nuke the clutch plates. Slip could be passive via spring force or active via the actuator *partially* reducing clamping pressure.
