24 replies to this topic

[gloomyandy]

In this weeks autospaort there is an artical speculating that the gearbox used by McLaren gives seamless shifts abd may be the edge they have over Renault. Soooooo....
1. How does this gearbox work?
2. Does it really give them an edge?
3. When did they start using it?
4. Who else has anything along the same lines?

Thanks

Andy

[zac510]

Wouldn't that be illegal?

[Chubby_Deuce]

or so BAR was told..

[peroa]

I read that it should provide 0.36 per lap but that it is a bit "optimistic" I think.

Since when is seemless shifting illegal?

CVT is, like the Williams 92, 93?!


McLarens gearbox uses gears, just that they are shifted fast and without loosing speed.

[scarbs]

I read the piece and Mark Hughes was trying to prove if McLaren use this set up or not. I don't think he really reached a conclusion.

To my knowledge several teams including BAR, Renault and McLaren have adopted new gearboxes this year to provide faster shifts.

In conventional gearboxes the gear is engages using dogs (square teeth on the side face of the gear) , meshing with dogs on the selector ring. One only one dog is engaged at any one time and the gearshift involves the deselection of the current gear and the engagement of the next

While I don't know of the exact design of these systems, I know that a system that splits the dogs into two (up shift and downshift) that are sprung against the dog ring. Each dog has a slope on its outer face, this allows the next gears dog to jump over the gear dog when the current gear is being driven. This set up works like a ratchet, when the load is taken off the the gearbox momentarily the current gears dog deselects and the next gear fully engages. engine power is reapplied with little loss in drive.
This is not constant drive, but a large amount of the selector fork movement and dog engagement is removed from the shift, there is a patent out for this design by a company called Zero shift, as explained in much more accurate detail in Racecar Engineering magazine.

I was told that simulations predict a lap gain of as much as 0.3s for fully seamless shifts,


Scarbs...

[zac510]

Originally posted by Chubby_Deuce
or so BAR was told..

Yeah, I thought that would have set the precedent.

If it is worth (roughly) that much per lap then every team MUST have it by Round 1 next year. It is usually once this is realised that the FIA bans it!
Maybe not this time, as the ZeroShift is a setup that really could filter down into road cars.

[J. Edlund]

McLaren has also worked on a twin clutch gearbox which was said to give very smooth gearchanges. There was however some legality issues, and also some problems concerning reliability.

[Christiaan]

I remember some time ago when TC was illegal Ferrari said that they had seen a loophole which allowed then to engage a form of TC during the gear change, which was quoted to be a period on 200ms. Now I wonder that with shifting times of 200ms how much can you gain by effectively reducing that time to zero.

I hear the Audi TT has somewhat of a seamless gearshift. Dunno how it works though

[RDV]

Christiaan -Now I wonder that with shifting times of 200ms how much can you gain by effectively reducing that time to zero

...at Monaco assuming 28 shifts , at 2/10 per shift= 5.6 seconds.... mmm.. we can scrap this thought experiment... as for that gain a lot of F1 designers grandmothers would have already been sold... (note gain is only on upshift)

drive loss during shifts closer to 0.040 sec even in clunky boxes.. nearer to 0.025 on a good one... maybe Ferrari was quoting 0.020 secs or 20 milliseconds?

but at 0.040 still 1.12 secs at zero loss... not bad gain....

[Christiaan]

I think you are assuming that the car is stationary during the time of the shift - which it isnt. To accurately work it out you would need to take the %age of speed lost during the shift. For a total of 5.6sec spent shifting per lap, I still wonder how much time is actually lost.

[Greg Locock]

For a rough check on a mythical track:

In a 90 second lap 60% is at full throttle, then that is 54 seconds at full throttle. say you average 800 hp at full throttle

To be extremely crude, that gives us an average power input of 480 hp.

if you can give me 5 seconds of extra full power that is an average power of 524 hp, an increase of almost 10% .

I'll have it.

Obviously a more complex analysis will generate better results, but energy methods are fairly robust.

[-RM-]

Lapsim with the F1 example car at Donington gives:

40 ms 1:13.78 304.2 km/h
20 ms 1:13.65 305.1 km/h
0 ms 1:13.53 306.1 km/h

[RDV]

Christiaan-To accurately work it out you would need to take the %age of speed lost during the shift.

Guilty as charged m'lud... was doing a very fast thought experiment without full investigation of data... there is a lack of acceleration, not total lack of movement.

Would say -RM- s approach nearer to truth.

[desmo]

I was told some time back that the upshift times in F1 were such that there was honestly little point left in trying to improve them as the current interruption was barely sufficient to accomplish the change in the rpm necessary without actively braking the engine which on its face sounds to me like a dodgy idea.

I'd played with Lapsim and got similar numbers some time ago.

[DOHC]

Originally posted by Christiaan
I hear the Audi TT has somewhat of a seamless gearshift. Dunno how it works though

It works by having two clutches. In this pic 1st and 2nd gear are simultaneously engaged. Clutch 1 is engaged but clutch 2 is disengaged, so power is routed through 1st gear. "Shifting" consists of disengaging clutch 1 and engaging clutch 2 (in principle this requires very little time, say tens of milliseconds). After this upshift, while accelerating in 2nd, 3rd gear is engaged and the process is repeated. The trick is that all the odd gears (1, 3 and 5) are on clutch 1, while the even gears (2, 4 and 6) are operated through clutch 2. This means that "shifting" is no longer a matter of changing gears, but only of "changing clutches."

[rhm]

I don't have any information or speculation about McLaren's gearbox, but there's a short news article in this month's Racecar Engineering reporting that Geoff Willis admitted at the MIA "networking dinner" that BAR used a seamless shift gearbox throughout the 2005 season. There's not much information except he states it works on a similar principle to the Weismann Quickshift and Zeroshift systems. But since those systems are quite different, that only tells you that it doesn't use a twin-clutch system.

Twin-clutch gearboxes are thought to fall foul of the rules regarding brakes in two ways. One is that a car may only have one brake per wheel and two is that they have to be passive - taking only power from the driver's foot. This is because by engaging a gear on both shafts and then engaging both clutches, the gearbox can act as a brake. Of course it would be hugely impractical as a brake, but rules is rules.

There was a lot of speculation from various pundits as well that any seamless shifting gearbox would be regarded as a CVT and thus illegal, but perhaps that's not the case if BAR have actually been running them. AFAIR Mosley is quoted as saying that there must be a measurable break in power delivery during gearshifts. Personally I disagree with that - as long as there are a no more than 7 distinct ratios the gearbox should be legal.

[John M Cannon]

Pehaps this is what you're looking for.

http://www.zeroshift...ducts/index.htm

[rhm]

Hmmm, I don't think anyone is running the zero-shift system in F1. I don't think they will either based on Geoff Willis's comments that instantaneous gearshifts are not really a goal. He has basically admitted that BAR were running a derivative of the Weismann Quickshift system on the BAR007. It's previously been tested by Williams.

The Quickshift system works by coupling the ratios to the output shaft via a roller-clutch (Sprag clutch). There's a complex mechanism that runs inside the hollow output shaft to control the engagement of each roller-clutch (they'd work automatically on upshifts, but you'd never be able to shift down without it). Although the system allows for almost instantaneous shifts, Willis flags this as a potential problem rather than a benefit because it could put a lot of strain on the drivetrain as well as potentially upsetting the tyres. There needs to be some cushioning in the drivetrain (perhaps by releasing some pressure on the clutch during gearshifts).

The real advantage of Quickshift is that by doing away with the sliding dog mechanism of a conventional manual transmission, the gears can be placed almost adjacent to eachother making the gear cluster a good deal shorter. This was important in the past when Williams tested it due to the aero rules then. It's probably not as important now (especially with the shorter V8 engines).

http://www.weismann....uickshifts.html

[Tomecek]

Is this
http://www.autosport...rt.php/id/49822
Williams' gearbox same design as that of McLaren and BAR before? Is it same idea as CDG Audi's concept?

[rhm]

I believe it's the same concept as BAR's. I've probably mentioned in this thread or another one on this board, that Williams tested the Weismann Quickshift in the mid '90s. Honda are thought to have acquired the technology when McLaren tested it in the late '80s.

[gmw]

I saw the Williams transmission on the test rig in Pete Wiesmann's shop. It was being driven by a 2.2 L Chrysler engine with a 3:1 step up gear between the engine and the tranamission in order to get the correct input speed. They were cycling the transmission up and down through all of the gears. It was interesting, but understandably Pete wasn't giving any details of what was inside the box. It was tested in a car later I believe.

GMW

[anbeck]

So which teams will use it in Bahrain?

McLaren, Honda and Renault (if it's true they already used it in '05) and Williams. Anybody else?
Could Ferrari have missed an entire year of technological progress (well, not as bad as Smith's "I've developed the RB2 for the Antarctica GP only)?

A.

[bluetentacle]

Originally posted by anbeck
So which teams will use it in Bahrain?

McLaren, Honda and Renault (if it's true they already used it in '05) ...

Source? I've never heard of Renault using a continuous torque transmission. They did however have one less gear in their unit, being a six-speed transmission, as opposed to the more common sever gears.

This year Renault have reverted to a 7-speed unit, due to the peekier torque curve of the V8. On the surface at least, Renault have lagged behind others in terms of cutting edge transmission technology. For one thing, their unit is fully made of titanium, whereas Ferrari, McLaren, and Honda have been using carbon or carbon/titanium gearboxes.

[davegp3]

Comparing McLaren and Renault in 2005, I think Renault had edge in better traction out of slow corners, because they had more weight on rear wheels . On the other hand mclaren was meant to be better in changing gears thanks to their seamless gearbox so it evened out the situation partly.

[Nathan]

Could Ferrari have missed an entire year of technological progress

Who's to say Ferrari didn't use such technology years ago? They were alot quicker than many for a while and this could be an example of one reason.