Stripped F1 Gearbox

[hardingfv32]

zeroshift has a torque spike if I remember right. Hence, it wouldn't be logical to call it seamless shift.

Well the plot, as shown on their web page, looks very smooth with only small bumps. I would say it is as close as you can get to seamless with cog style gear set.

Brian

[Caito]

I totally agree. My point was that it's not as seamless as a double clutch or cvt can be.

Their is a transitory between gearshifts, but at least torque does not go to 0.

[jenkF1]

Ultimately CVT would be true seamless - with constant engine revs.

Unsure why the roadcar world haven't followed this, believe its in the Prius? Suppose a linear delivery of low engine power ain't worth the investment. Still surely a belt is cheaper than a gearbox

[Just_a_fan]

Ultimately CVT would be true seamless - with constant engine revs.

Unsure why the roadcar world haven't followed this, believe its in the Prius? Suppose a linear delivery of low engine power ain't worth the investment. Still surely a belt is cheaper than a gearbox

Because the normal mechanical gearbox is effectively "fit and forget". It's simple, robust, reasonably fool proof and the only moving parts are easy to make reliable. CVTs require control systems to effect the 'change'. There is probably more bits that are likely to break in a CVT. Manual boxes are, in normal use, bomb-proof.

[Caito]

CVT failed in car because people didn't like them. How they are, how they sound.

A CVT can in fact be quite simpler than an automatic gearbox with all its hydraulic system

[strad]

likely to break in a CVT

The chain is the weak point...

[Just_a_fan]

Or how about the IGS from Xtrac which also has two gears engaged at the same time (something that is impossible, apparently):

“The IGS is as fast as it can be, given that for a nanometrical split of time both gears are engaged,” said another judge, Alvaro Sauras technical editor CAR & Tecno magazine. “To tune the mechanics of the gearbox to make it able to work with two gears inserted at the same time, without blowing up, must be extremely difficult!”

http://www.xtrac.com/latestnews.htm#Adam

[lz2lps]

Racecar engineering March 2012: XTRAC IGS

[autogyro]

CVT failed in car because people didn't like them. How they are, how they sound.

A CVT can in fact be quite simpler than an automatic gearbox with all its hydraulic system

CVT transmissions whether Van Doorne cone,toloroidal disk or the three way torque balancing act in the Prius, all suffer from needing to use high levels of energy to operate the control systems. This makes them inefficient.
Most CVT systems can only achieve around a 1:1 to 1:4 ratio range and need extra gearing anyway.
I see the Xtrac boys are now in on the seamless act.
Their gearboxes still have power spikes like any other layshaft gearbox.
They might be very fast but they are not seamless.

[autogyro]

I suggest readers take a look at the Catterham thread in F1 cars.
There is a brilliant video on there posted by Catterhamrules.
Well worth a watch.
I am banned from posting on there but then I was only a consultant for the real Lotus F1 team not todays copy.

One thing is absolutely certain, the gearbox in this car is NOT seamless.
Try using your ears to confirm this fact!
I am confident that the speed of gearshifts shown from the F1 cars of today can also be equaled with a manualy operated dog box but then I have driven such boxes and not just played on computer games and talked about torque spikes!

[pitlaneimmigrant]

I am confident that the speed of gearshifts shown from the F1 cars of today can also be equaled with a manualy operated dog box

Laugh? I nearly cried.

Depending on how you define the gearshift (time to move the dog rings, or request-to-completion) an F1 gearshift take 50-150ms (on average), let´s see you match that.
And in an overlapped upshift the gearshift is effectively instant, the target gear takes drive off the current gear which then extracts before it gets collected by the overrun dogs. THAT is not humanly possible

The torque spike that would occur is managed either by an engine torque reduction or by slipping the clutch for a few milliseconds.

Also, (to other posters) don´t confuse F1 seamless shifting with dual clutch systems (aka DSC)
F1 boxes do alternate gears on different barrels and will momentarily have two gears in an engaged position (but only with drive on one, otherwise it gets messy), dual clutch systems also alternate gears, and have two gears in the engaged position but switch the torque transfer path by opening/closing clutches making the timing much easier.
As mentioned earlier in the thread this system was in use in F1 for a couple of seasons before the current generation of gearboxes came around

[ppj13]

Upshifts:

Zero shift engages two gears at a time. It is possible without gbox grenading if dogs cluthes are not infinitely rigid (nothing is). Both are engaged a short period of time. During this short time (less than 10ms) torque transmitted ramps from one to the other.

This mechanism equals instant(read, really fast) traditional dog clutch travel.

With former system (and almost with latter) seamless torque transmission is possible. The problem is that torque has a great spike, when engine speed is forced down in very short time. If all components are designed to stand that spike, (and they could do this) they will be many times (3 at least) too big for the rest of the time, a lot of wasted weight.

The optimum is to reduce this spike. During upshift, torque is sum of engine torque plus inertia torque. Easiest way to reduce the spike is reduce or eliminate engine torque (ignition cut) and "absorb" inertia torque with flexible components (like springs absorb bumps). That's why racing boxes are so flexible, (common trick is to stack two hollow shafts internally to "simulate" double lenght of the shaft) and that's why revs go (something easy to hear) up and down inmediatelly after upshift or sudden throttle input step.

In traditional boxes, ignition is cut inmediately BEFORE shift so dog is overrun and free to move, so driver or actuator doesn't need huge force to disengage. Then driver or ecu moves the dog so it select next gear and shift is made and then ignition comes back gradually. Shifter serves two purposes, both equally important: reduce actuator effort and torque spike. With zero shift, by design, the former is not needed, the latter is still needed.

So, to sum up. Seamless shift is possible, but not fast.

Edit: Seamless as per "no acceleration hole". Acceleration step will be inevitable.

[autogyro]

I am confident that the speed of gearshifts shown from the F1 cars of today can also be equaled with a manualy operated dog box

Laugh? I nearly cried.

I suggest you learn how gearboxes work, it will give you little time for mirth.

Depending on how you define the gearshift (time to move the dog rings, or request-to-completion) an F1 gearshift take 50-150ms (on average), let´s see you match that.

No problem

And in an overlapped upshift the gearshift is effectively instant, the target gear takes drive off the current gear which then extracts before it gets collected by the overrun dogs. THAT is not humanly possible

If it is instant why do you quote the following?

The torque spike that would occur is managed either by an engine torque reduction or by slipping the clutch for a few milliseconds.

Also, (to other posters) don´t confuse F1 seamless shifting with dual clutch systems (aka DSC)
F1 boxes do alternate gears on different barrels and will momentarily have two gears in an engaged position (but only with drive on one, otherwise it gets messy), dual clutch systems also alternate gears, and have two gears in the engaged position but switch the torque transfer path by opening/closing clutches making the timing much easier.

It is you who is confusing other posters.

Neither types have two gears FULLY engaged, they both use shift components that use a range of 'partial' engagement. They have to, the shift components are not strong enough to transfer full prime mover torque when their geometry is changing through the shift overlap.

As mentioned earlier in the thread this system was in use in F1 for a couple of seasons before the current generation of gearboxes came around.

The first semi/ auto shift gearbox in F1 was in Mansels Ferrari.
I was a consultant on the project.

[Richard]

autogyro - how quickly can a human physically move a shift stick between gears on a manual box? Assuming the clutch isn't used.

[bhall]

I am confident that the speed of gearshifts shown from the F1 cars of today can also be equaled with a manualy operated dog box [...]

And I'm confident that one day I'll be King of Zamunda.