PhillipM

Because you should be blipping the throttle and only engaging gear when the revs. match, you'll be wearing the edge of the dog rings otherwise.

madcows

There should be false neutrals.

PhillipM

I'd hope not!
They're from when there's enough play in something for it not select either gear properly, which means something is fooked already..

bhop

Parts of my statement were sarcastic. Sounds like you are referring to me when you mention, "I don't think retrofitting the RX-8 with an automated gearbox is going to be an appropriate solution (unless you REALLY want the RX-8, and money is not much of a factor - as it will be pretty damn expensive). I recommended the Evo earlier, but if you want a RWD car, BMWs, and Porsches are being offered with good automated manuals too". Are you talking to the OP, I own a RX8. I'm just saying how well would a sequential system cooperate with a stock RX8. I have yet to see one, well at least with one like TeamRX8 has mentioned. Thats all....

madcows

I'm sorry.. For some reason, I though you were the OP.

A dogbox with a sequential shifter is usually reserved for race cars. You can have one on a "stock" rx-8, but the driveability due to the shifting requirements would make it somewhat impractical for regular street use, as all shifts need to be performed as quick and hard as possible. The point of such a system is to be able to shift as quickly as possible without the need for a clutch (except for launching into first). Just about all top levels of motorsports utilize such a trans/shifter setup.

Phillip, all dogboxes should have "false neutrals" between each gear changes. Otherwise, it would mean that at certain points between shifts, two gears would be activated, which obviously would be a very bad thing. It's ecause of this inherent design that allows you to be able to up and downshift (accurate rev matching necessary) without use of a clutch.

madcows

I should point out that the axial gaps between two engaged dogs is pretty large in order to provide as large a window for mismatched revs as possible.

Here are some reads:

http://en.wikipedia.org/wiki/Sequent...l_transmission

http://www.carbibles.com/transmission_bible.html (scroll down to the dog-box section)

PhillipM

Yes, but the definition of a false neutral is the transmission not going into a gear when you've had one selection event....

Even then, you should still be unloading the dogs by pulling the clutch on the way down, otherwise you're wearing the dog as it comes out of gear, on the way up the box you're giving it time with no torque through the box to change cleanly, hence the split second you roll off the throttle.
You should be doing that on the way down with the clutch, not putting torque through the box and selecting, it will do it, but every time the edges of the dogs meet (which they will), then you'll wear the corners off from the torque load.

madcows

The links weren't intended for you, Phillip. I'm also not trying to argue your credibility.

When I say "false neutral", I'm referring to the point in between gear selection where not gears are actually selected (outside of actual neutral). This is purposely exploited to shift without the clutch. And while for longevity and reliability, you should be using the clutch, it's very common for drivers to not use it. And it you're very proficient at rev-matching, then very little wear might be occurring at all. If you shift properly, the edges of the dogs might never even touch. Do the automated shifters press the clutch? I don't think so.

Realistically, you can up and down shift with a standard synchronized transmission. I do it every once in a while just for the hell of it with no grinding (maybe a little extra wear on the synchros)

PhillipM

Actually the automated shifts do actuate the clutch on the way down the box, and execute a rev-match too, you don't need to on the upshift as simply cutting the power for a split second releases the dogs and allows the next gear to be selected with no torque through the gearbox, blipping the throttle on the downshift without using the clutch doesn't allow that.

Even shifting with an automated air shifter as we use results in some dog wear, there will always be a point of edge to edge contact, it's minimising the time for that event and the load (torque) on the dog when it happens that keeps the wear managable.

madcows

In the case of my motorcycle, I can accomplish clutchless downshifts by briefly cutting back the throttle while simultaneously beginning the shift to help disengage the dogs, followed by a throttle blip while in false neutral and into gear. Of course, as you suggested, upshifts are as easy as rolling back the throttle while shifting.

Now looking on Geartronic's site (http://www.geartronics.co.uk/paddleshift.htm), their system does not actuate the clutch, and they specifically say that the clutch is used only for standing starts.

PhillipM

You press the clutch yourself on the way down the gearbox, we work with the Geartronics bloke he's about 5 minutes down the road, I made some of his parts.
There's no benefit to powershifting down a gearbox, it just upsets the chassis if you don't get it spot on, I know it does say you can do clutchless downshifts - you can if you have to, it does sometimes get you out of a tricky situation - but you won't believe how much work he's had to put into it to get it reliable and low in wear and tear, there's no way you could do it manually.

I've seen a cheap system on an Elite box munch the dog selectors within one single meeting because they were trying use clutchless downchanges and it simply couldn't keep things in synch fast enough.

madcows

I wouldn't argue that that safer on the gearbox components to use the clutch. The only point I was trying to make is that you *don't have to*.

PhillipM

Well, yes, but then, you don't *have* to use one to set off either, doesn't mean it's a good idea

madcows

Oh, come on! Now that would be just nasty!

PhillipM

I've done it before when the cable's snapped on a stage....

madcows

Cable? Clutch cable!? How old is the damn race car?!?! lol.

And you said "stage", so I presume you're rally racing?

PhillipM

Hah, you might laugh at a cable, but put it this way, cable snaps from fatigue/vibration, 3 minute job to change a cable.
Hose bursts from vibration/fatigue, fluid everywhere - 20 minute job, might need new clutch plates...



We only use hydrualic thrust races when we're getting above 2700lb on the pressure plates, no real choice then.

madcows

Good point.... So now I got to keep a lookout for brits with huge left legs? lol

NeilW

Hi Guys, Neil from Geartronics here. Just to claify a few points...

Firstly, it's better for dog boxes if the clutch isn't used - this applies to up and down shifts. That said, the driver must be skilled in timing the gear lever movement to a throttle lift on upshifts and a throttle blip on downshifts. If the driver can't do it properly then it's best to use the clutch. If the driver just bangs the lever without the appropriately timed lift or blip then you can end up with a box of swarf within just a few laps! Seen it done

Intelligent paddleshift systems time things to perfection and so gearbox wear is much less - up and down the box. It's not true to say that dog wear is kept relatively low - it's virtually eliminated. Again, there is a qualifier - smooth and non-damaging shifts are more related to engine response than any other factor. We need to get the engine to change speed as quickly as possible so that the revs are a better match for the next gear. Also, the system doing the engine cut and throttle blip needs to be intelligent and have closed-loop control. It's no good working on fixed timers because every shift is different. You need to monitor the gearbox and modify the behaviour of the shift system in real time. That's what makes the difference between proper paddleshift systems (of which there are probably only 6 or 7 in the world that work properly) and the crude electric solenoid systems which shall remain nameless

Clutchless downshifting, when done properly, does NOT upset the chassis. In fact, if you have a responsive low-inertia race engine, downshifts are usually faster AND smoother than upshifts. One of my big customers sometimes used to get feedback from their customers claiming that the paddleshift system was missing downshifts - in actual fact, the shifts were so smooth that the driver didn't realise that they had gone in. Examining the data logging verified this. These cars use proper race engines that are capable of very rapid RPM changes. Anything using production based engines (big flywheels, long stroke, or, sorry to say, rotary engines) will never be as good because of the large rotating inertia.

Neil.

PhillipM

Cheers for clarifying things Neil, but as you say, the proviso on it not upsetting the chassis on the way down, is it being done properly, and I still think you'd struggle to do that manually on a bike to perfection without upsetting the rear wheel, especially when braking, given how little load is on the tyre and how easily it could lock or spin up, you'd only need to be a fraction of a second out to get the tyre past its slip limits, and I've seen how much your systems have to keep an eye on to get it right when you've got far more grip and weight available to quell any slight discrepancies, never mind doing it accurately enough by hand + foot to keep any torque load off the dogs just as they're engaging...


Your thoughts on the one on the previous page might be quite illuminating Neil, solonoid one that attaches to a normal synchro H-pattern box....

NeilW

Philip, If you're referring to the KAPS sequential gear lever conversion then I don't know much about it. By all accounts (and bear in mind this is heresay) the one made by IKEYA is better. I've seen these in the flesh and it doesn't look a bad bit of kit - but I suspect that setup of the cables is critical. Regarding using them on a standard syncro box, I would just as the question "what's the point?" unless you're just interested in bragging rights down the pub i know a few guys that are using them on H-pattern dog boxes and apparently they work very well and aren't a million miles of a proper sequential box. Forget about trying to paddleshift them though!

madcows

Thanks for your input, Neil. I had a feeling that the proper systems were as you described. Apparently even more interesting. I would imagine that throttle blips would take different amount of time depending on the rev range, in which case the closed-loop system also utilizes a crank/cam speed sensor (wow, that math would have to be performed very quickly).

And again, Phillip, my point was simply that up/down shifts can be performed without the clutch. Whether I'm more proficient at it than a machine is a whole different story - I'm not the Terminator I also don't race my bike, nor ride it at the limits, and any down shift is performed well before a turn, so the chassis is stable by the time I actually do turn.

PhillipM

And a very accurate barrel sensor to detect if it's mis-shifted and recalculate and retry the shift without you noticing (so make that very, very quickly)

NeilW

You would be surprised how much time we have available to make the electronic decisions. An engine spinning at 10,000 RPM rotates once every 6ms (6 thousandths of a second). A typical gear shift will be completed in less than 10 revolutions of the engine, which allows us to execute 1.8 million instructions during that time!

Even the fastest race engines are going very very slowly in terms of computing power .

Neil.

PhillipM

So can you put Pacman on there to give me something to do too?

It's still a wee bit quicker than ye olde right foot...