Automatic Transmissions

[enginesrus]

Are normal automatic transmissions going to be made with more speeds than what has already been offered?
I think 10 speed Ford and GM is the most at the time? I'm thinking if they used a splitter then it would be 20 speeds.

 

[BrianPetersen]

I have a couple of customers that build parts for the 10R80.

Honda also has a new 10-speed automatic, and that one is for transverse front-drive applications. I haven't seen the inside of it and don't know anything about that one.

Ten is the most at the moment for a stepped-ratio transmission. It is certainly into the diminishing-returns regime as far as efficiency is concerned. I wouldn't exclude the possibility of someone ever building a transmission with more, but it's not really a design priority right now. As far as I know, ZF/Chrysler are not re-engineering their existing 8-speed to have more.

The ZF unit has 4 planetary gear-sets and 5 clutches. The 10R also has 4 planetary gear-sets but 6 clutches.

What allows these transmissions to be more efficient, aside from having a tall overdrive, is that the gear steps are close enough that most of the shifts can be done with the torque converter locked. In the case of both the 10R80 and the 8HP, the 1-2 and 2-3 steps are big enough to warrant leaving it unlocked to avoid a big jolt or bog through the gear change, but once it's partway through 3rd, the torque converter can be locked and stay locked. That objective has already been achieved, which means now there's not so much incentive to add even more ratios.

 

[NormPeterson]

Once you get the 10R80 up into 3rd gear, most of the rest of the ratios are spaced less than 20% apart. Cutting that spacing in half kind of begs the question of why not go straight to CVT.


Norm

 

[BrianPetersen]

A CVT to reliably handle the torque of the engines that those transmissions are connected to, has not been developed. And, gear-to-gear transmissions when properly designed have lower frictional losses than CVT mechanisms that have been developed thus far.

Also, these transmissions have a greater ratio spread than a CVT has without auxiliary gear sets. There are a couple of CVTs out there which have the CVT mechanism in series with a two-speed planetary gearset to get more range between low and top, and Toyota has a new CVT transmission that includes a "launch gear" which is a lower gear reduction than the CVT provides on its own, then it does a "gear shift" to CVT operation after starting off, essentially treating the launch gear as 1st and CVT operation as "2nd". And they all still use a torque converter.

 

[enginesrus]

In my opinion a cvt is like a Wankel engine its cool for awhile, it can not take torque for heavy hauling or towing. It is a design that is against all logic in the automatic transmission world as the way it works is to be in a constant slippage mode the belt has to be moving between and all the time being squished between the pulleys. That would be the main reason for the higher frictional losses of a cvt over a conventional clutched geared automatic transmission.
Again though adding a splitter would increase the speeds to 20 if closer to cvt operation is required, and still be able to obtain good load handling capabilities of a conventional automatic transmission.

 

[NormPeterson]

I wasn't at all trying to promote CVTs, don't much care for them personally. I only mentioned them because once the gear steps are only spaced 10% or so apart (and the transmission actually uses them all), you're almost there except for an awful lot of discrete shifting in daily use. Like (on average) every 3 mph on an easy run up to 60 mph/100 kph and a downshift or two every time you dropped a couple of mph and added a little throttle to recover the lost speed sounds like a recipe for generating consumer complaints of 'hunting'. If the shift logic was set up to intentionally skip every other gear under most circumstances to avoid complaints of 'hunting', it might as well remain a 10-speed.

I suspect that CVTs in relatively powerful CVT applications get a fair amount of torque management in the small drive pulley ratios, and it has to take hydraulic pressure to force the pulley 'halves' together in order to develop the requisite friction for transmitting X amount of torque. More torque, more psi, more parasitic loss.


Norm

 

[gruntguru]

So we either tolerate the slippage in a CVT or the slippage in a torque converter. I guess the torque converter is only functional for a percentage of the time and "bypassed" with a locking clutch for the remainder.

je suis charlie

 

[enginesrus]

The difference of a torque converter or even the clutches slipping is a lot different than a pinched metal belt being forced to turn in that pinched area and the constant movement of pulleys to accomplish ratio change. The slippage of the torque converter is fluid no friction interaction with metals or other materials thus no wear, and with clutching in TC's now a lot of loss is eliminated, even the CVT's have a TC. With pulleys pressing on the belt its a constant ripping for lack of better term, of the belt out of the pulley, so that as well as the diameter change imparts a lot of slippage and potential wear. I am surprised of the longevity of some of them though. In the old days they didn't last long.

 

[GregLocock]

It, it's called a lockup clutch. It has a fairly small torque and heat capacity, enough to transmit full engine torque, but you don't want to use it to rpm match, ie it needs to be either in or out . Falcon first had them on the 4 speed auto, say 1991, and I doubt they were the first.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[BrianPetersen]

CVTs do not operate by "slipping". There is negligible slip between the "belt" and the pulleys at the contact points.

A lot of torque converter lockup clutches nowadays are designed to have controlled slippage - they're not just on or off. A lot of them have a PWM-operated control valve and both the input and output speeds are monitored (the input speed is from the engine, there's another speed sensor on the transmission input shaft).

Torque converter lockup clutches in the early days got a bad reputation for causing driveability hiccups. Modulating the pressure to the lockup clutch allows it to be locked or unlocked progressively to avoid sudden jolts.

I know the Chrysler 62TE is like this, and it's not the only one.

 

[GregLocock]

Oh , that's right. Heat buildup is still an issue, so you can't go too crazy with the slipping.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[JayMaechtlen]

Early lockup converters?
My 1951 Packard with Ultramatic had a lockup clutch in the converter.

Jay Maechtlen

http://www.laserpubs.com/techcomm

 

[enginesrus]

I was going to mention the dual range hydramatic had a direct drive as well.
The act of the pulley pinching on the belt and the act of it being pulled out of the pulleys is a slip point. And then the decrease and increase of the diameter of the pulleys while still pinching the belt equals more slip. The metal belt can not just sit flimsy in those pulleys and expect to transfer torque it has to be squeezed very snuggly by both drive and driven pulley to accomplish it. And of course the squeezing also tightens the tension of the belt. There are no drive cogs to prevent slippage.

 

[BrianPetersen]

By "slippage" I mean the ratios of the input and output speeds not corresponding with the mechanical position of the tapered pulleys, i.e. the CVT tapered-pulley mechanism is set for (say) 2:1 reduction but the input is spinning 3 times the output with the balance taken up by the surface speed at the pulley having a substantial difference to the surface speed of the belt. *That* is "slippage", and that does not happen (aside from perhaps a minuscule and negligible percentage) by design in any normal CVT. It *will* happen if someone fills the transmission with the wrong fluid, such as normal ATF instead of the special CVT traction fluid that they're supposed to use. The whole CVT mechanism operates in an oil bath, and it's a special fluid that's meant for this application (and using the wrong fluid, such as normal ATF, WILL blow up a CVT).

The "slip" up or down the pulleys (not tangential slippage as described above) between the tapered pulleys only happens in conjunction with the whole assembly rotating while changing ratio. (Belt-type CVT mechanisms cannot change ratio at a standstill!) If you actually calculate the amount that each element of the chain ("belt") moves with each revolution while it's changing ratios ... it's insignificant and may actually be accommodated by the inherent deflection of the chain assembly (which is designed to flex slightly). Bear in mind that the clutches inside a normal automatic transmission "slip" with every gear change, so it's not like a normal transmission is completely free of wear points, either.

I'm not entirely sure what your point is. Are some production CVTs unreliable? Oh yes; there have been more than their share of bad ones. Are there some reliable ones? Absolutely; most of the newer ones seem to be pretty good for their intended applications. And you can say EXACTLY the same thing about conventional stepped-ratio automatic transmissions. Some are good, others are explode-o-matics. Honda built a whole design generation of bad 5-speed automatic transmissions attached to V6 engines around the turn of the millennium give or take a couple years. The Volkswagen 01M has a host of designed-in flaws. Chrysler had enormous difficulties with Ultradrive (a.k.a. the A604/41TE, their first electronically-controlled 4-speed auto-box). Yet the 41TE morphed into the 62TE, which after a few initial minor teething troubles, has mostly been OK. Biggest problem ... People filling the transmission with the wrong fluid! That transmission wants Mopar ATF+4, not Dexron.

Some CVTs are bad. Some are good. Some conventional-design automatics are bad. Some are good. Heck, some conventional-design MANUAL transmissions are bad. Some are good. What's your point?

 

[enginesrus]

You mention fluid. I can't believe even the know better specialist transmission shops have a propensity to use a one size fits all automatic transmission fluid on conventional automatic transmissions. But then there could be a not so nice reason for that as well. I agree that the use of the wrong fluid plays huge in premature transmission failures.
Those mopar transmissions you mention that spec ATF+4 don't happen to be the same transmission found in Toyota's / Lexus ?

 

[BrianPetersen]

No, not the same. Toyota is part owner of Aisin, so Toyota tends to use Aisin transmissions (I have no idea if they are an exclusive supplier to Toyota, though). Chrysler mostly builds their own transmissions, some of their own design (e.g. 41TE, 62TE), some derived from Mercedes (although I believe these have all been phased out), lately ZF designs that are built under license (the 8-speed longitudinal and 9-speed transverse transmissions). The Aisin longitudinal 8-speed transmission is entirely unlike the ZF 8-speed inside, and the GM 8-speed is a different design again.

It is hard to keep track of who owns whom and who buys from whom. Honda mostly makes their own transmissions, but they buy some transverse 9-speeds from ZF. ZF supplies transmissions to plenty of manufacturers. Ford and GM lately have been sharing designs. Aisin supplies not only Toyota but many others as well. It's confusing.

As far as your comment about transmission shops ... I'm inclined to have major servicing done by either the dealer or a shop who specializes in that brand, as opposed to being done by the corner lube shop (in my area, these shops employ low-paid teenagers who don't give a crap, and I don't want someone who doesn't give a crap to touch my vehicles). The ATF+4 is for the Chrysler-designed 41TE and 62TE and quite a few others. I am not sure, but I think the ZF-derived transmissions use a different fluid. I have a vehicle with a 62TE, so that's why I know what that one is supposed to have ...

 

[pontiacjack]

enginesrus- Dual Range Hydramatics almost had a direct drive in 3rd gear and 4th gear. But a small fraction of the torque still passed through the fluid coupling in those gears.