Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Why a torque converter?

  1. Dec 20, 2004 #1


    User Avatar
    Science Advisor
    Gold Member

    After being studied some about automatic gearboxes, I haven't really understood why is put a torque converter just at the gearbox entrance shaft.

    Surely as any gearbox, it needs some type of clutch. The torque converter can play the same role as a clutch when is functioning as fluid flywheel. But is the fact of reducing the r.p.m what I haven't understood yet. A common clutch doesn't function so (except some slipping between disks that reduces the r.p.m.)
  2. jcsd
  3. Dec 20, 2004 #2


    User Avatar
    Staff Emeritus
    Science Advisor
    Gold Member

    Modern torque converters can actually be mechanically locked by the transmission control system so they no longer function as fluid couplings when you're at cruising speed.

    - Warren
  4. Dec 20, 2004 #3


    User Avatar
    Science Advisor

    Well its a very elegant slipping clutch with a very gental transition between the two very different RPMs. The introduction happened about the same time power assisted brakes and power assisted steering came to the market. All combined together were marketed heavily to women as an empowering development to motivate them to drive and it worked wonders for sales.

    Now its just extremely useful sitting in traffic - so much for modern developments. :frown:

    Also if you think of the power to RPM relationship of a typical engine the ability to control the slip speed precisely with a device that doesn't wear out from large amounts of slip, it allows engines to be at their power peak over a larger range of transmission input speeds than a friction clutch provides. This is useful in more than just drag racing where its exploited. its used as well in earth moving equipment and other heavy industrial applications. Its far cheaper and simpler than the generator/electric motor and has nearly the same benefits.

  5. Dec 20, 2004 #4
    Does it sometimes produce more torque at the output shaft than is applied at the input (at the expense of slip obviously)? If so, then it is doing something that a normal friction clutch could never do.
  6. Dec 21, 2004 #5
    Comparing torque multiplying potentials of hydrodynamic and friction clutches

    According to howstuffworks.com, torque converters at low speeds typically double or triple the torque received from the motor.

    Friction clutches multiply torque every time they are employed, and especially when entering first gear. (One might not often see someone pop his clutch into first gear. When he does, generally either the engine dies or the tires break traction and squeal. People accelerating manual-transmission cars from a standing start engage their clutches smoothly in order to take advantage of the torque conversion capability inherent in friction clutches.) In my experience, they don't multiply torque as smoothly as do liquid couplings.

    Redlining the motor and accelerating the car by gradually backing off of the clutch pedal is a standard drag-racing trick. (If you have ever had teenage kids and manual-transmission cars at the same time, your kids might have done this to your car(s) many times.) This is why the 0-60 MPH acceleration times (for manual-transmission cars) listed in car magazines are always lower (better) than the 5-60 times.

    For the 0-60 tests the tester redlines the engine and slips (smokes) the clutch, massively multiplying the torque driving the transmission.

    For the 5-60 tests, the tester handicaps the car by trying to accelerate with the clutch fully engaged (left foot completely off the clutch pedal; car initially idling or almost idling in first gear while cruising at 5 MPH). The torque is not multiplied at all in this case and the relatively pathetic 5-60 acceleration times that are recorded are evidence of this:

    • 2005 Corvette Performance Review
      0-60 mph (sec) 4.3
      5-60 mph (sec) 5.2
    Last edited: Dec 21, 2004
  7. Dec 21, 2004 #6
    A friction clutch can't increase torque. If you use the technique of slowing the engine as you release the clutch, then you are still not transmitting more torque at the clutch output shaft than exists at the input - you are merely creating more torque at the input shaft by extracting energy from the flywheel (and other rotating components).

    I did consider putting the clause 'at constant or increasing engine rpm' in my post for just this reason, but you answered my question anyway. A torque converter can amplify torque even when both input and output speeds are constant. A friction clutch can't do this.
  8. Dec 21, 2004 #7
    The friction clutch - a type of continuously-variable transmission

    A friction clutch, by slipping, can reduced driveshaft speed relative to crankshaft speed. After accounting for friction losses in the converter, it is not possible at a given mechanical power output to reduce speed without simultaneously increasing torque. (The gears in a transmission can do it, and a clutch, by slipping, can also do it.)

    More generally, it is apparently not possible in any natural system (again, after accounting for friction losses in the converter) for speed reduction to not accompany simultaneous force increase.

    Every friction clutch is a type of torque converter and function as such.

    A friction clutch, by slipping, functions as a continuously variable transmission. The sole function of any transmission is to modulate ratios of speed and torque. You can prove to yourself that a clutch function as a continously variable transmission (and thus a type of torque-amplification device) by driving a manual-transmission car up a steep hill and either shifting to a sufficiently steep transmission speed or backing off of the accelerator. As the engine revs spontaneously start to drop toward stall, the engine either will or will not be savable from stalling by way of letting the clutch slip.

    If you manage, by smoking (slipping) the clutch, to save the engine from stalling, and you observe that you can proceed up the hill as long as you keep your clutch smoking, you will have just proven that your clutch can amplify torque.
    Last edited: Dec 21, 2004
  9. Dec 21, 2004 #8


    User Avatar
    Science Advisor

    hitssquad - the torque converter is specifically designed to multiply torque and the function of the size of the vanes and stator are such to affect the STR (stall torque ratio) but as the RPMs increase the effective lockup allows it to then behave as a friction clutch. The STR is maximum at the effective stall (maybe 2.2:1 or 2.8:1) where the output shaft is stationary and the input at maximum, but as soon as the output starts rotating the STR quickly drops and approaches 1:1.

    A poorly designed torque converter can offer very little torque multiplication and poor lockup. Its efficiency is junk and the losses to heat are massive. All it does is offer a lossy coupling with little other benefit. Even in that case, it offers the same advantage of a slipping friction clutch in terms of engine powerband management but suffers from some of the same excessive heat problems from all the losses.

    Using a friction clutch DOES NOT offer torque multiplication. It simply allows an engine to rev to a better location on its powerband if slipping and when dropped it converts the stored momentum in the flywheel.

    You don't redline the engine nor do you drop the clutch in a Z06 for best acceleration from a standstill. You'll get wheelhop so bad it'll clatter your teeth and shaking heads of surprised onlookers at the dragstrip. You rev the engine to the correct RPM (say 3k depending on how well the track is prepped) and let the clutch out in a rapid but smooth progression so as to not break traction. The slipping clutch is throwing away power but its not power you can utilize anyways so its still maximum acceleration.

    Think about it - where is the mechanical advantage with a friction clutch? But a hydrualic pump is pretty easy to reduce to a force/area physical explanation.

  10. Dec 21, 2004 #9
    A mechanical lever as merely a means to spread power over time

    It is apparent that the friction clutch is also specifically designed to multiply torque. A grabby clutch is no good for standing starts and one reason is its grabbiness is actually a function of acute-spikes-in-speed/acute-dips-in-torque-back-down-to-baseline. A grabby clutch is likely to kill an engine at low engine-speed because what is happening every time it grabs is the torque is falling back down to baseline instead of being amplified through clutch slippage.

    And if you drop the clutch at idle, the engine will die. This is because too much speed and not enough torque will be presented to the driveline; the engine will find itself up against a resistance that its meager force generated at idle cannot match, and it will stop turning over. Yet, even at idle, if the clutch is slipped just enough the engine will not only not die but the car will gradually accelerate. This is because the meager torque generated at idle by the engine will be amplified at the driveline by the slipping clutch.

    The mechanical advantage with a slipping friction clutch lies in the spreading over time of the power applied. This is all any lever does. A clutch allowing 2-to-1 slippage is functionally the exact same thing as a 2-to-1 gear system or a 2-to-1 lever. If you have one unit of torque being fed into that clutch, allowing for friction losses two units will come out. (And if you have two units of speed being fed into that clutch, allowing for friction losses one unit will come out.)

    Ditto with a friction clutch. Hydraulic pumps and friction clutches are both continuously-variable mechanical levers. They both spread power over time. Spreading power over time simultaneously reduces output speed and increases output force.
    Last edited: Dec 21, 2004
  11. Dec 21, 2004 #10


    User Avatar
    Science Advisor
    Gold Member

    Thanks all for trying to clear up the subject.

    But is clear that all these quoted advantages can be offered by the version of "fluid clutch" or "fluid flywheel". I mean the same torque converter but without the itermedium stator which provides the torque multiplication, or the same torque converter functioning at low rpms without torque multiplication as a hydrodinamic clutch.

    Either I am too stupid or I haven't understood why such intended torque amplification is necessary at gearbox entrance shaft.
  12. Dec 21, 2004 #11
    The practical need for slippable clutches and torque converters

    Internal combustion engines are not, by themselves, able to deliver torque to drivelines at zero driveline speed; and they are not able to deliver practical power at engine idle (hence they need to be allowed to rev up independently of the driveline speed). In order to (continuously) bridge the gap between zero driveline speed and the driveline speed of powerband-range-RPM engine-speed reduced by the first transmission speed, a slippable clutch or a torque converter is used. This allows torque to be provided by the engine to the driveline even at zero driveline speed and continuously on up to low-cruising-range first speed; and also allows the engine to be revved up to powerband-RPMs independently of the driveline speed.
  13. Dec 21, 2004 #12


    User Avatar
    Science Advisor


    If you feed in a certain amount of power, lets say x torque and y RPM then you get out a maximum of x torque and 1/2 y RPM in your example. That is 1/2 the power!! Therefore the rest is lost as heat!

    Unlike a fluid example where we can go 100:1 based on drive/driven areas exposed to fluid, a friction coupling cannot offer anything greater than 1:1 for power OR torque.

    Time is on both sides of the equation and can be cancelled out. There is no storage of momentum nor kinetic energy but there are losses that can be accounted for and thus is the result.

  14. Dec 21, 2004 #13


    User Avatar
    Science Advisor

    Think of a diesel earth mover. It uses 200 gal of fuel a day. Now instead we optimize its engine to run at one narrow RPM range.

    Naturally under load the machine will move slower. As it slows down, the torque converter increases the torque and therefore applies more force and thus counteracts the force applied to the wheel. Our engine needs to only maintain its one set RPM and produce consistent power over a range of loads. Maximizing that one aspect means a smaller engine with the same power (variable torque to fit loads by means of torque converter) could save 10% of the fuel. Think of the reduction in fixed operating cost per year and it might be equal to 2 lease payments on the machine! Pretty easy sell on the ROI!!

    In some tractors, the operating RPM and torque peak are misaligned for a similar reason, so that under load more torque is available as RPMs decrease to allow for a wider operating range.

    In cars, the first automatic transmissions were two speeds. It wasn't called "slip-n-slide with a powerglide" for nothing as there was a pretty big gap from first to drive. For the 60s-70s the 3 speed auto was a 2 speed with a second gear added between.

    The convenience factor cannot be overlooked, and most OEM torque converters don't attempt to multiply torque much at all past 1.5:1.

  15. Dec 23, 2004 #14


    User Avatar
    Science Advisor
    Gold Member

    Right. Thanks Cliff and hitsssquad. Now it seems I have understood it yet: the advantage of a extra torque available at rpms in which the torque exerted by the engine is a bit low.
  16. Dec 23, 2004 #15
    Actually, Cliff's and my assessment of clutches did not agree. But Cliff's correction of me was correct, I think. That is, a slippable fiction clutch (like a torque converter) can continuously reduce down to zero the speed delivered by a spinning engine crank to a driveline while still allowing that spinning engine crank to couple torque to that driveline, but (unlike a torque converter) that slippable friction clutch cannot amplify the torque of the engine.

    BTW, I arrived at my conclusion that Cliff was correct by conducting a thought experiment involving two weights hanging from a single pully by lengths of rope connected to each other by a slippable friction clutch. There seems to be no way that one weight's force could be amplified by the clutch without the other weight's force also being amplified, and it seems also nonsensical for the force of both weights to be amplified by the clutch. Therefore, the clutch can reduce speed while still allowing force to be coupled, but - lacking the not-directly-coupled levers of a torque converter - it cannot translate that speed reduction into amplified force.
    Last edited: Dec 23, 2004
  17. Jan 3, 2005 #16
    I'd like just to go over and review what everyone has said, and maybe give an example. A "Torque Converter" can be seen as simply a "lazy-man's clutch", though in some ways, it is more efficient than a clutch (Torque multiplication being an important example of that).

    Basically, what you must do when shifting a standard gearbox is the following:
    a) Ease off (when accelerating) or increase (when decelerating) the accelerator until the torque across the gears approaches zero. Most of us do this by instinct.
    b) When the torque is zero (the input and the engaging gear are at the same RPM), disengage the clutch (press the pedal). If you are very good at it, you don't really even need to disengage the clutch, but I'd strongly advise it for the sake of your gear teeth.
    c) When the clutch is disengaged, it is then easy to disengage and re-engage the gears, because there's (almost) no torque across them.
    d) When the gears are shifted, we must then decelerate (or accelerate) the engine, until the RPM out of the engine (nearly) matches that of the gearbox for the next higher (or lower) engaging gear. (Most of us do this without thinking about it.)
    e) When there is (nearly) a match in RPM, re-engage the clutch. If this is done smoothly, the input and output RPM of the clutch will match - - but this is rarely totally the case, so clutches have spring loading to help us.

    This is basically what we do to shift gears with a clutch, in a standard transmission. With an automatic, we use the torque converter so that we don't have to do the steps above; in particular, "step d" in which we must accelerate or decelerate the engine, so that its RPM matches that of the gearbox input for the next selected gear. The Torque Converter will literally 'pull' the two RPMs (mostly the engine) until they match. In a slightly worn automatic, you can feel the "thump" when this is happening. First, the slippage across the converter allows us to compensate the torque across the gears in order to allow us to disengage without "jiggling" the accelerator. Next, when the gears are shifted, the converter drags the engine RPM the last little bit it needs to match that of the next output gear. Finally, it provides for a smooth resumption of the acceleration/deceleration.
    Essentially, the converter is a "clutch" which also serves to fine-adjust the input (engine) RPM until it matches that of the output (the road, via the gearbox). This takes that task off the driver.

    To this point, we haven't mentioned, starting from a standstill. This is the case in which the clutch is particularly weak. There is no good way to start the engagement with the engine and the output at the same RPM. Each time the vehicle is started moving, there must be some shock to the spring loading and grinding wear to the plate. (A good driver can minimize this.) The Torque converter, on the other hand, by its design, can handle this imbalance in torques. This, for example, is the reason why automatics are heavily recommended for heavy towing.

    On the other hand, fuel economy will not generally be as high with the automatic and its torque converter (because of the slippage through the converter).

    This is extremely useful in starting from a stop, where the converter input and output RPMs differ widely, providing quicker acceleration.

    Useful in determining the design of a good torque converter.

    For the reasons I stated above, and - - -

    (A clutch can reduce the engine speed, but with gradual abrasion of the surface, which eventually results in clutch failure, not ideal, especially with heavy urban driving.)

    Finally, I'd like to say, and this is just my opinion; that Today's automatics are still crude compared to what they will be. To now, change has been slow and evolutionary, but I see major changes just ahead. This change will lead to elimination of both the clutch and the torque converter, replacing these with microcomputer-controlled servo-actuated transmissions. It is possible to perform all gear-change operations quickly, smoothly and efficiently in this way; and in hybrid autos, in particular, with their high start-torque motors, it will be equally as easy to move from a stopped condition. Elimination of the clutch will get rid of a reliability and longevity problem, and elimination of the torque converter will eliminate a cause of fuel-efficiency loss (as well as of reliability loss). It will also allow the use of smaller, simpler “crash-box” type gearboxes, another cost and weight saving. (Some people say that hybrids don’t need transmissions {re. Toyota Prius}, however I disagree. Whereas un-geared systems will work acceptably in hybrids, they cannot match the efficiency of those with transmissions.)
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook