Why do overdrives work?

[stillsearching]

What I mean is... I don't see the point of gearing a transmission 0.70 to turn a 4.10 rear axle into a 2.87 rear axle. Wouldn't it make more sense just to have a a 1:1 connection with a 2.87 rear axle, and lower first gears to compensate? The efficiency of every gearset saps power some, i'd think something overspeeding would sap more power than 1:1 ever would. Playing with my lego sets as a child with the gears, any time I tried to make a larger gear, turn a smaller gear (so it went faster than 1:1 rotation on the input) had horrible resistance compared to anything equal or slower output speeds, so intuitively it feels wrong.

Cars like the Dodge Viper have a 0.50:1 OD ratio in some of the 6 speed transmissions. Large trucks had similar, some had double overdrives, or possibly even triple i've heard of with the brownie box at the very end after an overdrive transmission and an additional OD range selector, before hitting the two speed axle or the brownie box in the rear for slow speed mining off road crawling speeds up in michigan. Wouldn't the most efficient connection be the least # of gears to sap power?

[Frank Lee]

I agree and don't know why tall ODs are paired with short final drives, other than the desired bias is towards off-the-line punch.

[IsaacCarlson]

It gets the high torque load off of the ring and pinion.
Towing is where you really need a numerically higher rear end.
I have a 3.55 and 5 gears. I sometimes with I had another overdrive gear when empty. A lower rear end would only hurt me because I haul a lot of weight.

[darcane]

You have to consider the torque loads of every component in all situations.

Overdrive trannies allow you higher gearing with lower torque loads on the driveshaft, tranny housing, output shaft and related parts.

Imagine a 100 ftlb engine with 3:1 1st gear, .75:1 OD, and 4:1 final drive. 3:1 overall in OD, 12:1 overall in 1st means you get 1200 ftlbs at the wheels but only 300 at the tranny. Change that to 4:1 1st, 1:1 OD, and 3:1 final drive and you still get 1200 ftlbs at the wheels and 3:1 OD, but you now see 400 ftlbs of torque at the tranny, driveshaft, etc. and therefore need to be larger and more heavy duty.

[Big Dave]

On the face of it I agree with stillsearching. I would love to substitute a 1.6:1 rear axle into my F-350 and a seven-speed Spicer with a straight-through top gear. One gear mesh between flywheel and the tire. Optimal, but alas! Unavailable.

Ever try to find axles under about 2.73:1? My 3.08 is practically a one-of-a-kind. Even for dinky little GM 10 bolts, 2.56 is the lowest I ever heard of.

In my younger days you could get custom-hobbed hypoid gears. Maybe you can today, but I don’t know where.

Part of the problem is that the differential typically nests inside the ring gear. If you reduce the ratio the ring gear diameter has to reduce and then where do you put the diff? You want the diff as close to the vehicle center line as possible to avoid having unequal length drive shafts. (Unequal drive half-shaft lengths is what made the torque steer on old Saab 900 Turbos so wicked) You can’t reduce pinion size, that would allow engine torque to shear off pinion teeth.

FWIW slowing down the engine speed at a given road speed improves MPG because it reduces the engine friction loss. This is somewhat of a misnomer. The mechanical friction is almost negligible. Where the loss comes in is pumping air and exhaust gas through the engine. The torque needed to pump air goes up with the square of volume moved, hence the power goes up with the cube of volume moved. All air paths have a low range where the viscous resistance is greater than the compression power requirements. For My International 444, that is about 1800 RPM. From there the pumping loss skyrockets. At 3300 RPM power actually drops off as the engine consumes a greater and greater amount of power just pumping air.

So the slower the engine (a positive-displacement air pump) runs for a given road speed it pumps significantly less air and requires significantly less engine power to do so.

The theory works beautifully for me, even to the point of overcoming the inefficiency of having two gear meshes between the flywheel and tire.

[stillsearching]

Torque loads makes sense.. it's not an efficiency issue but a durability issue. Still 0.50:1 overdrives seem strange to me...

GM had a 2.29 10 bolt in the 78up Monte Carlo with certain engines, I think the 305. (not the stock V6) They might have used it until they had the TH200-4R, or might have stopped earlier being so steep but it was out a few years at least. It is the lowest rear end ratio I am ever aware of having seen in any car, ever, from any maker. (I went hunting once for a super-OD'ed lugger without a tall budget of adding a gearvendors onto an existing 4 speed) I was planning to use one along with a TH200-4R and a 6.2 diesel in an El Camino. :P

[darcane]

Originally Posted by stillsearching

Torque loads makes sense.. it's not an efficiency issue but a durability issue. Still 0.50:1 overdrives seem strange to me...

Yep, Torque is what kills trannies... Too much, and they break. Torque loads are a critical aspect of gearbox design.

This is exactly why Ford's transmission naming convention includes the torque rating in the name. i.e. a 4R70W is a 4-speed, RWD tranny that can accept a peak input shaft torque of 700 ft-lbs (not engine torque, but torque after the torque converter). Other manufacturers also reference torque ratings in the tranny name, though usually less directly.

[gone-ot]

...an overdrive "shifts" (pardon the pun) the engine RPM's at highway speeds down into the 'slower' 1000-2000 RPM range where the engines' SFC is "least" for "power" produced, ie: the "best" fuel-economy area of the engine operation.

[stillsearching]

Actually to be contrarian for a moment, even if torque kills trannies it also kills axles.

Look at a picture of a 2.29 ring/pinion then look at the tiny spindly little 4.56 ring/pinion... and don't tell me the latter is as strong as the former. It seems you either add weight in the axle or in the transmission, maybe not as much necessarily but it's still not completely simple.

Still with all other things being equal, assuming the trans isn't overloaded, am I wrong in assuming a 1:1 straight through ratio is the ideal? I guess what i'm really wondering is if anyone knows of an efficiency to compare the loss through a 0.50:1 overdrive vs straight through for instance. Just for curiosity's sake to sooth my instincts telling me it's wrong. Or because i'm curious if it's even worth engineering for for an economy project. (numerically low axles and lesser OD ratios)

[darcane]

Originally Posted by stillsearching

Actually to be contrarian for a moment, even if torque kills trannies it also kills axles.

Look at a picture of a 2.29 ring/pinion then look at the tiny spindly little 4.56 ring/pinion... and don't tell me the latter is as strong as the former. It seems you either add weight in the axle or in the transmission, maybe not as much necessarily but it's still not completely simple.

Yes, it kills axles too... but you have the load going through one set of large, continuously meshed gears and some rather stout components (input shaft and axles). Compare your ring and pinion gear to the sprag clutches, clutch packs, etc in your tranny and you'll quickly see a big difference. Plus, these parts are engaging and disengaging instead of in constant contact.

Originally Posted by stillsearching

Still with all other things being equal, assuming the trans isn't overloaded, am I wrong in assuming a 1:1 straight through ratio is the ideal? I guess what i'm really wondering is if anyone knows of an efficiency to compare the loss through a 0.50:1 overdrive vs straight through for instance. Just for curiosity's sake to sooth my instincts telling me it's wrong. Or because i'm curious if it's even worth engineering for for an economy project. (numerically low axles and lesser OD ratios)

You have a loss through any gearset. 1:1 allows you to eliminate at least one gearset because you can just lock the input and output shafts together. This would be most efficient if it were your only consideration.