I too, do know that some tranny's do a straight through the mainshaft (1:1) in a higher gear for my old BMW 323i did just that. Unfortunately, the way I learned was through a bad bearing on the secondary layshaft that created quite a whir in the lower gears but the tranny was quiet as a church in fourth. As described previously, the input and output shaft were locked by a sliding collar, which really is pretty much how gears to the layshaft for other ratios are engaged (sliding collar). This then cut the layshaft and its noisy bearing out of the loop in 4th gear and I could imagine I had a good tranny:D

And the layshaft, not being driven by any engaged gear, did not turn at all, so I imagine there might be some small increase in efficiency by the straight-through 1:1 method, avoiding two driven gears and whatever added frictional losses that may entail (98% per gear plus added bearing losses...?)

Of course, that doesn't necessarily mean that is how the RX-8's box is configured, but the above method for achieving a 1:1 ratio is the simplest and thus, I would imagine, the most likely.

 

What year BMW rhumb?

I've plumbed through every shop manual I could get my hands on (mostly cars from 1990 onward) and none of the transmissions behave as described. I'd really like to see how its done.

Here's why I'm having trouble visualizing this:

1. The input shaft from the clutch is on a particular axis. This connects to one of the shafts in the transmission directly. This shaft (main, lay, whatever you want to call it, we'll call it the main shaft), the mainshaft has a collection of gears on it that can be engaged/disengaged from the shaft using sliding collars.

2. Running parallel to the mainshaft is another shaft (let's call it the countershaft). It is connected to the transmission output shaft which drives the prop shaft to the differential (in the case of a RWD). While the countershaft is parallel to the mainshaft, it is on a _different_ axis. Like the mainshaft, it mounts a number of gears which mesh up with the mainshaft gears, but these gears are fixed.

So, in the type of tranny just described (and as demonstrated in the link I provided), the input and output shafts of the tranny are on parallel, but different axes. Thus, it would be impossible to simply lock them together, you'd need a gear set to mesh them - even if its a 1:1 gear.

This raises the question then of "how do you design a tranny such that the input and output shafts are on the same axis?" I know of one way, which is to use a planetary gear setup like an automatic, but that's not how manual trannies are setup (at least most production trannies). The only way I can see to do it is with extra gears, which will increase losses, not decrease them.

Again, if you could tell me the year of your car, I'll see if I can dig up a manual because I'd like to learn about it.

SC

 

All the rear-wheel drive transmissions that I've opened up (RX-3 4-speed, early RX-7 five-speed, and Ford Cortina 4-speed) have had the input and output on a common axis, with a secondary parallel shaft (the layshaft). In all but the 1:1 gear, the power is transmitted from the input shaft to the layshaft via selectable gear ratios, and then back to the output shaft via a second, constant-mesh gear set. While the input and output shafts are concentric, they are not joined and obviously turn at different speeds. Except in the 1:1 ratio, when the sliding collar locks the two together rather than engaging another ratio to the layshaft. Here's a cutaway of the Ford MT75 gearbox as an example. If I find any others, I'll post them.

 

Why do people say 3rd gear anyway? In most modern 5sp gearboxes, 4th gear is the 1:1 gear...

Here's a powerflow diagram for the FD rx-7 gearbox in 4th gear.

Not sure but it looks to me like the input and output shafts are on the same axis and 4th gear is simply an interlock between them. On the gearbox description page there is mention of cogs for all other gears except 4th...

 

Tremec T56 - 6 speed


service manual - there is no "4th gear"


Most RWD tranny's are this way, with a direct drive layout

 

Thank you gentlemen. Interesting that an automaker would choose to make every gear but the 1:1 less efficient by requiring an extra gear interface to return to the output shaft.

Good show.

Nonetheless, it doesn't explain the less than 1% variation in load controlled tests in multiple gears

SC

 

Actually I'd find it more interesting if they chose to implement it in a more complicated/inefficient manner than neccessary. Interesting, but probably not surprising...