AT to Semi-AT?

This is going to be a long and wild speculation so I didn't want to inject it into the ongoing AT to MT thread here:

http://www.msprotege.com/forum/showthread.php?t=123633926

First of all, I agree with pretty much everybody in that thread that it's nuts to do the the work to convert an automatic into a manual, since what you end up with is something you could have for less by trading the car. But what if you want to convert the AT to something you can't buy? For lack of a better word let's call it a SemiAutomaticTransmission (SAT). There probably is already a term for this beast but I don't know it, some race cars already have transmissions more or less like this.
EDIT. Looks like the term is semi-automatic, see for instance:
http://en.wikipedia.org/wiki/Semi-automatic_transmission

To go from an AT to a SAT these would be the conversion steps:

1. Replace the torque converter with an electrically controlled clutch. This clutch would also have RPM sensors on both sides. At the same time it would seal up the AT with a plate of some sort (to close the holes that previously let AT fluid pass between the AT and the torque converter.)
2. Reflash the car's computer with the MT code, so that the car's computer won't know or care what the SAT is doing.
3. Install a controller (small computer with a bunch of IO ports) and wire to it:

At this point it should be possible to have the car shift automatically (in D) or semiautomatically (using the sportronic part of the console shift). To shift gears the new controller releases the clutch, then shifts the existing automatic transmission via the solenoid lines, then engages the clutch while monitoring the input and output RPM sensors. In other words, the computer controls the clutch instead of the driver via a clutch pedal.

This new setup would allow for some features not provided by either the existing AT or MT transmissions. For intance, on the control panel G there could be a little dial or switch to control the point at which the transmission shifts (high or low RPMS), so you could set it to run sporty or economical "automatically". The sportronic could be used to downshift all the way to neutral. For instance, for those situations in really slow stop and go traffic where in a manual you just get up to 3-5mph and then put it into neutral and glide. The sportronic could conceivably also be used to go into Reverse so long as the new controller enforces some safety precautions. For intance a single downshift into R might beep (through the new control panel) and a second downshift would actually put it into reverse. Maybe best not to do this though. (Shifting from N in Sportronic to D should leave the car in N, only shifting N to D in the regular part of the transmission should actually put the car in drive. Otherwise manually downshifting to N there's no way to get to R or P without going through a forward gear.) The cruise control range could be extended downward, for instance, for those trips where the speed limit is 30 and there isn't much traffic. There are probably others I have not thought of.

At the very least this conversion would result in a car that drives just like a regular automatic but gets much better mileage in the city. If the price of gas keeps going up there might well be a point where it makes $ense to do this conversion, if it cost, for instance, $1000. A conversion kit might fit more than one car, so there may be some commerical potential here. $1000 probably should be within range since an installed replacement clutch is often around $500, and that leaves $500 for the other parts.

I don't see anything on this list that's impossible mechanically or electrically, or for that matter even all that difficult to do. Some parts aren't off the shelf but neither are they rocket science. For instance I knew a guy in college (way back in the late '70s) who put a solenoid on the clutch on his Fiat so that it could be shifted electrically. And electronic controllers are really cheap these days. In terms of installation it consists of one major surgery, replacing the TC with the fancy clutch, and some rewiring. On the down side, when all is said and done, it's still just a 4 speed transmission, albeit a really high tech one.

Just a thought...

i think at this point just sell ur automatic and buy the sport auto that comes w/ the 03's

I thought he had a sport auto.

oops, my bad.. didnt read all the way.. so the whole idea is to get a ferrari enzo like tranny?

mpf1v3 said:

oops, my bad.. didnt read all the way.. so the whole idea is to get a ferrari enzo like tranny?

Click to expand...

Sort of, but very much on the cheap, by reusing the existing AT gearbox.

Just thought of one other thing, with the proposed rebuild the sportronic action could be reversed so that pushing the lever forward shifts up, instead of down as it does now. The way Mazda built it always seemed backwards to me, pulling back to upshift.

What you just described would cost at least five figures to engineer and build. I don't understand why you'd replace the torque converter with an automated clutch if you're going to keep an automatic trans behind it. There are already transmissions like this in high end cars and even they haven't all been perfect. Launch control is a big issue/problem. What you're talking about is incredibly complex.

Jack Daniels said:

I don't understand why you'd replace the torque converter with an automated clutch if you're going to keep an automatic trans behind it.

Click to expand...

Couple of reasons. First, the AT already switches gears under electronic control. If a 5 speed MT was put in instead a very complex electronic to manual shift mechanism would have to be constructed. Second, if the 5 speed went in the complex clutch would have to fit into the space the normal clutch used, and it might not fit. (To be fair, flip that around, and some sort of activator could be attached to the normal clutch where the pedal cable normally goes on.) Third, the AT is already in the car, so keeping it should reduce the cost of the conversion.

While the initial engineering might be pretty expensive the per unit cost after that doesn't need to be very high.

Jack Daniels said:

There are already transmissions like this in high end cars and even they haven't all been perfect. Launch control is a big issue/problem. What you're talking about is incredibly complex.

Click to expand...

Those cars tend to have 400hp, have high compression, and are often turbocharged. The P5 should be a lot less finicky.

pasadena_commut said:

Couple of reasons. First, the AT already switches gears under electronic control. If a 5 speed MT was put in instead a very complex electronic to manual shift mechanism would have to be constructed. Second, if the 5 speed went in the complex clutch would have to fit into the space the normal clutch used, and it might not fit.

Click to expand...

the MT and AT bell housings are pretty much the same size.

i think it would be easier to just do the AT to MT swap if you're gonna try to do all that.

pasadena_commut said:

(To be fair, flip that around, and some sort of activator could be attached to the normal clutch where the pedal cable normally goes on.)

Click to expand...

Turns out there is already such a beast designed. I couldn't find it for sale anywhere though:

http://www.valeo.com/automotive-supplier/Jahia/pid/868

Anyway, as Valeo envisions this product there is a pedal for the driver that is connected electrically to the clutch. It's not too far from there to removing that pedal and controlling the clutch and the transmission together electronically.

Seems like Valeo may have already done a big chunk of the key development work to make this fly.

pasadena_commut said:

Couple of reasons. First, the AT already switches gears under electronic control. If a 5 speed MT was put in instead a very complex electronic to manual shift mechanism would have to be constructed. Second, if the 5 speed went in the complex clutch would have to fit into the space the normal clutch used, and it might not fit. (To be fair, flip that around, and some sort of activator could be attached to the normal clutch where the pedal cable normally goes on.) Third, the AT is already in the car, so keeping it should reduce the cost of the conversion.

While the initial engineering might be pretty expensive the per unit cost after that doesn't need to be very high.



Those cars tend to have 400hp, have high compression, and are often turbocharged. The P5 should be a lot less finicky.

Click to expand...

I think more HP & TQ would actually work to your advantage with a setup like that. My P5 bucks under low RPM conditions in 1st. I always have to feather the clutch. My GTO can withstand going below 1000rpms in first gear without any problem.

I don't want to shoot you down, it's interesting to talk about stuff like this. However, in reality it's a very complex, difficult setup you're talking about. Just think about all the varying conditions the clutch would need to 'know' about in real world driving. I think you'd be better off with replacing the torque converter with something that stalls a little higher and putting in a manual valve body.

pasadena_commut said:

Turns out there is already such a beast designed. I couldn't find it for sale anywhere though:

http://www.valeo.com/automotive-supplier/Jahia/pid/868

Anyway, as Valeo envisions this product there is a pedal for the driver that is connected electrically to the clutch. It's not too far from there to removing that pedal and controlling the clutch and the transmission together electronically.

Seems like Valeo may have already done a big chunk of the key development work to make this fly.

Click to expand...

That's interesting for sure. It just seems to me that this is a bit of over-engineering and over-thinking. I will always prefer an old-fashinoned manual transmission.

I was just taking a quick read over your subject and it kinda sounds to me like your wanting something like lentech makes. The main downside to any auto trans it the take off from start I see no reason for having a clutch for all gear shifts. All you need to get around this is a Trans brake and if you want to change the RPMs a different torque converter will help. Lentech sells a plug and play MANUAL SHIFT MODULE but im not sure exactly how it would work. They dont list all that much info on it. If and when I blow up my tranny from racing I will get one of the Lentech ones with all the bells and whistles.

http://www.lentechautomatics.com/4f27eproducts.html

And then if you want more control over shifting times just get something like this.

http://www.powertraincontrolsolutions.com/content-4.html

This one is not made of our car but im sure one can be found.

Jack Daniels said:

My P5 bucks under low RPM conditions in 1st. I always have to feather the clutch

SNIP

I think you'd be better off with replacing the torque converter with something that stalls a little higher and putting in a manual valve body.

Click to expand...

The clutch I described measures the rpm on both sides. The CPU controlling it would know where the Mazda's engine bogs down and would modulate the clutch to keep the motor side above that RPM point. Imagine the worst case scenario - the car is at a complete standstill, there is no pressure on the accelerator at all, the engine is idling at 500 RPM, and the driver shifts into 1st. Obviously if one let the clutch out on a manual at that point it would jerk forward slightly and stall. The system I'm describing wouldn't even begin to let the clutch out since the motor's RPM are too low. Bizarrely, the car could idle in 1st and not move. However, when the accelerator is finally pressed the RPM sensors are consulted and the clutch eased out so that the motor never stalls. The electronics should be able to feather the clutch faster than any human could. So stomping on the gas when stopped in first wouldn't VROOM the engine as one might expect, rather the clutch would let out as soon as possible and load the engine at relatively low RPMs.

Similarly, when stopping let's say the car is in it's most "manual" mode, the transmission in third, and the driver just steps on the brake and goes all the way to a stop. When the car has slowed down to the point where the (computer knows) the motor would start to stall the clutch is released automagically. In a more automatic mode the car would automatically downshift through the gears and the clutch would let out as above at about 4-5 mph. In either case the transmission should then shift to neutral, so that the clutch can be released, just as it would have been on a "real" manual car which is stopped. To start again in this mode requires an explicit upshift to 1st.

I don't understand what would be gained by installing a manual valve body.

By manual valve body, I meant simply keeping the trans an auto with a valve body that requires manual shifting. Then possibly changing the torque converter to one that brings the car more into it's powerband. That would be a decent, easy setup IMO and you could spend all the savings on a turbo or some other power adder.

Honestly, if I were looking for performance, I'd just buy a faster car. We're in the golden age of HP right now. I used to be a hot rodder, but nowadays you can get incredible performance and a warranty to go with it. :woot:

Here's a Toyota with more or less the transmission setup I described:

http://www.wheels24.co.za/Wheels24/NewModels/0,,1369-1374_1523600,00.html

The relevant part:
The Verso isn't sold in the US but seems to be some sort of Corolla variant. I didn't see this transmission listed as a Corolla option in the US though.

Check out the 1960's technology. Ever driven one of these?

www.vwar.org

My intention on posting here is not to argue and say people have funny ideas. I Understand all that you are saying and I am only contributing information that others and yourself might find useful.<?xml:namespace prefix = o ns = "urn:schemas-microsoft-comfficeffice" /><o></o>
<o> </o>
Imagine the worst case scenario - the car is at a complete standstill, there is no pressure on the accelerator at all, the engine is idling at 500 RPM, and the driver shifts into 1st. Obviously if one let the clutch out on a manual at that point it would jerk forward slightly and stall. The system I'm describing wouldn't even begin to let the clutch out since the motor's RPM are too low. Bizarrely, the car could idle in 1st and not move.<o></o>
<o> </o>
This same effect is somewhat achieved with a high stall torque converter.<o></o>
<o> </o>
However, when the accelerator is finally pressed the RPM sensors are consulted and the clutch eased out so that the motor never stalls. The electronics should be able to feather the clutch faster than any human could. So stomping on the gas when stopped in first wouldn't VROOM the engine as one might expect, rather the clutch would let out as soon as possible and load the engine at relatively low RPMs.<o></o>
<o> </o>
This sounds good and all but having the clutch engage at a relatively low RPM will do nothing for performance but put us back at the stage of where we are now with the stock auto. We need a higher engagement RPM to get a performance gain. <o></o>
<o> </o>
If you install a PCS I have been told that you can calibrate it to automatically downshift to suit your current vehicle speed but it will not upshift automatically when you start to accelerate. This would be up to the driver. If you can program it to do this I am sure you can have it return to neutral when you come to a complete stop. <o></o>
<o> </o>
I do see you point about not having a torque converter. Removing that behemoth power sucking fluid flinger would help the engine performance. I wish they made a light weight flywheel for our autos but I have not yet seen the stock one up close to say that it is heavy.<o></o>

Mazdacub said:

I do see you point about not having a torque converter.

Click to expand...

That would be the main benefit. I'm not really after better performance, just better efficiency. Still, the sportronic would probably feel a bit more sporty if it didn't have to draw power through a fluid coupling.