Accidental Transmission discovery (bump started my automatic)
 

Lately, I have been throwing the Auto trans in my car in Neutral when approaching a stop light, and if I know it will be a long one, I shut the engine off. Typical hypermiler stuff. I usually will turn the key off for 5 seconds and then turn it right back on. Anything less than 3-5 seconds means the engine has a chance of refiring.
Here's the incredible part, last night I was talking on my cell phone to my wife while driving. As I came up to a light I knew was going to be a long one, I shut the engine off. When I turned the key on five seconds later, the car re-started! I had shut it off in drive and the automatic trans had kept the engine turning! I didn't think that was possible, but I had just bump started my car with a slush box!
On the remainder of the trip home, I performed my discovery again and again and ran through a few other scenarios. It will easily restart after long periods (10-15 seconds) as long as I am above 30 mph. It seems to want to downshift at that point. It seems that the trans will keep turning the Stator as long as the TC housing turns fast enough to keep pressure on the clutch packs. It will also not bumpstart if the engine was shut off in neutral and then put into drive after the engine is off.
It doesn't coast as far (naturally). I have a 2000 Cirrus with a 2.4L and a 41TE (A604) trans.
I am not recomending anyone try this as I have no idea if this damages the trans or not. You should also not do this if you are close behind someone.

There are varying opinions about coasting in N hurting the transmission but I think everyone is in agreement that coasting engine off in D will hurt it. N is ok for some. Look in the owners manual and see if flat towing is allowed and that will give you an indication of whether N coasting engine off is all right.

I noticed this myself...quite by accident as well. I did about the same thing you did, I was coming to a light and just killed the engine without thinking about the shifting and I then clicked the ignition back and it fired again...I was pretty excited as well but I decided since people say that turning the engine off while in D regardless of whether you're moving or not, is a bad idea that I would not do it intentionally. It's definitely worth more looking into though because if it doesn't harm anything it would be useful for sure.

I could understand towing in drive would be harmful if you didn't have clutch pressure to start with. In this case, the pressurized clutch packs turn the stator which turns the TC housing which turns the oil pump which keeps the the clutch packs engaged. Also, everything is getting lubricated as well with the oil pump turning.

Hmmmm....

What is particularly cold out (or was the fluid cold?)

RH77

Hrmm... Makes sense to me... Bear in mind that you'll be causing a helluva lot of friction material wear as the pressure slowly drops off and you start getting slip on the bands.... I'm saying this because you mentioned you must be going over 30mph. So something tells me the bands are slipping at even higher speeds, but the pressure is high enough that it can still turn the TC...

Something to think about... You're turning the TC with a whole bunch of extra load on it (engine brake).... If your TC is designed for this (that is, keeping the TCC locked for engine braking above XX speed) - no problem. Otherwise, there could be reason for concern.

Just tried it + Why?
 

I just tried it, and it worked -- presumably when the TC is locked. It will downshift and unlock with the engine off; which is when it didn't work.

Next item up for bids on the Price is Right:

Why? I'm trying to think of a situation where this would be applicable (to turn the car off in Drive and key-back on before a stop). Generally on decel, the injectors are cut anyway, or you could downshift to keep the revs up into the fuel-cut range...

RH77

Some valid observations and questions.

On the transmission/bump starting phenomenon, here is the deal. The Automatic transmission's in use, typically have one pump, which runs off of the front main shaft on the transmission. If the engine was running and you leave the car in gear, so the pump continues to turn, then the engine is still turning over, with no fuel being injected. However, once the car speed slows down to where it tries to downshift, the electronics are off, the oil pressure drops and at some point you have the potential for slippage in the transmission.

If, in most cars, with FI, if your in gear, you let of the pedal completely, the ECU will turn off the injectors, so no fuel is flowing. Once your engine RPM gets down to around 1200, or so, the ECU starts firing the injectors again, so their is a smooth transition and the engine is now available to turn the pump on the input shaft of the transmission.

The amount of fuel savings is basically the same, but you haven't placed your transmission in a scenario where it might get a little low on oil pressure. If it gets low on pressure, you might not notice it, right away, but you can be assured you have some degree of accelerated wear on the clutch's and at some point it will fail, sooner, rather than later.

I probably won't make this a habit, but I might try this for 50 miles or so to find if FE increases a bunch. However, I beg to differ on everyone's claim that the injectors shut off when you let off the pedal. They still inject a small amount to keep the engine running, even on deceleration. My proof? If they did shut off, the effect would be exactly the same as turning the key off in gear. Your trans would be driving the engine.

I've experienced this myself - engine shutdown with the tranny still in D will keep the engine turning over from tranny drag. I don't think there's any harm to the tranny at all - it's all turning from end to end, so the lubrication is still working. Same for the engine, for that matter. It also keeps your brake vacuum and steering pressures up.

If you live in a really hilly region as I do, a long downhill with the engine off allows you to save brakes and keep the speed down, so I don't see a big issue here.

I agree with elhigh. As everyone has experienced before, during deceleration, engine rpm's will not return to idle unless you throw it in neutral. This means that the trans is driving the engine, which is a common occurance.

A quick calculation shows that stator rpm in top gear at 30 mph is 1090rpm. If the stator is running faster that the engine by 25%, then engine rpm is 812, which is probably lower than idle. Also, the slower the TC speed, the less efficient it will be. There is probably an exponential drop off in which the TC housing would finally quit turning.

More quick calculations based on 75% engine to trans ratio;
50mph = 1814rpm trans, 1360 rpm engine
40mph = 1451rpm trans, 1088 rpm engine
These engine rpms should be adequate to provide full prssure to the clutches as it is the engine/TC housing that turns the pump, not the input shaft.

One more reply. The correct way to do this is to set up a kill switch. That way I could leave the key on and still watch engine rpm's. As long as I would restart it before engine rpm's went below 1000rpm, I should be fine.

My shop manual states:
________________________
Fuel Cut-off Control

*During deceleration with the throttle valve closed, current to the fuel injectors is cut off to improve fuel economy at speeds over following rpm:

• B18B1 engine: 970 rpm (Canada model: 1050 rpm) My Engine -- U.S. Version
• B18C1 engine: 970 rpm (Canada model: 1050 rpm)
• B18C5 engine: 1300 rpm (Canada model: 1400 rpm)

___________________
2 years ago, I rigged up an injection deactivation device. On decel above 950 rpm or so, there was no difference in engine operation.

RH77

EDIT: Adding...

So, then the engine is still operational providing power from the alternator, and all other ECU-controlled functions. The resistance between the compression of the engine and the turning transmission may be a point for wear if IGEOC (in-gear engine-off coast)...

Believe what you wish. I'm basing my comment on the fact that on my 89 Honda, by monitoring the Oxygen Sensor Voltage, it is clear that the injectors completely shut off, if I am in gear, my foot is off of the pedal and my rpm is above 1200. You are correct, it is just exactly the same as having the engine off, with the key, while being in gear.

You might have a point, in an Automatic. On my wife's Chrysler, I think the injectors do inject some fuel, largely to give people the impression that the car is able to coast farther than it really is.

My primary point was that if you want to risk killing your automatic transmission, knock yourself out. It will have an effect on it, and you will have to replace it sooner, rather than later. I personally don't like replacing or rebuilding transmissions any sooner than I can possibly get away with. I shared the information I did with you, so you'd have a better idea of the risks associated with it.

AE340 does the same, it will even keep to motor turning if you downshift (either manually or electrically). however once you put it in neutral with the motor running, it is all over.

how can i make the engine go to idle?
 

99 windstar 3.8L auto
i pop the trans into neutral and coast engine on. it coasts a lot farther this way.
Question how can i get the engine to idle, it still tracks the vehicle speed even in neutral.
If i put a momentary interrupt switch in the VSS circuit do you think the engine will idle? this bugs the daylights out of me, i know it uses a lot less gas at idle. and i know it won't hurt the trans. coasting in neutral as long as i let it idle.

i believe the ECU is adding air with the Idle Air Control (IAC) which increases fuel because its MAF system.

John, my Cavalier keeps the engine turning around 1100rpm when I NC. But the Scanguage reports idle fuel consumption rate, 0.4gph, regardless of rpm. As long as I'm burning fuel at idle rates, I'm not concerned what the rpm is.

I don't think interrupting the VSS will get your engine to idle. I think the ECU is sending idle-amounts of gas, and our engines are just getting a little extra push from the not-completely-decoupled slush boxes.

I think it's VSS and Vacuum. I've wondered the same...

I'll dig out the shop manual and see if I can figure out what mine does...

RH77

Interp
 

This is the best I can come up with from the shop manual...

This is on the '98 Honda/Acura 1.8L base Integra engine (B18b1)

The PGM-FI system regulates idle based on:
Overview

1. Sensor Inputs
2. Control Module Program (the brains)
3. Result/Output (sensor or device)

Process
Sensor Inputs Relevant:

• Vehicle Speed Sensor
• Alternator Charge Signal
• Primary Heated O2 Sensor
• Transmission Gear Position Sensor
• Batt Voltage
• Fuel Tank Pressure Sensor

These and others are fed into the ECM. The "Electronic Idle Control" program is part of the computer, and likely outputs to:

• Alternator Output Control
• Idle Air Control Valve
• Ignition Control Module
• Fuel Pump

Hope this helps.

RH77

What it doesn't really get into is the following observation:

Lots of cars I drive will increase engine speed based on brake vacuum at idle. With the vehicle stopped and in P or N, pump the brakes quickly and note if idle increases.

On your newer vehicles, doesnt turning the ignition off turn off the SG as well?
Some vehicles have electronic speedometers vs cable and you mite not see any #'s for mileage by doing the engine off trick..

keep in mind, cars weren't designed to drive in motion with the engine off..
The transmissions weren't designed to do this and the results could be good or bad..
The manufacturer wouldnt know what to say if you asked if it was ok, they'd just take the easy way out and say it's bad insted of hard facts..

Talk to some transmission shops and ask people that work on these if there would be any harsh effects..
thats my mindset.

or get a manual transmission ;)

during coast down, engine on, neutral, the engine maintains high idle to either to burn hydrocarbons(keep O2 sensors hot)(or cats) or keep the speeds in transmission matched, should i pull it back into gear. if i unplug the VSS for a test, i wonder if it will shift at all.

My transmission is dumb; no connection to the ECU. So my high NC engine rpm's can only be coming from input from the trans. I hadn't considered computer controlled trannies, so we may have been apples vs. oranges. But I think that even if you can fool the ECU into thinking the engine should be idling, you'll still get increased rpm induced by the transmission. As I said, my ECU is feeding idle quantities of gasoline, but my engine is turning 2X idle rpm on the highway. Engine rpm reduces as the vehicle speed reduces.

Ok, I'm back from the weekend and I have to eat a little crow. On my Cirrus, I have come to the conclusion that the injectors do not shut off on deceleration, but my wife's Kia Minivan does shut down as the deceleration is much more than in my Chrysler (the Cirrus coasts farther). In my car, there is more deceleration by turning off the key compared to coasting with my foot off the pedal. Also, I lost FE by shuting the engine down and bump-starting.

My engine idles around 700rpm. Smaller engines idle faster, YMMV