The FINAL word on DFCO (sorta)
 

A while back I started this post looking for saturn specific parameters for dfco..
http://ecomodder.com/forum/showthrea...ers-26812.html

Turns out exact answers are non existant because the variables are so stinkin' high. I talked to an executive in advanced engine developement (from one of the domestic auto makers) about DFCO. I was probing for some more answers but left with more information but no actual answers. He said that there are thousands of lines of programming for DFCO alone in the cars ecu.

I struck out on my own to find some answers. I borrowed a listening device and hooked it to one of my fuel injectors. This thing is cool. It has long leads with clips on one end that plug into a small box in the car that also has headphones to listen with. It is often used to find a bad wheel bearing for example. You clip one lead near each wheel and can flip back and forth and listen for the offending item. I clicked it on a fuel injector so I could hear when it was pulsing (clicking) and when it was silent (not squirting the liquid gold we are all trying to conserve).

At first my ears were getting sore from the cheap heavy head phones until I discovered they had enough volume to hang from my rearview mirror and hear the clicking or lack of. I ran around for about the last ten days or so (which is three fill-ups, I drive a lot) with this thing cllicking away. Listening to the fuel injectors alone will get your foot out of the throttle. Learning the DFCO of my saturn was my main goal and I learned a lot but nothing super concrete. Here is the basics for my car. Yours will have its own parameters of course but you will get the idea.

It seemed to drop out of dfco at about 32 mph regardless of rpm
It dropped out in the 1500-1300 rpm range (a few times 1600 or 1200-ish)
If I backed off the throttle slowly it would not go into dfco (or not for a long time)
If I jumped off the throttle (fast tps closed reading) it jumped into dfco right away
If I wasn't fast enough on the closed throttle it would take 2-3 seconds (sometimes longer) before going into dfco.
If i got off the throttle quick enough and was in dfco but touched the throttle (even slightly for a second) it came out of dfco and would not go back in.
In 5th gear it would drop out from low rpm (usually about 1400-1500 rpm).
In 4th gear it would drop out from low vehicle speed (under about 32mph).
In 3rd gear it would drop out from low speed as well.

It was satisfying to be driving along at 55mph and jump off the gas and hear the annoying high speed ticking on my mirror go silent. I knew for certain that as I coasted down quietly that my fuel injectors were not shooting any gas. Another method for seeing this in action (aside from the listening tool or expensive instrumentation) would be to put a little led inline with one of your fuel injectors and watch the light blink or go off.

Another thought on DFCO that I have not read here is you go a little further than you think. I have heard mention about all that compression slowing you down. It is but not quite as much as you think. For compression to happen in the first place you need air to compress. With the throttle plate closed, very little air is entering the cylinder on each intake stroke. I absolutly know the value of EOC as referenced in my above mentioned post. I have avoided it on this car because the fuel pump is marginal and will leave me stranded someday I suppose if I don't replace it. I have driven over 3000 miles in the month I've had it and not been stranded yet however.

Last note (sorry this post is so long),
I now can easily feel when my fuel injectors kick back on if I have been in dfco.
I can not feel when dfco kicks in because it is usually exactly when my foot comes off the accelerator so car is slowing down anyway. Interestingly, while listening to my injectors, they kick back in (start clicking) a second or two before I "feel" the difference in the car.

I hope this helps someone.

Very interesting!

The first thing that I learned with my Ultragauge was that my car never entered DFCO. Except, I was wrong. In http://ecomodder.com/forum/showthrea...tml#post391443, California98Civic told me how to detect lean burn using my Ultragauge. I did not find the information in the manual, but I did see that you need to tell it to detect DFCO, and I just realized that in http://ecomodder.com/forum/showthrea...tml#post372073, SentraSE-R told me four months earlier. It is on page thirty-seven in the Ultragauge manual.

Please forgive me if you see me updating other threads with this information.

DaveLobi, this is some very cool research that you did. I never would have thought of listening to the fuel injectors! :) Thanks for sharing this with us!

for automatic transmissions..
 

I forgot to mention that I played with the listening device the other day also on my sons college special car..
1994 Toyota Corolla with an automatic trans and 251k miles. Didn't have much time in the car but found that it easily went into dfco with foot off gas. When we lost enough rpm to lose dfco I got it to kick back in right away by turnig the overdrive off (factory shift selector mounted switch) thus bringing rpm back up into range for dfco. When it dropped off once again, I downshifted into 2nd and it went back into dfco yet again.

Maybe I'm buying simpler cars ...

They go into DFCO if you let go off the gas, and start feeding again if rpm drops too low, a bit above idle, i.e. the engine idle control is taking over and will keep the car running @ idle engine speed.

ha, I didn't think there were any cars simpler than a saturn sl (well, at least any that have obd2 at least).

Remember that the idle rpm will be held higher if vehicle speed is up. Mine idles around 1100rpm until just before coming to a stop then finally drops to around 7-800rpm. I would like to have it drop to its lowest rpm anytime clutch is in but that would likely involve re-programing an ecu. Anyone ever beat this high idle at higher vehicle speed?

I'll have a look when coasting from higher speeds, but I haven't noticed anything like it so far.

Thanks for posting your results. I noticed you mentioned abruptly closing the throttle activated DFCO more quickly. That could be part of the reason why I have so much trouble getting DFCO to activate in my Saturn. I'm pretty smooth with the throttle transitions. Plus, it seems to take 2000RPM+ to get DFCO to activate, anyway. And I rarely rev the engine past 2500. So, yeah... ;)

I wish we had access to the computer programming, so we could tweak stuff like that.

Ya, funny that snap closing of the throttle plate (or tps) seemed to activate the dfco immediatly where a ease up on the throttle (as we are used to doing) made the computer think for a little while before deciding to activate dfco. Even more intriguing to me was if I was in dfco and I touched the throttle for even a brief second. The dfco would cut out and not come back on. The computer must be thinking something like "this guy may be ready to accelerate quickly" or maybe "doesn't know if he wants to accelerate or decelerate, better leave injectors on and ready". I don't know.

It would be fun to be able to mess around with our cars computers and play with things like this. I would make dfco way more common and user friendly (at least on my own car). A lot of the controllable items like this could all be done manually like a switch to kill the injectors but I don't want to go there.

Or maybe the programming looks at the delta TPS to help determine if it should activate DFCO. Tapping the throttle gives a very small delta TPS when you let off again. Just a guess.

Colder outside? WAI? Winter fuel?
 

Still playing with the listening device ocassionally.
It seems my car does not drop into DFCO as easily or for as long as it did only a few weeks ago.

What changed? Outside air temp? Winter formulation gas already at the pumps?

Maybe time for the warm air intake I've been thinking about trying. I'll update it I get around to it and if it makes a difference or not as far as dfco is concerned.

Anyone have ideas about this? Anyone have evidence that a wai may help here?
thanks,
dave

I made a IAC blockoff for Saturns quite a while ago. I sold a bunch on sixthsphere.com, and I also gave the drawings to someone else to make more. There are a few floating around on there if you want to pick one up.

The main reason I made it was so the car was way less annoying to shift - RPM drops like it should when you let off the gas and clutch in.

Also, given that you are in Michigan, you could run the car with an MS, or OBD-1 swap the car and tune the OBD1 computer.

complicated logic on DFSO is definately a reason why i like neutral coasting so much.

I already have a warm air intake, although it isn't as "hard-core" as some I've seen. I don't have the intake sucking hot air straight off the exhaust. But the air cleaner box is modified to suck in air from the engine compartment instead of outside. Combined with a partial grille block, it does increase intake temps, although not drastically.

Does that affect DFCO? I have no freaking idea. And to be brutally honest, I don't think it's worth the effort to install a stock air cleaner box and find out. But don't let that stop you from trying it yourself.

It was probably 25 years ago when I first listened to an injector and experienced DFCO.
The car was an early fuel injected Datsun 280 Z. They started using FI in 1975, under license from Bosch. You could hear DFCO with a automotive stethoscope while revving the engine with the car sitting in front of the main door to my shop.

Basically Nissan used DFCO in their first FI cars to eliminate the unburned hydrocarbon spike when the throttle was released. The throttle position sensor had to be at idle position and any imput would automatically stop DFCO. The early Z cars (75-78) had no catalytic converters in the Federal versions, while California versions had converters. My 76 even had no EGR system from the factory while still passing Federal emissions.

I believe the initial purpose of DFCO was to eliminate the air injection systems that were necessary to burn the otherwise unburned fuel when the throttle was released without DFCO. They even had fuel shutoff in some of the last carbureted versions of Nissan-Datsuns of the same era as the first FI Z cars. DFCO accomplished the same function as air injection had done before, but with the added advantage of 0 fuel delivery.

regards
Mech

Interesting. The Saturns do have air injection. That might explain why their DFCO parameters are so stingy.

The air injection didn't start until around 2000 on the "S" bodies.
My SL is a 1997 so no air pump.

Oldmechanic, How did they accomplish fuel cut off on the carberated cars?
I loved the "Z" cars. I had a 72 240, a 74 260 and several of the 280Z's

They put a solenoid in the idle-low speed circuit which shut off fuel flow when the throttle was closed until engine RPM dropped to a little over idle speed. The last of the NAP-Z 4 cylinders (200SX and trucks) had a heated screen in the carb to help with atomization (early mid 1980s). Also a "gulp" valve that had reed valves to allow air to pass through the engine when decelerating, without any fuel.

regards
Mech

Most older throttle cable efi engines have a throttle damper, I believe for emissions & something to do with reducing cylinder temps slowly.

These are a simple device attached to the throttle body that dampens the closing action. They are dead easy to adjust for a faster closing action. Might be worth have a play.

ProDarwin, what did the iac cutoff do to your throttle plate closed idle? I also hate when it revs up a little between gears, small but a waste of fuel and the rpm will be lower when next higher gear engages anyway.

Also what did you mean about running my car with an ms?

My Aygo dropped into DFCO with no throttle at any engine speed over 1400. It would then allow RPM to drop to about 900 before engaging the idle circuit.

At idle it would rev at about 0.85K rpm.

If I selected a gear (1st or reverse) at standstill the RPM would raise to about 0.95K which I assumed meant the tiny engine would not stall when drivers pottered about a car park and would also not stall when pensioners held their steering wheels hard over causing the PAS pump to draw extra energy from the engine.

Digging up an old thread but interesting (to me anyways) new to me info on dfco.
While my old Saturn was picky when it would go into and drop out of dfco my newer Pontiac Vibe loves going into dfco. The Saturn cut out below about 1400 rpm or below 32 mph. This (Toyota 1.8 liter) Vibe goes in easily, stays in, if out-jumps back in easily, will stay in all the way under 1000 rpm, doesn't seem to care about speed. Basically anytime (in closed loop) I want it, it is there. It is a 2007, maybe that is a big part of the difference. Better programming for more often used to help battle the CAFE standards?

My Del Sol ('93 Acura engine) drops into DFCO whenever you let off the gas, in any gear, and the RPM is above 1000. It's really as simply as that. Warm or cold, 1st or 5th (or even reverse!). I believe it detects "letting off the gas" via throttle position, because I was having some trouble getting it to go in for a while and found the screws on my TPS were a bit loose. After adjusting it back to the position it was supposed to be in, and tightening it, everything again worked properly.

It seems a similar system is used in my Insight but I've only had the car a few weeks and am not 100% sure yet. I only use the onboard instrumentation, and not an MPGuino, which reads the injector grounds and tells you when they're not firing.

Well, the Vibe is gone and now I'm driving a 99 Saab 9-3 (5 speed 2.0L turbo).
I don't know what the manufacturers fuel economy ratings are but the first tank gave up just over 32 mpg. I have since fixed the pixel screen (replaced ribbon wire) and am using the average mpg on the dash. I have eeked it up just over 34 over the last couple hundred miles (mixed driving) with a tiny bit of eoc and some dfco.

This one, I don't think Saab was ever known for high fuel economy, is a little pickier as to when it wants to cut out the injectors. I got out the chassis ear again for a listen today. It will not go into dfco at all in first or second. I even tried running second way up (to get speed above dfco range in third gear) to around 4k rpm and it would not go into dfco. It seems to want to be real close to 2000 rpm or higher to go into dfco and will drop back out around 1300 or so. If coasting all the way down (below 1300) like I do approaching my driveway I now know to dfco down to about 1300 rpm then bump to neutral rpm (just under 1000) for the rest of the coast.

By far, the most aggressive that I have driven (easy to keep in dfco) has been my sons old (270k miles) 1993 Toyota corolla (automatic trans).

My Aygo wouldn't go into DFCO under 1400 either, although it was happy to stay in it until just over idle (~950).

Be careful of shutting the engine off when running due to Turbo temps - once the engine is off the oil is no longer circulating and oil in the turbo can overheat and cause seal problems.

I know very little about turbos. Is the turbo temp close to the same at all ranges of boost? I mean, with my light driving the boost gauge on the dash, for what its worth, stays low almost all the time. My driving doesn't ask for the boost or horse power. I also rarely eoc, often once per trip for the last half mile of my road where I have to come to a stop anyways. I find p&g/dfco added to neutral coasting more than adequate to appease my desired mpg.

The temp will go up with the level of exhaust boost employed so yes a lot of use will raise temps.

In the mid-late 80s a lot of high performance cars with Turbos had issues when drivers would blast down the motorway, then stop for a pee and shut the engine off straight away - the oil would stop circulating and the oil in the turbo would "boil" and cook. A few Escort RS Turbos had this issue early on.

http://fastestlaps.com/photos/Ford_E...4708925d79.jpg

Eventually this would lead to premature issues with oil seals becoming brittle, fail to seal, and then lots of blue smoke as oil leaked into the turbo on boost - and eventually turbo failure.

Translate this to ecomodder style driving - maybe a harsh pulse and glide with EOC might not be too great for a turbo, although the material seals are made from is much better these days.

When driving a car with a turbo I still tend to slow down gradually from high speed so the turbo has time to cool before shutting off completely.

On the positive side you are driving a SAAB. Although it is based on a GM product it still has a SAAB engine (not a GM one Thor be praised) and SAAB were early into turbos and making them reliable for large mileages. People here regularly get 400hp+ from a 2.0 SAAB without it going pop even after 100K+

If you want to see how GM did it in a previous generation, the ecm for the 86-87 turbo 3.8V6 Buick Grand Nationals has been completely cracked and decoded. It is OBDI, not II. Go to Turbo Regal Web Site, resources, tech info, ecm and sensors, ecm secrets revealed (this is the disassembly of the code in the rom and the 2K of code in the removable 4k chip). Then get chip secrets revealed and more chip secrets revealed, these are the 2k of data tables in the chip with descriptions and formulas, in spreadsheet form. This ecm used both dfco, and for throttle closing conditions not quite vigorous enough for dfco it also did dfe, decal fuel enleanment, where it cut the fuel 10-15% but didn't turn it off. Anyway, you can see the control parameters in the spreadsheets and the code if you want the real explanation :-). GM tended to reuse code so I imagine other ecms from this time frame were pretty similar in their control strategies. No telling how it evolved later, though.

Two other web sites with lots of gm ecm info is the Welcome to diy_efi/efi332 site and Mark Mansur's site, TunerPro and TunerPro RT - Professional Automobile Tuning Software. Tunerpro is a freenware (basic version) and shareware) full version chip editing program, and he has lots of definition files for various ecms so you can browse for yours and see what data tables are accessible. I've done lots of chipmaking for the buick TR's, starting with a calculator and doing manual hex conversions and editing and then using several freeware packages as they were released, and the free tunerpro is very good. I've never had to deal with obdII so I don't know what its capabilities are there, sorry.

No matter what brand of vehicle, I've always been told that you should allow at least 30 seconds of very light throttle or idle after making any boost, to allow the turbo to spin down and cool off. I've always tried to do that and so far have never lost a turbo seal, knock on wood (25 years of spirited driving and drag racing a Buick TR and 6 years so far with a turbo Duramax diesel pickup).

Also, a spark ignited gasoline engine has maximum pumping losses when the throttle is closed. It isn't that the engine compression is slowing the car down, it is the pistons being forced down with the intake valve open and the manifold at maximum vacuum, so the vacuum is pulling upwards on the pistons. Any air that is ingested will be compressed as the piston goes up on the compression stroke, but then on the expansion stroke the air will be re-expanded and minus friction losses the work down on compression is recovered, at least until the exhaust valve opens and you lose the rest of the stored energy. Diesels do not have throttle plates (in general, I know there are a few but ignore them so we can keep this simple :-)) which is why they are so efficient at idle and light throttle; they don't have the pumping losses of the spark ignited gas engines. That's also why when you lift off the accelerator on a diesel you don't get nearly as much engine braking. Diesels move the same amount of air at a given rpm no matter how much fuel is being injected, at least without a turbocharger. The power level is controlled by how much fuel is injected. That is one of the advantages of the direct injected gas engine, which is why manufacturers are working so hard to develop it; it lets them do away with the throttle and inject the fuel at the last possible moment to effectively raise the octane so they can raise the compression ratio to raised the efficiency. You could test this with a fuel injected engine and a kill switch, just do a short coast down test in gear with engine off and throttle closed, and then repeat with throttle wide open. Yes, you will be injecting fuel at wot but if you only do this once each way it shouldn't hurt any thing, and you should coast a good bit further with the pedal to the floor which will demonstrate how great the pumping losses are at idle. Just don't turn the kill switch back on with the pedal still on the floor :-).

Had to go out for awhile, and right as I was leaving I realized I didn't give a complete description of the coast down test to show throttle closed vs. wot with a kill switch. If the kill switch is wired into the injectors then no fuel will be injected and the test is completely safe. If it is wired into the spark side of the ignition system then fuel will flow while the pedal is down and the engine is being driven by the car so when you turn the kill switch back on there might be a backfire from the fuel in the exhaust system, so make sure the throttle is back at idle position before restarting. If the kill switch is wired into the crank sensor then the ecm won't see pulses and won't inject fuel, so again the test is completely safe. Sorry to be so longwinded.