Question about Euro Carrera 3.2 DME
 

Can some of you engine gurus explain to me exactly how the euro DME controls the decel function (rapidly dropping RPMs)? I drove a friend's car at an autocross on Sunday that had a very disconcerting quirk that I have never experienced before. When threshold braking from high RPMs in second gear, the engine would seem to "stall" or cut out at a certain point as the revs dropped. The exhaust note would change from a rapidly lowering "burble" to more of a "hum", as if the engine had stopped firing, and I could feel the rate of deceleration increase as the dead engine pulled the rear wheel rotation down another notch--as if I had suddenly dialed in more rear brake bias.

I talked to one of the local mechanics who told me that this is normal for a DME engine--that there is a fuel cut-off at 1700 RPM on rapid decel. Is that correct? How is that signalled to the engine and why is it necessary? For economy, or to prevent fuel overrun or something? That may be fine for a street car, but is there a work-around to prevent this in a race car setup? You can actually hear this happening when just free-revving the car at a standstill--the engine note changes at a certain point during the decel cycle, and the revs will drop more rapidly, to below normal idle speed, and then bounce back up to idle as the fuel comes back on. Is this fuel cut contained in the chip programming (fuel maps) or is it controlled by the manifold pressure/RPM and a microswitch that shuts off the fuel pump or something? I know little or nothing about Bosch DME functions, as I have never owned an engine like this.

When this happened on course, if I was trailbraking into a corner, it was sufficient to upset the car by breaking the rear tires loose, and I would have instant oversteer. I basically had to drive around this by abandoning my normal trailbraking and braking earlier, in a straight line, and being back on the throttle slightly before turning in, to make sure that it wouldn't happen on my way to the apex. Even braking in a straight line, some cornerworkers told me they could see the rear wheels locking up momentarily at times. The car has stock brakes (S calipers in front, M in the rear) so this added rear bias is coming from the engine drag when it cuts out. The engine is in a lightened '74 RS-bodied car running 245F/275R Kumho Victoracers, so this is not a stock Carrera, BTW.

Thanks in advance for any explanations,

TT

I would say it is probably in the fuel maps for two main reasons fuel economy and emissions. You could probably have it tuned out of the software with the right chip programming.

They cut fuel ABOVE ~1700 rpms on closed throttle. I don't think that's limited to the euro models.

Thanks for the replies. I guess I misunderstood what was said about the cutoff point. So anytime the butterfly closes and the engine is above 1700 RPM the DME cuts fuel off? How fast does that happen? By what mechanism, exactly? When the butterfly closes in the throttle body, does the DME just pull fuel out of the map, or do the injectors stop firing completely at some point? It seems to me that you still need a certain amount of fuel to keep the engine firing, even with little air coming in, and that this would decrease with engine speed until you got to the idle map. Something weirder is happening with this engine--it is like it is not getting enough fuel to run at some point.

I wasn't looking at the tach at all on this busy autox course, just going by ear, but when I ran this engine to 6,000 or so in second gear and then lifted and went to heavy braking, it would decelerate normally for the first few seconds in the brake zone, and then seem to die when the RPMs were somewhere below 3,000, I'm guessing, and not always consistently in the same range, I don't think. This is when the sound changed and the extra engine compression braking would start. When I got back on the gas, it would pick right up again. It was almost like someone turned off the key. If I was already turning in when it happened, it made for very unpredictable handling, to say the least.

TT

It might not cut the fuel totally off but lengthen the injector pulse. Are you sure that the airflow sensor just didn't have a bad spot on it. It would be interesting to figure which input is used for the Idle Air Control valve if it is the micro switch for closed throttle or the TPS to show throttle plate position.If can figure that out you might be able to figure away to trick the DME not to do the fuel cutoff.

On decel (idle microswitch activated) fuel injectors are shut off until the rpms drop down to about 1050, when the injectors come back on. This is for fuel economy and emissions. The turn on point can be altered, to eliminated altogether in the programming. Turning it off is good for a more rapid warm up of the engine in cold weather situations, or in racing for slightly improved throttle response. The drawback is a 10% reduction in fuel efficiency.

HI TT
some paperwork for you to check on your car and the DME unit, which may help your car for the track days.

http://forums.pelicanparts.com/uploa...1194963763.jpg

http://forums.pelicanparts.com/uploa...1194963790.jpg

http://forums.pelicanparts.com/uploa...1194963818.jpg

http://forums.pelicanparts.com/uploa...1194963845.jpg

for your educational purpose.

regards mike

HI just found this also

http://forums.pelicanparts.com/uploa...1194964416.jpg

this info' is for "euro" cars.

regards mike

Thanks, Steve. Since it's mostly a race car, it can live with less fuel economy. What happens if you do not allow the idle microswitch to activate at all? Can it just be removed, or its position adjusted outside of the activation range, or does this wreak havoc with other things, like a smooth idle?

Having talked with the owner about it some more, I have additional info about the engine. It is not a stock 3.2--it has been modified to a 3.4 with JE pistons (10.3:1), with remachined and plated 98mm barrels by EBS, and 964 cams. The pistons are domed with machined valve clearance reliefs, as the plan was to eventually put carbs and "S" cams on it, but for the time being, the stock induction was retained. It spent a lot of time on the dyno at a local EFI shop having a custom chip programmed for it, but I don't think the kind of "real-world" race usage I experienced was anticipated or accounted for. I can't help but think it is going too lean on decel to fire the mixture at some point, and it is going dead at times--at least that is what it feels like.

My only experience with EFI has been an SDS system on a Honda S2000 engine, so I know little to nothing about the Bosch Motronics. All my 911s have been either CIS or carbs so far. The SDS system used a TPS and MAP sensor to control mixture, along with a crank position sensor for timing--there was no idle microswitch. Why does the Motronics use this?

There is no 02 sensor on this car, but there is a port for one in the headers, so I am thinking of hooking up my LM-1 wideband A/F sensor to it to see what exactly is happening. How lean should the mixture be going on decel? What A/F ratio would be too high to support combustion anymore?

The owner wants me to co-drive the car with him, but this particular quirk makes the handling so evil it is impossible to really attack the course as I am used to doing, and it is not as much fun for me as my own cars.

Thx,
TT

Thanks, Mike. This engine started as a type 930/25, but the DME box has this part number on it, which I don't see in the images you posted: "911 618 111 20 RX". It says "Remanufactured" on the sticker. Since the chip has been re-programmed, the part number may not be very useful in telling me anything about the mapping though, no?

TT

If you disconnect the idle switch, you'll achieve what you want, but idle will be unstable to high, and possibly hunt, especially when you cold start the car.

The mixture on decel with the idle switch activated will be basically nonexistent, as there is no fuel being burnt. The LM-1 will read open air as the engine is now effectively just pumping air through the exhaust system.

The DME numbers indicate that it is a brain from a 87-89 Carrera U.S car. The RX is not OEM, and was probably added by a DME repair place. Since it was remapped as you said, this won't tell you anything about how it was mapped. The only way to tell is to dump the code and read it to determine the mapping.

I will try that, then. Since it is only a problem when the car is driven at the limit, it seems like you could leave it plugged in for warming up and daily driving and just unplug it for autox runs or track sessions, then, since idle quality is unimportant for those limited times?

I am having trouble understanding why this is a good idea, other than for economy. If the fuel is cut completely, isn't there quite a bit more engine braking being done at that point? If there is no combustion at all in the cylinder at TDC to help push the piston back down in its stroke, then the full force of friction and compression in the rotating and reciprocating parts will slow the engine (and the car) much faster, no? Especially with a tight, higher compression motor? What is weird is that this engine does not seem to "shut off" right away when the throttle closes at high RPM--there is quite a delay from the initial lift until it does the "extra compression braking" thing, as I tried to describe above.

OK, then it is not even a Euro DME, as I originally thought. Aren't the US spec cars supposed to run with an 02 sensor? Is this to stabilize the idle, or just to avoid ruining the cat. converter?

TT

You can try it - I never did so I can't tell you exactly. The proper way really is just to turn off the function in the chip software.

The fuel shuts off the instant the idle switch is activated. Perhaps there's something causing the switch to hang momentarily until it activates. All fuel injected cars are programmed this way to save fuel and emissions. I suppose it can work for or against you, depending on how you drive. Other guys will complain that not having the fuel cut off will cause them to have to brake harder on the track, and not allow their car to rotate the back end as before on throttle lift.

The O2 sensor is mostly to keep the mixture at low and mid rpm part throttle, where most people are 90% of the time, within the stoichiometric ratio of Lambda=1 where the cat is most efficient at reducing the three main exhaust emissions components, HCs, CO, and NOx. If it's not connected, the car's stock chip is programmed to be as close to Lambda=1 as possible during low and mid part throttle, but if something is off, such as an air leak, it will know know how much to compensate. It has nothing to do with idle.

That must be what's wrong, then--either the flapper is not closing all the way to hit the switch when lifting off the throttle at high RPM, or the switch itself is not activating instantly to cut the fuel as it should. This delay is what's causing the problem in driveability--there shouldn't be a change in the decel rate like I experienced if the fuel was cut instantly, and the engine compression portion of the braking rate would be constant. I guess it could be a throttle adjustment issue, perhaps, with too tight of a linkage not allowing the flapper to close all the way, or maybe it is binding at some point? Or else the idle switch is defective or out of adjustment? Something like that could explain the inconsistency of the effect, as well, given the differences I felt in various corners, depending on the speed, camber and the bumpiness of the surface. Chassis flex and engine movement could be contributing to a linkage or adjustment problem.

I will have to inspect the throttle body and linkage action and then put the A/F meter on it to see when it goes totally lean in the decel cycle.

Thanks for all the help. I've certainly learned a thing or two.

TT