3.6 liter conversion stalling at low speeds

[Bill Verburg]

Oh and on the '95 I still need to replace the head temp connector(when I find what out one it is )

it's not the one that is supposed to be on a '95 993

[Jim Williams]

The stalling issue comes up many times when transplanting the 964 engines into a earlier car, and replacing the dual mass flywheel with a lightened one. When researching the 964 engine stall problem, I found that there are three connections from the climate control panel on the dash to the DME. I think I uncovered a similar scheme on the '84 - '89 Carrera, but much simpler (1 signal). Even though the info is a little old, I thought I would share it. I don't know how much of this might be applicable to the 993 engines.

All three of these inputs from the 964 climate control box to the DME are ground signals. I wired a ground via a switch to each respective DME pin, and tried all the combinations of switch closures in a effort to see if there were any resulting problems. I found none, and left all the three pins to the DME listed below grounded. However, here is a formal disclaimer: DISCLAIMER -- the author is not responsible for any damages resulting from these modifications. (BTW, the engine is still running well for it's new owner after 2 years.)

Did it work? Even with the Cyntex chip which was supposed to help the problem, the stalling still occurred before the wiring mods. After, I had no more problems. YMMV.

That said, here is what I think happens (somewhat long):

1) Pin K 23 on the climate control sends a ground to pin 41 on the 964 DME when the controls are set to enable the Air Conditioning function (also enables relay for slow speed condensor blower).

2) Pin K 13 on the climate control sends a ground to pin 50 on the 964 DME when the Thermostat (temperature control knob) decides the compressor should be running.

3) Pin K 24 on the climate control sends a ground to pin 40 on the DME when the A/C compressor relay is energized.

From information I discovered in the Charles Probst book -- "Bosch Fuel Injection and Engine Management" -- some DMEs (the 964 maybe?) had the capability to slightly delay the application of the signal to turn on the A/C compressor for a 1/3 second until the idle positioner had opened up some amount to anticipate the upcoming load which the compressor was about to place on the engine. This was apparently enough to ward off an engine stall situation. Of course the SC has no such provision with the HVAC controller, plus the A/C compressor on the SC is not controlled by a relay.

How would this work on the C2?

Assume that no controls on the climate control unit are initially set. DME operation is not altered, which it doesn't need to be since the flywheel nice and heavy. (I think the double mass flywheel on my engine weighed 43 lbs.)

The temperature knob, if turned CCW below cabin ambient, would signal the compressor to turn on, IF the A/C switch were turned on. The DME can know both these facts as there is a connection there to accomplish this. (Pin 50 for the temp control, pin 41 for the A/C control.) The compressor could be turned on by the DME (pin 40) OR by the HVAC control box (pin K24).

Scenario 1: (engine at idle, A/C not on)

a. temp knob calls for A/C compressor = ground on pin 50
b. A/C switch turned on = air conditioning mode = ground on pin 41

Result = idle positioner gets a slight bump up, A/C compressor turned on after 1/3 second delay by either the DME or the HVAC box.

Scenario 2: (A/C compressor on, engine not at idle)

a. Accelerator pedal released, clutch in, engine suddenly isolated from inertia of transmission/road wheels.

b. As soon as DME knows compressor is running (ground on pin 40) and idle has been called for by idle contacts on throttle body, idle positioner "coasts" engine down to idle speed.

c. If Temp knob was calling for air, but the Evaporator sensor had shut the compressor off because this temp had been reached, pin 50 could still be ground but compressor could be off. If compressor was off, there would be no need for the positioner coast. The voltage at DME pin 40 would reveal that the compressor was off.

That's my take on it. I could get no deeper as no one I talked to about the DME internals knew any more, or if they did, they weren't telling. I talked to chip builders, race car shops, you name it, I probably talked to them. However, if anyone has info on the DME internals, please share.

[Bill Verburg]

Just as an aid I think that these are the pins that Jim is discussing


Jim, Please let me know if I am wrong

[Jim Williams]

Bill,

Yep, that's them. Thanks for posting the connector pin layout. Is that the 993 DME connector? Looks just as I remember the 964 connector.

[einreb]

Jim,

Thanks for the post. Thats great info.

I have a 964 3.6 in my '88. Ive gotten the stalling sorted out (it still is a little sensetive with the lightweight rs clutch/flywheel package, but does not actually die), but have yet to fully reinstall my A/C. I'll be curious to see how the engine acts under the load of the compressor.

I seem to recall reading that faking out that the A/C was on was how the chips handled the lightweight flywheel issue. I'm not suprised you got such limited info from the chip makers. They seem to hide the data quite well. It must be a good business.... burn a chip, charge $300-$600.

Thanks again,

Bernie

[Kevin Stewart]

is your car smoking? Kevin

[Joeaksa]

Excellent info guys!

JoeA

[Lorenfb]

"The stock program cuts fuel at idle." - Moses -

Not true! The fuel cutoff only occurs during decel with the idle switch closed
AND the RPM > than about 1200 - 1300 RPMs.

"a custom chip like the cyntex chip offered by instant g will solve this problem as it takes the lightened flywheel into account." - Moses -

Cyntex and others make many claims related to idle undershoot which I and my many customers
have yet to see!

"I really think that the stalling issue is more related to ISV sticking or malfunction or vacuum leaks or electrical faults than any thing else.

I have made both of the ISVs adjustable

on the 3.8 i can adjust it to idle well cold buut then it idles fast when hot or idle well hot then stall when cold" - Bill Verburg -

1. The stalling issue when using a LWF is VERY common and is NOT related to
ISV problems.
2. When modifying the ISV to be adjustable, you then have to deal with cold/hot idle problems
as mentioned above.

"1) Pin K 23 on the climate control sends a ground to pin 41 on the 964 DME when the controls are set to enable the Air Conditioning function (also enables relay for slow speed condensor blower).

2) Pin K 13 on the climate control sends a ground to pin 50 on the 964 DME when the Thermostat (temperature control knob) decides the compressor should be running.

3) Pin K 24 on the climate control sends a ground to pin 40 on the DME when the A/C compressor relay is energized. "

- Jim Williams -

Use of these DME inputs to solve the idle undershoot on a 964 with LWF justs
masks the problem by basically always forcing the idle to a higher value even
when NOT de-accelerating.

For further clarity:

pin 41 - increase the idle, because the blower is on causing an increased current
draw on the alternator (increased load) reducing the idle

pin 40 - increase the idle, because the A/C switch has been depressed causing the
compressor to be on all the time resulting in an increased load reducing the idle

pin 50 - increase the idle, because the climate control system is in the temp regulation
mode CYCLING the A/C compressor on/off

The real solution is to modify the DME idle algorithm to allow a slower decel for the LWF.
It's not a difficult solution.

[Jim Williams]

Lorenfb writes:

Use of these DME inputs to solve the idle undershoot on a 964 with LWF justs
masks the problem by basically always forcing the idle to a higher value even
when NOT de-accelerating.

Although it may not be an optimal solution, use of the DME inputs doesn't mask a problem, it deals with one. It helps prevent the idle stalling. Also, it does nothing to force the idle to a higher value when not decelerating.

For further clarity:

pin 41 - increase the idle, because the blower is on causing an increased current
draw on the alternator (increased load) reducing the idle

pin 40 - increase the idle, because the A/C switch has been depressed causing the
compressor to be on all the time resulting in an increased load reducing the idle

pin 50 - increase the idle, because the climate control system is in the temp regulation
mode CYCLING the A/C compressor on/off

Your explanations would be a little more accurate if you had stated "increase the idle *air*". The idle *speed* or RPM, is not increased. At idle, the ISV is in closed loop with the flywheel sensor still attempting to keep the idle RPM constant at 880. More air may be flowing through the ISV to cope with the added engine load, but the idle *speed* remains at 880.

I stated only what signals the climate control box sends to the DME, not what the DME does in response to one of these single inputs. Your statements oversimplify what goes on when the inputs are received at the DME.

For pin 41: this signal by itself only indicates to the DME that the low speed for the condensor blower is enabled.

For pin 40: depressing the A/C switch, on it's own, does not cause either the blower or the compressor to come on. Other inputs from the climate control are required to happen in parallel with this event. For example, if this signal is present, and the temperature control knob signal is not present, the A/C compressor is not on, and there is no added load on the engine.

For pin 50: this signal by itself only tells the DME only that the temperature knob at the climate control is set such that the cabin temperature is higher than the dial setting. This signal alone will not turn on the compressor. It is sent to the DME even when the A/C switch is off.

The grounding of these pins do not increase the idle *speed* as might be implied from what you state.

1) Unless the *accelerator pedal* is *at idle*, I think the inputs do nothing.
2) The inputs affect the idle *air* in a proactive fashion to reflect the added engine loads *before* the idle drops too low to recover. They don't affect the idle *speed*.

The real solution is to modify the DME idle algorithm to allow a slower decel for the LWF.

A statement of the obvious.

It's not a difficult solution.

Maybe it's not a difficult solution for a software person who understands engine dynamics and has access to the DME block diagrams. Pretty difficult solution for the run-of-the-mill hobbyist who has transplanted a 3.6 liter engine with a LFW into his car.

In summary, I did my posting on this thread at the request of others, not to get in a position of defending of how much I know (or don't know) about DME functions. It is plain that I am not alone with my lack of detail knowledge of what goes on inside the DME to deal with the A/C issue. I only know what worked for me, and that it didn't upset the normal idle operation. The reader can make up his/her own mind if they want to try this out.

[Bill Verburg]

Jim, Thanks!

Is that the 993 DME connector?

No, that is a 964 connector.

Any other thoughts on things that affect idle, other than a stuck ISV?

Is there an electrical test, that can be done to check the function of the ISV?

[Jim Williams]

Bill,

Alas, as I stated in my last post, I could find out little about the guts of the DME. The closest I found was a PDF file that shows the internals of a Motronic DME for the BMW. This I think is a close relative of the 964 DME, about the same time frame; it had a 1.3 number. I'm not sure what else the ISV relies on for position other than the flywheel speed sensor.

The Bosch "Hammer" was able to test the ISV, but there was another test that required a duty cycle meter where you could monitor the duty cycle, and load the engine, such as with an added electrical load (or mechanical load such as switching on the A/C!) and checking to see that the duty cycle changed while the RPM stayed (more or less) constant.

There was also another test for a faulty ISV that required the use of an o'scope to monitor the waveform to the ISV. It is a 100 Hz signal, almost a square wave (with no specified peak-to-peak voltage), that wouldn't change when the ISV was unplugged if the ISV was "faulty" (open?). This was just an "ignition on" test, engine not running.

I had another thought for a test, not checked out BTW. I don't remember exactly how hard the components involved in this might be to get to, and no longer have a handy engine to go check. If one could short out the idle switch (if it was normally closed, don't remember), or open it by removing a connector if it was normally open, to fool the DME into thinking the throttle was remaining closed, then just cracking open the throttle to let a little air through, then watching the duty cycle to the ISV to see if the ISV tried to close up a little, this might also give a hint if the ISV was working. If the ISV was already at its mechanical stop, this wouldn't tell anything.

Maybe someone else would have some other thoughts.

HTH

[Lorenfb]

"I only know what worked for me, and that it didn't upset the normal idle operation." - Jim Williams -

That mod will now AFFECT how the idle functions with the A/C "on" AND
the throttle decel with a LWF, i.e. it won't function dynamically as the stock
964 does, e.g. hunting. Besides, it's all a function of what LWF is used,
e.g. some LWFs (Andial & et al) have different affects.

"The closest I found was a PDF file that shows the internals of a Motronic DME for the BMW. This I think is a close relative of the 964 DME, about the same time frame; it had a 1.3 number. " - Jim Williams -

The BMW and Porsche (964) DMEs are totally different!

[Bill Verburg]

Jim, Thanks! there is some additional info here in a Rennlist thread.

It's very discouraging dealing w/ the lack of info available. i have considered a Motec or DTA system just because of this sort of nonsense. My 964 was perfect, both 993s have idle issues, the ISV is mechanicaly clean and moves freely but the idle is either low or high on both, no stalling just out of spec.

I may just buy a new one just to see what effect it may have.

[Jim Williams]

Bill,

Thanks for the link. Even though my 3.6 SC has gone on to new owner, it helps to review some of my old notes occasionally so I don't forget about my experiences with different Porsche models.

It's also reaffirming that most posts are on a positive note in this forum environment in the face of a few negative "contributions". Instead of trying to help, there always seem to be a few posts that seem to only find fault with, or "one-up" those of others.

[autobonrun]

My 96 993 has the LWF but does not stall. The idle does drop quickly and very low. The idle is so low it feels like it will stall but never has. I understand some have fixed this problem by fooling the engine management system into thinking the A/C is on all the time. I understand that if the A/C is on, the engine idles higher.

Try cutting the A/C on and see if the engine does indeed idle higher, to the point where it doesn't stall. If so, you might try inputting a permanent signal that the A/C is on. It won't affect the A/C but may help the stalling problem.

I don't know the mechanics of where this input is brought in but if this experiment works, perhaps others are familiar with this trick can chime in.

[GarryW]

Jim, thanks for your input and fading some heat. I, like so many am dealing with the same problem. My knowledge and experience is minimal so for me and perhaps others, please detail exactly how you make these grounded connections to the DME. I have a 91 3.6 transplant and after installing the NBD chip, although better, I am still having stalling issues on deceleration..........Garry

[Jim Williams]

Autobonrun,

The engine won't idle higher with the air on. The DME adjusts the idle air to compensate for added engine loads, but the idle speed stays the same.

To compare the 993 engine in a 993 car, with the 993 engine in an SC, for example -- the first has a heavy flywheel and the climate control panel that talks to the DME. The second (in most cases) has a lightened flywheel, and no DME connection to a climate control box. The second gets a double whammy. When the stock 993 gets a LFW, there is still the climate control-to-DME interface to deal with the A/C problem. I don't know of any way to fake the "A/C On" without having the A/C actually on. So in order to the A/C On trick, you would need to disable your normal A/C function, seems to me....

Garry,

I'm sharing this mod on this thread for the first time since I did it on a 3.6 transplant somewhere around 1998. So I'm having to rely on my notes alone; I don't have the car any longer. -- Just to put all this in context.

Bill Verburg posted a scan of the connector to the DME box. I'm assuming that the DME resides under your driver's seat where it is in the stock installation. You would pull the connector from the DME and remove the cover from the connector. My memory is hazy on exactly what this entails, but it should be fairly obvious. You would use the connector diagram and find the DME pins mentioned in my earlier posts, 40, 41, and 50. These pins will probably already have wires on them, but I have no way of knowing if that is the case with the harness that came with your engine. If they have wires on the them they will either end in another connector on that harness or they could have been cut. If you got the complete harness, you will have a 14-pin connector on the harness near the DME connector. Pin 7 will be the connection to pin 41 of the DME. Pin 9 will be the connection to pin 40 of the DME, and pin 13 will be the connection to pin 50 of the DME. If you already have the 14-pin connector, you can ignore taking off the cover of the 55-pin connector, and just verify the connections with an ohmmeter.

You will need to find some extra pins that will fit the mating connector to the 14-pin connector of the DME harness, or find another way to make a connection to these wires which are then simply grounded to a handy ground connection under the seat. I initially put a simple switch in series with each of these 3 leads, and did all sorts of combinations with closing them with the engine running to see if any problems resulted. I didn't find any.

I then left all the switches closed and did some test runs in the same conditions that had caused stalling problems when running the air as described earlier in this thread. I recall this as taking care of my stalling problems.

Be advised that you need to be careful when chosing the proper pins on the connector. I don't want to be responsible for anyone trashing their Motronic box.

[autobonrun]

Jim, thanks for the clarification. I'll have to ask the Porsche mechanic who told me this to explain the trick to me again. I know he said he could fool the car into thinking it was under the additional load; but I don't recall the particulars. And, as you rightly point out, his "fix" was for a 993, not a 993 engine in another vehicle.

[GarryW]

Thanks Jim, for the explanation. I do have a complete engine harness and with that I think I can get it done.