Not really fixed

[corsachili]

When we last left our hapless hedonist, he had figured out that his beloved CSL's long suffering malady was, as he always thought, electrical in nature. It seemed that it was related to the headlight switch as turning the headlights on would cause the problem to go away. He hoped against hope that it was the switch itself, but as it turns out, that's not the case.

I've done lots of testing, cleaning, soldering and thinking over the past few days. Some have opined that the problem is a poor ground somewhere and so I've gone to great lengths to make sure all my grounds are clean, making good contact, that the connectors are tight, and that the wiring is healthy. This includes, but is not limited to, the ground wire on the fuel pump relay under the right rear seat, the three ground wires connected to the fuel sender unit, the ground in the trunk, and the engine block to chassis ground at the rear of the engine to the firewall tab. I also have a massive ground wire (4 gauge) which connects to where the battery ground would normally go and goes back to the aforementioned firewall tab. The ground for the battery is 0 gauge welding cable which bolts solidly to the rear seatbelt boss on the parcel tray. Positive power is carried by another piece of 0 gauge welding cable which goes directly to a binding post in the engine compartment. I've cleaned and checked the connections at the coil as well.

A friend of mine things the problem is too much voltage, which I'm doubtful of, especially after measuring only about 13.5 volts with the engine running. He thinks the Optima battery is producing too much voltage. Again, I don't buy it. Carl Nelson says D-Jet only starts having problems in the 14-15v range, and I'm not close to that, especially at idle.

We tested removing each headlight and trying to determine if turning the lights on would still fix the problem. Then we went directly from the binding post to the headlamp (circumventing the wiring harness) and still the additional draw of the headlamp would fix the problem. Then we went directly to the battery, and this time connecting the headlamp did not fix the problem. At this point however, the problem went away and we could not get it to come back.

There is yet another running problem, perhaps related, perhaps not. It feels very much like the original problem I had, and occurs at partial throttle openings. For instance, driving slowly in second gear the car has an off idle misfire/stumble that clears up as soon as you accelerate. It happens at speed too, with a miss or too lean feeling sort of stumble that may just be a manifestation of this problem which at idle feels like the car has a lumpy cam in it.

I am getting closer. It's frustrating but it seems that maybe someday I will figure the problem out. I am going to check the back of the fusebox and make sure I've got good connections there as well.

Maybe with a little luck the car will actually run properly.

 

[Malc]

More things to throw out.... Some whacky, Some you may of checked already.....
1]...Is the Throttle Position Switch set correctly, is it working??
2]...Is the "butterfly" seated correctly in the throttle body
3]...Throttle body Air bypass passage clear
4]...No air leaks in the cold start aux air intake system?
5]...Brake servo one way valve ok
6]...Fuel pump relay good? (next to ECU), ECU earth (ground) good
7]... Harness from ECU to Engine - all wires good, resistance on each line should be very low <0.1Ohm
8]...Throtttle body engine vapour return clean, pipe clear and not split
9]...All injectors, 2.2 Ohms
10]... Altenator voltage regulator ok.

If I think of anything else.....
Malc

 

[ScottAndrews]

Do you have a scope?
If not , you can borrow mine. I'll be home this weekend, lik eall other weekends.. stripping paint... :-( (man am I sick if this!)..THe scope is covered in soda dust.. so you will hae to clean it..Come to think of it..everything in my garage is covered in soda dust..

You can check the injector pulse form and the ignition pulses at the coil primary. At least you can see if there is some weirdness going on there.

I assume you have checked the fuel preesure. Have you tried slightly squeezing the return line (don't clamp it tight or you may burst a line and torch the coupe!). That would boost the fuel slightly at idle.

Also, a while back I suggested that you add some resistance to the cooalnt sensor line. I am assuming the D-Jet uses a similar temp compensation for the mixture. By adding resistance (I'd use a 5K ohm pot) you can fine tune the mixture and at least put the lean misfire question to bed (or maybe solve the problem). I just put the pot up at the ECU. I adjust it seasonally for optimum running.

 

[corsachili]

Hi Scott,

I don't have a scope.

I have checked fuel pressure.........29PSI. I bumped it up to 34 per Carl's direction.

I have tried squeezing the return line (effectively increasing the fuel pressure). The problem goes away immediately. We've just never figured out what the problem was that doing so solved.

Terry and I are going to take the engine, gearbox and driveshaft (and exhaust) out of his CS this weekend. Maybe we'll stop by in the CSL on Saturday and say "hi".

Malc-TPS is new, and I've played with it but I don't know for sure that it's set properly. I know it's supposed to be off at idle and off at max RPM.

Butterfly seated in body? Never really checked, how could it be otherwise?

Throttle body air bypass passage clear? Again, never checked, but I don't think it would idle if it weren't.

Aux air intake system has been checked for leaks.

Brake servo check valve has been checked.

Have not done any of the other things you mention, but will do.

Thanks for the advice!

 

[velocewest]

1. Remove all d-jet related equipment and place in a box.

2. Install a set of Zeniths, or as a last resort, Webers.

3. Get your ass busy on the S38.

 

[x_atlas0]

1. Remove all d-jet related equipment and place in a box.

2. Install a set of Zeniths, or as a last resort, Webers.

3. Get your ass busy on the S38.

Why not go Motronic. Cheaper than the Zeniths, and the S38 harness will meld much easier with an existing 1.3 setup.

 

[AndyM]

Have you checked your voltage at the coil when running? How 'bout when running poorly? What coil/resistor/ignition combo are you running and how is it wired?

The phantom headlight switch may not be the problem, but it suggests your ignition is very sensitive to any additional current draw. I've suggested this before, but just in case -- some coupes have a resistor wire in line to the coil in place of an external ballast resistor. If you use this wire to run a pertronix, crane, etc. -- or if you hook up a coil with an internal resistor (Bosch blue coil, for instance), you will get the symptoms you describe. The electronic ignition and internal resistor coil need a full 12v to operate properly and should not be used with an external resistor. If they are used with a resistor, the additional draw from the headlights will cause a miss. Also, if you use a coil intended to have an external ballast resistor without one, the coil will act in a similar manner (and will get noticeably hot to the touch).

 

[corsachili]

Hey Scott,

Terry and I are going to pull the engine out of his CSA this weekend. My CSL will be at his place, so maybe we'll come pay you a visit on Saturday. I'll contact you off list to discuss.

 

[ScottAndrews]

If squeezing the fuel line makes the problem go away, then I suspect the injectors are not opening long enough. Squeezing the return line will jack the fuel pressure way up. If the stumbling stps then the issue is that there is not enough fuel (the lean condition that I have been harping about).

It seems, hoowever,t hat the cause is not the fuel system, but the ECU or the injectors themselves. Since you changed outt he ECU to no improvement, then I suspect the injectors.

Are you certain they are the correct impedance?

Different injectors require different pulse shapes.

S

 

[ScottAndrews]

TJ;

Do you guys have a hoist and engine stand? I have both currently sitting idle

S

 

[corsachili]

Scott,

I just had all 6 injectors cleaned, tested and bench flowed. They do have different part numbers but after a bit of research I thought they were the same as the 043 part number injectors.

Thanks for the offer, I do have a stand and hoist.

 

[ScottAndrews]

Yeah, I just realized that you were doing this at YOUR place, not at Terry's. I didn't recal seeing the engine stuff in his garage.

There seem to be two basic types of injectors: Peak and Hold, and Saturated. From what I gather, Peak and Hold are faster, but they require a more complex pulse shape to operate. The L-Jet and Motronic injectrors are Peak and Hold.

I am uncertain which type the D-Jet uses. The D-Jet was one of the first EFI systems, so it might be the case that it used Saturated injectors. If this is the case, then running P&H injectors woul dwork, but the system would always be a little bit lean because the injectors woudl a) open slowly and b) would be getting this weird pulse.

This is a quote from Probst:
"In early L-Jetronics (Until about 1978) each injector circuit includes a resistor to limit the current flow. In these circuits, the initial current flow must be great enough to rapidly open the solenoid/needle valve, and then flatten out when only a holding current is needed. Since about 1979, L-Jetronics regulate the current in the control-unit, and so do not need separate series resistors. The resistance is buil tinto the injector itself, using a solenoid winding of brass (higher resistance than the previously used copper winding). This also requires less power than the series resistor type".

I have so far been unable to ID the 043 injector. The later L-Jet used a different number, and by far the largest numbe rof D-Jet systems were Porsche and VW, most with smaller engines.

So, a number of things may be going on here. The CSL presumably used the highest performance injectors available. It is concievable that your "compatible" injectors are not actually compatible . They may have too low a flow rate, OR they may have the wrong electrical characteristic.

Theory 1:
For example, if the CSL ECU is different from the CSi ECU, then it is likely that the CSL unit was driving the lower current brass coil injector. If this is the case, then the ECU is not equipped to drive the higher current injector, so it is current limited. This theory is counter to the observed headlamp fix. When the lamps are on the system improves, so it seems you are supplying TOO MUCH current.. not too little

Theory 2:
The CSL is designed to run saturated injectors, and you have P&H injectors. In this case the injectors are not opening fast enough, and the engine is fuel starved. It is unclear what the impact of lower system voltage wouldhave on this. Seems wrong.

Thoery 3: The CSL is designed to run higher current copper coil P&H injectors. The ones you have are later model brass coil P&H injectors. The higher resistance attenuates the initial current spike, and the injector does not open as fast or opens late. THis theory maybe wrong, but the implication is this: The injector drive circuit is "tuned" to run a specific type of injector. If you run an in jector that looks electrically different, then the system may sort of work, but not exactly as planned. The result is that the car wiull run, but it will either run rich or lean, depending on how the electrical mismatch occurs.

I now suspect your injectors. Do you have any detail on the original CSL injectors?

If you bring the CSL by, maybe we can put the scope on the injector lines and examine the pulse shape. If it is normal then I am wrong...

Some additional links:

http://www.sdsefi.com/injectors.htm


http://www.rbracing-rsr.com/fuelinjectors.htm

 

[AndyM]

Slow down a second. Good theories, but I think you can rule out most of it.

First, peak and hold are low-impedance injectors. The D-Jet used these. All D-Jets -- all flavors. Saturated are high-impedance injectors used on the Motronic systems and many other newer EFI systems. Your Probst cite seems to indicate that L-Jet may have used both low-impedance and high-impedance injectors at various times, but I haven't played around with those enough to confirm. The reference to a resistor in the injector circuit would be used with low-impedance injectors. The ECU control and brass windings (without resistors) suggests a change to high-impedance injectors.

You are correct that you can't mix and match the two. If you use low-impedance injectors with a high-impedance ECU (like a Motronic system), you'll fry the ECU injector drivers since it takes too much current to open the injectors. Likewise, you'd be lucky to get the motor to run using a low-impedance ECU with high-impedance injectors.

Anyway, I don't believe there are any significant differences to the D-jet parts fitted on the CSi vs. the CSiL. ETK confirms the same part numbers for the injectors, the ECU and the manifold pressure sensor. Thus, I think its safe to assume that it should have low-impedance (peak and hold) injectors installed.

The stock 036 injectors measure 2.2 ohms across the terminals with a digital voltmeter. If TJ's are similar, then they are the correct low-impedance injectors (though the flow-rate may not be the same). If they measure in the 10-16 ohm range, then they are the incorrect high-impedance (saturated) type and should be swapped out. Even if the injectors currently installed are the correct low-impedance versions, the flow rates could be signficiantly different which could have a major impact on driveability.

I still think the problem lies in the electrical system somewhere. Putting the headlights on and off shouldn't have any effect on the problem unless the ignition or injection wiring is screwed up somewhere. The intermittent symptoms suggest electrical too. If the injectors and ECU are mismatched, the problem should exist at all times. Clamping the fuel return line just made it run a little richer, but that does not necessarily mean its too lean to begin with.

I'm still curious about the ballast resistor/coil/ignition set up. I'd also like to know whether the MPS has been swapped for a known good one -- the MPS and the coolant temp sensor have the largest impact on the mixture at various RPM/MAP points. Do you guys have access to a WBO2? That and a scope would really help.

 

[ScottAndrews]

You are welcome to borrow the scope any time. It's a decent Tektronix portable.

I wish I had a WBO2 meter...Maybe if you buy one I can borrow it ;-)

My ruminations on the injectors were based on the two observations:

1) It improves if richened, so the misfire is probably not ignition related
2) It is affected by the headlamp current draw, whihc means it is somewhat electrically related (just not ignition)

So what can cause low RPM lean running?

a) Aging in the ECU components might change the current source to the temp sensor, and thus throw off the nominal operating point
b) The injectors might be somewhow mismatched to the ECU (although I agree that, in general, the injectors seem to have two basic variables: flow rate and type, and if the type is right, then flow rate should impact you all across the RPM range, since the amount of air going into the cylinder per cycle doesn't change appreciably with RPM).

Have you measured the voltage on the temp sensor with and without the headlamps on? Have you tried the simple fix of placing a 5 K ohm pot in the temp line? The fatory actually did this as a Tech Service item on the 635 inthe mid 80's to get it to idle properly.

S

 

[AndyM]

I wish I lived closer to you guys. With my WBO2 and Scott's scope, we'd have TJ's coupe running perfectly in no time. I also have a box of known good D-Jet parts off my CSi that I'm not willing to part with, but would certainly lend for troubleshooting purposes.

The low RPM lean condition could also be caused by a misadjusted or malfunctioning Manifold Pressure Sensor. That still wouldn't explain the sensitivity to voltage fluctuations. There's also the mixture knob on the ECU, but that's very limited and should only affect idle.

 

[corsachili]

As usual the information offered by you guys is greatly appreciated. Let me try to answer a few more questions.

Regarding the map sensor, it was tested against known good (by someone whose opinion I trust) and determined to be good.

I have not tried measuring the impedance across the injectors, but I can and will do, just as soon as I get my car back.

I'm not clear about the injectors and whether or not they're right. At one point I thought that they might be the culprit, but then later I was told, on some authority, that they were fine. They are not the original Bosch part number, I do know that.

I still believe that the problem is electrical (as I have all along) based on my experiences. It is amazing how very intermittent it is. You can drive it one time and it will behave perfectly the entire time (as it was yesterday) and then the next time you drive it, it's horrible. Terry drove it home last night and it put on a perfect display of how badly it can run (missing and stuttering subtly the whole way home for him).

 

[Garrett]

You are correct that you can't mix and match the two. If you use low-impedance injectors with a high-impedance ECU (like a Motronic system), you'll fry the ECU injector drivers since it takes too much current to open the injectors.

Andy,

What would be the effect of "frying" the ECU? Motor would run rough, or not at all? I definitely ran the low-impedence injectors on my 3.5L/Motronic 1.1 set-up. But as you say, the car barely idled much less ran, and certainly wasn't driven. I have the correct high impedence injectors now (or at least I think I do, 19#, 15.1 ohm) but the motor runs unevenly and down on power. Could that be explained by a fried ECU?

Garrett

 

[AndyM]

Yes, you have high impedance injectors in there now (19# 15.1 ohm). Those are the stock ones for 1.1 Motronic. You easily could've damaged the driver circuit by running the wrong injectors. I believe (but would need to research it to confirm) that the 1.1 Motronic is batch fired -- meaning that there are two sets of drivers and two injector sequences, i.e., three injectors fire at a time. If one driver circuit is bad, it'll run, but not well. That would easily explain your uneven running.

 

[ScottAndrews]

That woudl be mighty uneven!

All of the Bosch systems run 3 and 3 injector firing.

I would think if one set of 3 cylinders were not getting any gas due to a fried ECU, the car would run so badly that you would not consider it "running". How uneven is it?

More likely the injecotr transistors have cracked solder joints. This is a common malady in the Motronic ECU. You take it out, open it up and reflow the solder joints and you are good to go for another 75K miles or so...

You can tell a lot with a scope on the injector pulses

S

 

[andrey]

ou easily could've damaged the driver circuit by running the wrong injectors.

I doubt it... you can damage Motronic's circuit by running D/L-jet (low Z) injectors (without inline power resistors), but I don't see how it would work the other way around: higher imedence injectors in a low-Z driver will just make the injectors barely work (hmm) but they would do the opposite of increasing current though the drivers...

also, yes, Motronic does batch-fire (even/odds, I think)