Too high injection signal/injection pulse (¿found problem?)

[DaniTD]

Again with my every-day-getting-worse idle in my 318i with m40 engine
After injectors being rebuilt, problem still exists. Engine shakes a lot and feels like missing here and there at idle.
So, I connected the M40 to the BMW DIS and found some strange data. Viewing the live status of the engine sensors, timing, engine and air temperature etc seems ok, except for the injection time. TIS says injection time when warm should be around 2-3ms, but I am getting 4.5ms which seems too high. Load signal is ok at about 2ms.
What can cause the ECU to open so much time the injectors although load and idle speed is ok?
Thank you!

Edit: For the ones that say our cars are too old ;)

[Speedtouch]

Check the blue temperature sensor is working properly as a first step - this tells the ECU when the engine is warmed up, to adjust the injector opening time.

http://www.e30zone.net/e30zonewiki/inde ... emp_Sensor

[reggid]

also make sure the TPS is adjusted and working so that the idle switch is working / activated

[DaniTD]

Hi, thank you for the answers! Coolant temperature sensor is ok as DIS tells the engine is at 86°C, same with air intake temperature at 35°C. TPS is working properly as I can see on Component Status that idle switch is active when idle, and WOT switch active when flooring the accelerator. There must be something else but I don't know what to do a˜ž

[Tommek]

I see you have some oxygen sensor data. Check if it's fluctuating from 200mv to 800mv. I see it's at 400mv in that screenshot and could be stuck, but i may be wrong.

[DaniTD]

The O2 sensor is disconnected as my car has no catalyst and no O2 sensor, so that voltage is a break in the circuit and the ECU uses the CO potentiometer instead

[Tommek]

In that case, a high injector pulse width could be caused by an air leak or a low fuel pressure. Both of them could make the ECU increase injector opening time to compensate for lack of fuel, as long as the ignition system is in good working condition.

I have the same problem with my m40, a serious engine shaking and vibration at idle with a fairly stable idle rpm. Changed injectors with clean ones, checked for intake leaks and ignition leads and cap seem ok without obvious carbon tracking. I cross my fingers i don't have a compression problem because i haven't checked it yet.

[DanThe]

Any AFM data? Its the only other thing that can affect the fuelling really

[Tommek]

AFM can be seen in the screenshot in the LMM Up/Uv data. It's a bit under 0.2 and tehnical data it's between 0.2 and 0.3 at idle so it is a bit on the low side. Maybe a vacuum leak ?

[Brianmoooore]

Don't forget that there's no feedback loops on this primitive management system, other than idle speed. The ECU has no idea what the fuel mixture is, and only sets the pulse width from a preset map.
An air leak would form a type of feedback loop with the AFM, in that the bigger the air leak, the less the AFM output, but this would tend to narrow the pulse, not make it wider, unless the speed loop is confused and intervenes in some way?
An exhaust analysis would help.

[Tommek]

Yeah, Brian is right there is no oxygen sensor to make the computer change injector opening.
He is right, the drop in engine speed caused by a leak and therefore a lean condition could be compensated by a more opening of the icv and increase of injector opening time.

How does it feel at higher loads/engine speeds? Is the miss still there?
Another brainstorming idea would be to check the ignition coil, it may be too weak to deliver a good spark but i think that would be persistent even at higher RPM

[DaniTD]

Thinking on what you've said to this moment, I developed a theory...
As Brian says, because I don't have O2 sensor, Motronic doesn't have any way of knowing if the engine is running properly.

Say at the Factory they tune AFR tables with O2 sensor before leaving the line, even if they don't have the sensor, to make the engine run properly.
Now, say I have a... ¿maybe fuel pressure problem as you noted? Engine starts and Motronic assigns the "Factory" value for the idle (18º KW ignition fixed, 2.0ms ti fixed and so on). But, because of the low fuel pressure, although injectors are opening the time they are supposed to, they don't provide enough fuel, so making the engine running lean and dropping rpm down. Then, as Motronic only have the reference of the programmed idle speed (800rpm), it steadily increases injection time until the idle programmed speed is reached, although, it always will be with the mínimum required quantity of fuel so it makes the engine run at the proper speed (so for the ECU is ok) but running rough and lean (in fact, CO readings are way too low about 0.1% when the specified value is 0.7+-0.5%).

How do you see that? It's a theory...

[Tommek]

Yes, as long as rpm is the primary input to the ECU , when it detects a drop in engine speed caused by fuel pressure it shifts to another cell in the map with a grater pulse for injectors.

As for the AFM in this case the value can only be used for diagnosing an air leak and has no impact on mixture regulation because as far as i know the ECU ignores the AFM at idle and only uses it higher engine speeds (with the use of throttle switch) but i'm not sure about this.

The only impact on idle as i am thinking wold be from ICV, injector opening time and ignition advance.

[reggid]

Thinking on what you've said to this moment, I developed a theory...
As Brian says, because I don't have O2 sensor, Motronic doesn't have any way of knowing if the engine is running properly.

Say at the Factory they tune AFR tables with O2 sensor before leaving the line, even if they don't have the sensor, to make the engine run properly.
Now, say I have a... ¿maybe fuel pressure problem as you noted? Engine starts and Motronic assigns the "Factory" value for the idle (18º KW ignition fixed, 2.0ms ti fixed and so on). But, because of the low fuel pressure, although injectors are opening the time they are supposed to, they don't provide enough fuel, so making the engine running lean and dropping rpm down. Then, as Motronic only have the reference of the programmed idle speed (800rpm), it steadily increases injection time until the idle programmed speed is reached, although, it always will be with the mínimum required quantity of fuel so it makes the engine run at the proper speed (so for the ECU is ok) but running rough and lean (in fact, CO readings are way too low about 0.1% when the specified value is 0.7+-0.5%).

How do you see that? It's a theory...

it takes about 10min to test fuel pressure and fuel delivery. if you dont have a gauge or bentley get one

also how does the car drive otherwise?

[DaniTD]

Driving seems fine although the car always had this problem so I can't say... I tested with a smartphone app and sends that on the top of the rev counter has a lack of power, because although it measures maximum torque almost at the same speed as the original data, the power bands seems to fall after 4700rpm, having an estimated loss of 15-20cv. The power numbers doesn't bother me because of calibration, but the power loss after 4700 does.

[Brianmoooore]

Clapped out injectors will have a similar effect to low fuel pressure, and, if they are the of similar age to those found on the average E30, is more likely than a fuel pressure problem at idle.

[Tommek]

He said that injectors were rebuilt. If you loose power at the top of the rev range it cold be caused by a lack of fuel. Check fuel pressure and see what values you get. I have a vid from when i tested mine. I got about 2.5 bar at idle and 3 bar when i snapped the throttle open which seems good.

[Brianmoooore]

He said that injectors were rebuilt.

Still can't see where he says that, but if true, then discount my last post. Fuel pressure testing kit can be bought for around £10 on ebay, as I've posted on here many times.

[Tommek]

Again with my every-day-getting-worse idle in my 318i with m40 engine
After injectors being rebuilt, problem still exists.

All you need is a t piece a bit of fuel hose and a 3bar gauge. That's how i did mine.

[DaniTD]

Maybe that's the good way. Today I noticed that when idling the fuel pump is making a loud humming that was not there before. You can hear it all from the pump to the filter housing through the fuel lines. And the pump is only 3 Years old. Its not the same sound as when the tank is too low, its louder. Seems like its working bad because of wrong pressure?

[Tommek]

You only know what's going on if you do the pressure test and see the results. Check for fuel pressure increasing as you open the throttle fast or disconnecting the regulator vacuum hose. It should jump up from 2.5 to 3 bar.

[DanThe]

Try it with the electrics loaded, i.e. heater on, full beam lights on, heated rear screen etc, I've had bad fuel pumps that get worse to the point the car will cut out when loading up the electrics.

But a pressure test is the most reliable way to test

[DaniTD]

Try it with the electrics loaded, i.e. heater on, full beam lights on, heated rear screen etc, I've had bad fuel pumps that get worse to the point the car will cut out when loading up the electrics.

But a pressure test is the most reliable way to test

I tried that, connected everything (moving windows up and down too ) and the rpm would drop for a moment but then readjusts.
I think I should bring it to the workshop to make the fuel pressure test as I don't have the tools.

Another thing I noticed but that I don't know if it has something to do with...
If the engine is idling and I give it some gas and say it gets to 1200rpm, and then I maintain the pedal on the same exact position, after one second it would go up to 1300rpm, and maybe 2 seconds later to about 1400rpm. Is this normal? It seems a bit weird to me

[DaniTD]

So, I think I found the reason for the high opening time of the injectors...
Today, I connected again the car to the DIS, and read that CO pot value and its 1.803V. So, BMW says nominal value is 2.8V then I said let's put it to spec (1.80V tells ECU like the O2 sensor that mixture is lean so it opens up the injectors)... but the CO potentiometer DOES NOTHING. It stays at 1.8V no matter how much turns I give to it.
So, tested the pot with a ohmmeter and its working. TIS says that if CO pot is broken or the wire is cut, the ECU should store a fault and give the pot a default value of 2.77V, but I don't see any faults stored neither the substitute value is active.

Anybody saw anytime some kind of "bad fix" as cutting the CO pot wire and attaching a resistor for some kind of strange reason? Maybe the ECU is toast?

[Tommek]

Maybe the pot wire is disconnected. It's a black and white wire witch should have a plug somewhere in the engine loom under the bonnet. Check if that wire is plugged in. If you can't find it i'll make a photo of it later

[DaniTD]

Maybe the pot wire is disconnected. It's a black and white wire witch should have a plug somewhere in the engine loom under the bonnet. Check if that wire is plugged in. If you can't find it i'll make a photo of it later

Thank you! I read on the Electric Troubleshooting Manual that care must be taken with non-cat cars for the CO pot plug being connected, but I didn't know that one existed. Where is it located? Under the plastic black cover under the windshield?

[Tommek]

Yes, that's where it is ... it's a black plug

[Brianmoooore]

As above. White/black wires with single pole black plastic connectors, situated under the bulkhead wiring cover, near the middle of the car.
Wires are left unplugged for lambda sensor, and are connected together if no lambda sensor is fitted.

[DaniTD]

I can't believe it.
I found this:



I checked continuity between the terminal of the last photo and the CO pot terminal on the AFM connector, and indeed it is the CO plug. As can be seen, it is disconnected.
Questions now arise. Why? The plug seems to be factory attached to the loom to be not connected, or maybe it was done long time ago? If so, why DME doesn't give an error about CO pot disconnected and it's not using the substitute value of 2.77V?
The ECU coding is F01E which is prepared for (so it doesn't have cat neither lambda sensor) and I checked the lambda sensor loom and I have the plug at the connector under the battery.
Should I connect this to the other end? The hints says so but it puts me back because of it being well tied to the loom.

[Brianmoooore]

If the plugs are tied back, then this would indicate that the car left the factory with a lambda sensor, or the engine loom has been changed at some time.
One of those white/black wires goes straight to the rotor of the CO pot, and the other goes to the ECU. The CO pot will do absolutely nothing with the wires not plugged together.

[DaniTD]

Maybe the loom was replaced, the EPC says the car left the factory with the option "Without catalysator". I think I should plug the two ends then and adjust with DIS the CO pot to 2.8V... hope it doesn't burn anything when plugging...

[Brianmoooore]

You should connect the two ends, and adjust the CO pot on a gas analyser for 0.2% - 1.2% CO.

[DaniTD]

I think I can say... PROBLEM RESOVED!!!! After 5 years, I can't believe it.
I connected the cable and as I don't have a CO meter, I set up with DIS the position at it's nominal value (2.8V). No more missings at idle, before I had every 8 seconds or so a strong vibration like a miss in the engine, audible at the exhaust. It has dissapeared. And the overall vibration of the engine is so smooth, I can put the hand on the hood and notice a very faint vibration. Only need to verify the condition with the CO meter, but at least it seems improved a LOT!!
Thank you!

[Tommek]

Well done mate, i'm glad u figured it out. Mine is still missing, but i have a lambda so things get complicated a bit.