Rough idle - how smooth should it be?

[BristolE30]

The ecu is a possibility, you can use a 320i one and swap your 325i chip in it. If you try a different one, do some research and find the latest 320i ecu that was available

Yea I’ve been looking at a few used 380’s as I think they were the latest!

I have both numbers: originally 380 and a spare 173 I have run with for a significant time, and the only reason why my idle would be somehow erratic (not seen on the needle but feeling the car bouncing) was when the lambda was disconnected and the AFM mixture screw not screwed deep enough (hence idle running too lean). Adjusting the screw would instantly make my idle real smooth and stable, still with the 173 ÉCU.

Now I have some other issues so can’t confirm if there is a difference between the 173 and the 380, but I did not feel any so far.

I must admit I did not go through the whole post, but from your latest messages: do you have a lamba sensor? If so, have you connected it? Asking because I wonder if your problem could come from a faulty lambda so the ÉCU would adjust to an erroneous input. And indeed, when you disconnect the battery long enough you clear the lambda adjustments, which could explain why idle works fine for a limited time after reconnecting the battery, but not on a long run.

Other test possible: disconnect the battery to remove all lambda readings, disconnect the lambda sensor, reconnect the battery and see if idle remains stable after a correct adjustment of the mixture screw on the AFM. This adjustment I key to having a stable idle when running without a lambda.

Thanks for the input mate! Unfortunately I don’t have a lambda sensor, I checked for one earlier in the thread.

The only things I haven’t done in the idle circuit is: open up the AFM, clean it and find a new carbon track for the sensor arm. Change coolant temp sensor (although I did test it and was in spec for hot and cold). I guess I could also change the TPS but similarity I have tested it and it has continuity at idle stage

[Banban3]

For the coolant temp sensor, you did the check on the sensor itself. What about the wires? The issue I currently have is that the sensor is correct, but the wiring faulty, so the ECU reads a constantly changing resistor value. Therefore the ECU constantly adjusts the injection even when trying to drive at a stable speed, causing erratic behaviour of the engine. I hope I'll get the connector and other furniture I need tomorrow to get it fixed.

To check if the wiring is faulty, you need to unplug the ECU and measure the resistance value directly from the ECU plug. I'd recommend to slightly move the cables while measuring to check for any variation in the measurement, simulating driving conditions.

See the table below (pin 45 and ground):
- Pin table:

- Tests table:

As for the AFM, I'd say it would be surprising since after your reconnect the battery it runs fine. Yet, that could be another lead. On my side, I could not find a new carbon track, so I had my AFM refurbished by ATP Electronics (https://www.atpelectronics.co.uk/). Good for you: they are located in the UK! Very professional service, super fast. Rather expensive - still less than a new AFM - but refurbishment guaranteed for life. I am fully satisfied with their service.

But recomment you test the coolant sensor wiring first.

[HJ1981]

Additional thought:

If the butterfly stop has been fiddled with in the past, it could be set with insufficient gap between the butterfly and the body.

The ICV does not provide enough air for the engine to idle on its own, closing the butterfly down too much, even with the TPS correctly adjusted will cause the engine to have lower idle and hunt...the more you close the butterfly the worse it will become ( even with TPS set correctly) until eventually the car will start and then stall...pretty much with the butterfly closed at this point.

There's something in Bently about how to set the static butterfly position...I'll see if I can find it.

It says 15 thou between the throttle plate and the body...but this is rubbish - if you set this you will have at least a 1500 rpm idle and you won't be able set the TPS at all -

I've checked two throttle bodies which run the car correctly - so my advice is to set the gap between butterfly (plate) and body to 3 thou' it should look like this:


throttle butterfly opening.jpg

as you can see, the butterfly has no contact with the body which also stops it from jamming.


Paul

I apologise for raising an old thread, but i replaced the original 30 year old TPS due to it being faulty and found that the stop screw had been fiddled with and the plate was making slight contact with the housing.

Now, following the manual it indicates 0.0015" clearance, which is 0.04mm (approx). Yet you state that it is not correct? and setting it to 0.003" (approx 0.07mm) is correct? i found other threads where users used a 0.002" (0.05mm) gap. My question is which one is it? I know it is never truly closed and slightly open, but by how much?

[Brianmoooore]

Now, following the manual it indicates 0.0015" clearance, which is 0.04mm (approx). Yet you state that it is not correct? and setting it to 0.003" (approx 0.07mm) is correct? i found other threads where users used a 0.002" (0.05mm) gap. My question is which one is it? I know it is never truly closed and slightly open, but by how much?

None of the above.
It's set by substituting a fixed orifice for the ICV and adjusting for a specified idle RPM. In the absence of the special fixed orifice, I block off the ICV and adjust for a very low lumpy idle, with the engine threatening to stall.

[HJ1981]

Now, following the manual it indicates 0.0015" clearance, which is 0.04mm (approx). Yet you state that it is not correct? and setting it to 0.003" (approx 0.07mm) is correct? i found other threads where users used a 0.002" (0.05mm) gap. My question is which one is it? I know it is never truly closed and slightly open, but by how much?

None of the above.
It's set by substituting a fixed orifice for the ICV and adjusting for a specified idle RPM. In the absence of the special fixed orifice, I block off the ICV and adjust for a very low lumpy idle, with the engine threatening to stall.

I do not mean to be rude, but i have never read this kind of procedure for anything related to the E30. It sounds closer to adjusting the idle on certain M20 models, where one uses the special BMW tool to bridge the TPS in order to fully close the ISV and use the idle screw to adjust the rpms (during normal operating temperature) to 760 +/- 40 rpms, but again that is for certain models.

I followed the Bentley manual, as well as some German threads, the 0.0015"/0.04mm is "correct" for the throttle plate gap and housing (for the M20, M30 and various BMW automobile and motorbike models apparently). I adjusted to 0.04mm (about 1/2 a turn after the stop screw touches the arm), reset the DME/ECU by unplugging it for at least 10 seconds, replugged it and started the car. It idles funny for 2 minutes, which the DME/ECU uses the default ignition settings and learns and smoothed out. Everything is bang on now, starts up right away, idles fine cold/warm, no throttle lag etc.

[Brianmoooore]

I do not mean to be rude, but i have never read this kind of procedure for anything related to the E30.

It's something I read a long time ago. I have no idea where, but it makes perfect seance to set up the throttle stop in the way I describe.
You're trying to set the throttle for a certain RPM, so that the ICV can 'add' the amount of revs it was designed to do the most effectively to bring the idle to the correct level. With the very small mount of throttle opening you're looking for, tolerances and measurement difficulties will be a much more inaccurate way of setting it. With a substitute dummy ICV, the results would be accurately repeatable time after time on the same and different engines.

[coopman]

The part circled in red is the elbow that goes from ICV to TB (but you knew that!). It is a splined fit but the splines can wear away very quickly which leads to a dodgy fit & seal. I used a few drops of adhesive on mine and it's held up over the years.

Just saw this and I think maybe I have a worn spline issue. The part slides in and out as smooth as silk, which I just thought was the way it was meant to be. No obvious leaks, but I guess a bit of glue would be sensible?

[Brianmoooore]

The part circled in red is the elbow that goes from ICV to TB (but you knew that!). It is a splined fit but the splines can wear away very quickly which leads to a dodgy fit & seal. I used a few drops of adhesive on mine and it's held up over the years.

Just saw this and I think maybe I have a worn spline issue. The part slides in and out as smooth as silk, which I just thought was the way it was meant to be. No obvious leaks, but I guess a bit of glue would be sensible?

There's also another (or two, depending on production date) similar elbow on the other side of the TB, associated with the brake servo, that can come loose and leak air.