Re: 156 wonky idle - THE FIX !!!!
Hey guys, this is my first post on this forum, I saw a link to this thread on the Alfa156.net tech forum, and since I have also been suffering with this problem on my Selespeed for several years, I think it's a major step forward in the search for the <complete> solution for this problem...
As a bit of an electronics specialist I have some thoughts on this - first of all congratulations to Maverik for coming up with the idea to cut the wire in the first place. I think this proves that the problem is electrical. I think there is more to it though. The reason I think that is that on my car it did not fix the problem completely, but had a very interesting side effect.
my theory on what is the reason for the ecu to be fed with the speed signal from the ABS is that when the car is moving it invokes a special way of cutting off the fuel more gently than when the car is stopped, this is to avoid a 'jerky' feel when you take your foot off the gas when moving. On my car, cutting the wire from the speedo to the ecu partly fixed the bouncing idle, but also made the throttle feel more jerky when you take your foot off the gas.
What is more interesting is what happened when I looked at the car speed display in Aladiag after cutting the wire from speedo to ecu. Initially the speed was zero, as expected, but after a bit of driving, especially above a certian speed, the speed would suddenly jump up to a fixed value of about 20km/h, and stayed there even if the car wasn't moving until I switched the ignition off and then on again. When this happened, the idle bounce came back again, even when the car was stopped - again, this would be expected given what we know.
The question is, what caused the speed to give this erroneous value in the ecu, when the wire has been cut? I suspected this was because if you leave a wire which is an input signal 'floating' electrically, it can go to an undefined value that can be interpreted by the circuit (in this case the ecu speed input) as a false signal. What you normally have to do is to tie the line via a resistor, to either ground or the + rail (this is called a pull-up or pull-down resistor). So I tried this, but to my surprise the bouncing came back with a vengence, and all the time! Looking at Alfadiag, sure enough it showed a randomly generated speed signal... If i had the windscreen wipers going, it was even more crazy. I tried both pull up to 12V and pull down to earth, via a 1k resistor, and with the same result.
This indicates to me that electircal noise is getting onto this line, and being interpreted by the ecu as a speed signal, which confuses the system that is supposed to ramp the rpm down gently when the foot comes off the gas - hence the bouncing.
I think that, if this noise is there when you disconnect the wire and connect it to earth, it is probably also there even when the wire is connected to the speedo - ie, it could be there along side the real speed signal, and therefore confuse the hell out of the ecu.
The question is , where is this noise coming from? This is where I don't yet have the answers. Certainly I've listened to the pulses that are generated by the speedo signal on the cut wire and they are similar in frequency to the pulses generated by the ignition coil ccts, or the injector circuits.
One other thing I have noticed is that I started to get a lot more interference on the long wave radio band at a certain mileage, which was not there when the car was new. This did seem to correspond with the time when the bouncing idle problem started. I improved this recently by cleaning the ground connection on the aerial amplifier in the rear C pillar... but it's quite a well-known problem with many cars that all the earth connections can deteriorate over time and cause all sorts of problems. The source of that noise is still there, and we need to find it. It's bound to be a generic problem on all 156s...
I know someone on here said they'd had their earths checked, but it could be other things as well - for example the alternator... now there's an idea - disconnect the alternator and see if the problem goes away. Only problem is you have to get under the car to access it, and in this weather I don't fancy that!
Although my car is a Selespeed, I believe this particular fault has the same root cause in both the manual and the Selespeed, maybe just more sensetive on the Selespeed, as there is more load on the electrical system with this system.
So my new quest is to find out how to reduce this noise - is it a grounding problem, a noisy alternator, is it a bad connection in one of the in-line plugs, is it noise in the speedo area specifically (I'm going to check this by connecting my pull-down to a ground point nearrer to the ecu itself) - if anyone else has some ideas on this theory, lets hear them!
Ive been having a good look at the power and ground wiring on the workshop CD, and there are quite a few hidden earth points that could be removed, cleaned and tightened up. There's also the ECU itself - it's probably got an internal ground connection that could corrode, I saw this when I opened up my duff airbag ecu once.
A good way to check the effect of any improvements here is going to be to monitor the noise level on the long-wave radio.
By the way, referring to the posts above from AlfaBR and eggie: anyone with a Selespeed should NOT disable the ABS unit in order to get rid of the speed signal - there is another feed from the speedo to the Selespeed ecu which is essential and is used by the Selespeed to know when to change down gears automatically, and disabling the abs will take out all three of the speed signals (selespeed, main ecu and aircon) as well as being undesirable safety-wise. I havent tried disconnecting this line to the selespeed ecu (pin 14 I think), but the gearbox will definitely not be happy!
Sorry for the long post, I promise this is a one-off...but this has all been majorly exciting for me, and again, thanks due to Maverik for the great idea in the first place!