1.6 AEE Throttle Body and related

My problem is common to many. I just want to relate my thoughts and reflection as I do think every mechanical and electrical has a precise cause, it may be elusive, but it's there somewhere.

For the past 3 years my 1995 Polo 1.6 AEE with 140K on it has had the ever frustrating idling problem. There seems to be no ryhme or reason to what occurs. Generally it will cut out when you need it most - negotiating intersections. It accelerates unevenly. sometimes things will seem perfect then all of a sudden it's crap.

Since the garage hasn't found a solution nor do they seem to want to really to figure it out - I'm sure they see the problems with others, but probably some of the solutions work - just because of similar symptoms for different root causes.

Over time I've worked at this patiently - I know that in the end I will get it.

What I am able to determine so far -
Engine will start fine, idle will be high at cold as expected
Drops within 1-2 mins so this is within specs.
this should exclude sensor malfunction related to engine/coolant temp, etc
after 3 mins during which time the idle can be seen to exhibit typical warmup behavior - the exhaust and engine roughness evens out
the idle at this point will purr over nicely.

then one can feel that it misses a beat and then more
rpm will clearly start to fluctuate as the ECU tries to compensate whatever cause the system to go out of whack

I actually have two TBs. I put one in and run thru the basic settings to adapt the TB. This way I can still use the car while I patiently clean and see how to recondition the original TB.

Either the impact is fairly limited. With both the engine will still idle poorly. they are not exactly the same in terms of the extent of symptoms. Cleaning appears to make an impact, but how can it be that after 2 days the symptoms return because the TB are dirty again? The butterfly and throat have some deposit after 2 days of running, but I can't believe it is sufficent to be the problem. Movements are free and there are no hard carbon deposits.

Without the air filter on, I can clearly see the TB adjusting itself on its own when the idle acts up. From what I can tell it starts its adjustment when the engine rpm drops - so I think the rpm drop is not a consequence of a TB behavior. Infact as the TB opens the rpms rise.

So why is it the ECU thinks there is a change in system status if I'm not touching anything, ie accelerating? excluding some sensors because they should also exhibit an impact at other points, possible culprits are the MAP sensor, lambda, and knock sensor. Lambda is unlikely as it should also show its effect elsewhere if it were at fault. Likewise knock sensor. I've replaced the MAP sensor , but no significant change.

Parts of the ignition system has already been changed, so likely to be excluded.

So where does this leave me? I'm more than sure it's what we call unmetered air. It means there is a fault somewhere whereby air can get in past the vacuum. The sensors don't expect this, but the air mixuure will be different from what the ECU is expected. This kind of idling problem on carb systems was pretty easy to determine simply because there were no ECU who tried to compensate at superspeed.

Unfortunately locating a vacuum leak can be a nightmare. Finding them at joints, vacuum hoses,etc is easy enough, but on the Polo there is one other factor. The intake manifold itself is made of plastic. As strong as it might be, there can be faults at seal points as well as microscopic cracks. Clues that point in this direction - temp in the engine compartment will cause changes to plastic expansion/contraction.

As a home mechanic with very limited space I want to avoid trying to replace the manifold as a problem determination method. Besides it cost me a pretty penny to find.

Given the unpredictable nature of when the idle problemwill occur, but when it shows up the characteristics are most closely associated with small and inconsistent vacuum leaks.

A truly certified VW garage has the tools to run a leak down, but they are unlikely to do it as it is time consuming and in the end they can't make any money (their objective is not that to make a customer happy).

Hope to make some more progess this weekend.

I have read your post with great interest as you seem to have similar problems as me.
I have posted a query recently (Polo "hunting") about my 1.6sr Polo but so far no one has responded.
My problem is definitely temperature related - and by that I mean ambient temperature.
In cold weather the problem rarely occurs but in warmer weather it starts within 5 minutes of startup. I have removed, checked and refitted the inlet manifold making sure that there are no leaks at the manifold/head joint but since reading your notes I will be removing it again and checking for cracks and fisures.
I must say that I have experienced leaky manifolds on several other cars but have never know them exhibit such extreme symptoms.
Please keep me imformed if you make any progress and I will reciprocate.
Best Regards,
Steve

After several other iterations and analysis, my problem is definitely resolved.

I patiently returned to observing the symptoms to see if I could define some other defining characteristic. Indeed I found. When the uneveness appeared I perceived it is as being relatively random - perhaps more with hot than cold, etc. In reality it was cyclical in nature, but the cycle was more on the order of 1-2 mins. For example it would run fine for 40 secs then it would start the rough cycle, after 30 secs it would clear up and then start all over again. When you listen for this in the span of 90 secs you would get the impression of general uneveness. Actually it is quite difficult to be concentrated for 3 full minutes unless you know what you are looking for.

Why would it be cyclical albeit a little long in time. It means the ECU has been able to compensate, but then that compensation becomes an overcompensation. There's only one sensor that has an active component - the lambda sensor.

With that as the prime suspect I digged deeper. Went back and used the VAG diagnostics to take a closer look. No fault codes related to the lambda or otherwise. However with the VAG reader trace which plots the lambda voltage as the engine idles I could see it rising and falling, which it should, but the interval time was something like once a minute. It should have been closer to 30 times a minute. This is what is called a slow lambda. So the ECU adapts correctly but at the point the mixture will stay too rich or too lean for too long, all causing rough idling. It will cause acceleration problems too. Most importantly the minute you tap the accelerator pedal the engine will likely cutout as the mixture changes suddenly without immediate ECU compensation with a correct lambda reading. Beyond 2000 rpms the problem disappears because the lambda is no longer a critical sensor.

I had asked the garage in the past if the lambda could be the problem and the answer was no, it was fine. Sure, if they are lazy and look only for fault codes. A slow sensor will not show up in fault codes.

Why a slow sensor? it means that it is no longer performing correctly but is still providing voltage, hasn't completely stopped working. This should have been an easy suspect given two other indicators in my case. I have 140k and lambda sensors do have an age limit, probably anywhere between 80/120k in normal usage. I also had a thermostat problem in the past causing an overheat with consequent coolant loss. This required a new head gasket. It is possible that some coolant got into the exhaust which impacted the lambda for the long term.

With the new lambda and 1 week of running, the engine now purrs like new.