After two years I still haven't managed to crack the problem of my ECU occasionally claiming angle position sensor 2 in my throttle body is giving implausible readings. Had two different throttle bodies on it too, and even patched out the signal wire from that sensor including fresh connectors just in case. No dice. Looking around on the net suggests that this exact problem (or subtle variations of it) happens quite a lot with DBW throttles, and nobody has ever come up with a solid reason and reliable fix. Cleaning and running a full adaptation on the the throttle body often slaps a plaster over it, but it'll tend to come back. The workshop guide suggests that if you replace the throttle body, wiring and ECU (do the maths..) that it may fix it. Not will, but may.

I've been thinking about this and had an idea that I wanted to run past our more electrically minded folk.

I'm wondering if the TB might be acting as the electrical equivalent of a miners canary (go with me on this one) and becoming a weak spot in the event of a voltage drop across other parts of the system. The two angle sensors are mechanically connected and work in opposite directions, and so their outputs shouldn't be able to go out of sync unless the connecting gear breaks or the contacts are knackered. The power lines run direct from the ECU with a shared V+ and ground for both sensors, and individual signal lines back.

My reasoning is based on having 5v across the sensors with a tiny buffer zone at each end to detect earths or shorts. If sensor 1 runs from 0v > 5v and sensor 2 runs from 5v > 0v,then a drop in the supply voltage would leave sensor 1 unable to go below zero, but would result in an increased error from sensor 2 as the initial voltage dropped. I can't think of any other electrical systems in this engine that compare an inverted signal and might be affected in this way - all other errors would just scale. If this were the case, it would also explain why the error most often seen accompanying this is 'low voltage at base setting'.

Given that this problem occurs for me and many others at completely random times, it's not possible to reproduce it and test the theory. The only additional info I have is that I've noticed recently that at certain times the dash and aircon lights on my octy can dim fractionally, so little that you wouldn't notice if you weren't looking. The last time this happened it was accompanied by a tiny click that appeared to be coming from the centre of the car. I also noted that in icy conditions the headlights dimmed fractionally as the antilock brakes kicked in. At all other times it runs fine.

Any thoughts on this? If it were something along these lines it might also explain why all attempts at diagnosing this seem to fail - everyone's looking in the wrong place

Edited 17 December, 201015 yr by Interphase

. I can't think of any other electrical systems in this engine that compare an inverted signal and might be affected in this way - all other errors would just scale.

Throttle sensor (at the pedal) on a DBW car works like this. Your theory makes sense though, potentially. The two sensors are there as a failsafe so one going screwy doesn't give false info if it gets noisy.

Runnimg them a 5v reduces the fluctuations at the sensor (because the 5v is lower than the car electrics and regulated down), but a resistive connection from the ECU to the sensor could give problems.

Edited 17 December, 201015 yr by cjb

Doesn't the throttle pedal have two sensors with an offset, but both running in the same direction? A supply voltage drop there would mean the offset would remain but just scaled down.

Running a multimeter across all of the wires in that part of the loom gives solid and consistent results, even when yanking them around and wiggling the connectors. We took the precaution of patching out the sensor 2 signal wire based on the theory that only that sensor is affected, so the shared power rails are probably ok. In hindsight, and with reference to my current thinking, that might not actually be the case. The difficulty is in working out how the ECU arrives at that error so we can figure out more tests - my theory is that it compares both sensors with the voltage supplied to the butterfly valve motor and plays spot the odd one out

If it's a voltage drop because of increased resistance and it still happens after those tests then the only thing left would be the supply from the ECU. Given the number of people who are experiencing this, I'm not confident about the odds of such a specific failure there. Working out if the ECU might be dropping that voltage for some reason, or if it's a consequence of a load elsewhere is tricky.

Cheers for the feedback, much appreciated

Edited 17 December, 201015 yr by Interphase

Well, still cropping up at random intervals.

When it was serviced the other day, they noticed that the voltage across the battery was a bit lower than it should be; that might explain the 'low voltage at base setting' errors but it wouldn't explain why it can happen on the move when the ECU isn't running a basic setting on the throttle body.

The way I see it is that the battery can't be holding less than, say, 13v but still feel healthy & crank OK. If the throttle body sensors only need 5v but are still going out for the reasons I suggested, then that supply must drop as the feed to the ECU drops. I suppose the voltage required for a basic setting may relate to the butterfly valve motor though (possibly 12v?) but a drop there whilst on the move shouldn't cause an error. I'm wondering if something may be introducing a high load at seemingly random intervals, or maybe even there's an intermittent earthing issue that isn't throwing codes. No idea of how to test for those though as either really could be anywhere.

I started the car earlier and noticed that he interior light dimmed a *lot* for a couple of seconds after ignition. The car ran fine at idle while it was warming up, but then the dash lit up 20 seconds down the road and it went into limp mode. I managed to pull over, turned off everything, killed the ignition then gave it 30 seconds to run a basic setting on the throttle body. It ran fine after that, albeit with the traction control light still on - that'll clear the next time it starts.

I'm pretty convinced now this is power related, but I have a question which may have an obvious answer so apologies if that's the case. What happens to charge from the alternator once the battery is full? Is there a circuit or component that directs that charge to a sink and, if so, what does it use as a trigger?

Edited 29 January, 201115 yr by Interphase

the alt is internally regulated, the biggest current draw is going to be cranking the engine so maybe that is when its picking the fault up?

I've not got my wiring disgs here but will look tomorrow to see if the 5v feed to the tb is shared with anything else as most of the sensors are 5v

The problem does kick in when it's running as well as starting though, but I think it might be a different version of the same problem. If that were the case it would mean the 5v out from the ECU was sagging for some reason. Tracking that down might be the key. It might also be that the first fault is logged at startup and then it clocks up more throughout the trip until there are enough to trigger the code..

The feed for the TB comes directly from the ECU; there are six wires running between them - +/gnd for the angle sensor supply, +/gnd for the motor and one signal wire per sensor. If anything shares the supply it'll be within the ECU but I'm not sure if each subsystem has its own regulator. If anyone has cracked open the ECU and knows the answer it'd be hugely appreciated Thank you for offering to check the diags though!

Thinking about it, pretty much all of the power when the engine is running comes directly from the alternator doesn't it? Given that the headlights were dimming as the antilock brakes kicked in on ice over the winter, I'm wondering if the alternator isn't providing enough juice on the go - either because it's playing up, or because the famous cable resistance issue is starting to show up. I suppose that if the battery were starting to go at the same time that'd place an additional constant load on the alternator too. In that context, the intermittent dimming of the dash lights is intriguing too.

Edited 29 January, 201115 yr by Interphase

The problem does kick in when it's running as well as starting though, but I think it might be a different version of the same problem. If that were the case it would mean the 5v out from the ECU was sagging for some reason. Tracking that down might be the key. It might also be that the first fault is logged at startup and then it clocks up more throughout the trip until there are enough to trigger the code..

The feed for the TB comes directly from the ECU; there are six wires running between them - +/gnd for the angle sensor supply, +/gnd for the motor and one signal wire per sensor. If anything shares the supply it'll be within the ECU but I'm not sure if each subsystem has its own regulator. If anyone has cracked open the ECU and knows the answer it'd be hugely appreciated Thank you for offering to check the diags though!

Thinking about it, pretty much all of the power when the engine is running comes directly from the alternator doesn't it? Given that the headlights were dimming as the antilock brakes kicked in on ice over the winter, I'm wondering if the alternator isn't providing enough juice on the go - either because it's playing up, or because the famous cable resistance issue is starting to show up. I suppose that if the battery were starting to go at the same time that'd place an additional constant load on the alternator too. In that context, the intermittent dimming of the dash lights is intriguing too.

Could also be a bad ground in conjuction with the alt...

Definitely possible. I'm going to be going over it with a fine toothed comb this week, tracking all the major earthing points back and doing more continuity tests. I get the feeling there's a bit more going on as the fault doesn't show up on anything other than the throttle body, but you have to start somewhere

does this help or not

maybe it is an earth problem as they seem to be seperate on each side where the + is common if i'm reading it right

B169 is the throttle body

B138 is the pedal

Edited 30 January, 201115 yr by felicia16v

Coud you not run a 12V feed to the TB, and then use a 5v regulator chip to power the angle sensors? That should make it a bit less sensitive to voltage spikes and drops.

maybe it is an earth problem as they seem to be seperate on each side where the + is common if i'm reading it right

I think it shows the + and GND to each of the potentiometers come from the same source. The signal wire back from each is separate (pins 84 and 92) but the two pots are mechanically linked so that as one goes up, the other comes down. The motor to the right has its own feed, isolated from the sensors. I'm wondering if sensor one is the 'trusted' one and sensor two provides an inverse signal that can be used as a check? I think that sensor two starts high, which is where my theory about the effects of a voltage drop comes from.

Coud you not run a 12V feed to the TB, and then use a 5v regulator chip to power the angle sensors? That should make it a bit less sensitive to voltage spikes and drops.

Intriguing idea, but I'm not sure how the ECU would feel about it Because of the essential nature of that circuit, I think it switches in the voltages during the initial checks and requires the correct feeds to be in place before it runs a basic setting. There might need to be a constant load on the + feed equivalent to what would happen during normal operation.

Edited 30 January, 201115 yr by Interphase

I think it shows the + and GND to each of the potentiometers come from the same source. The signal wire back from each is separate (pins 84 and 92) but the two pots are mechanically linked so that as one goes up, the other comes down. The motor to the right has its own feed, isolated from the sensors. I'm wondering if sensor one is the 'trusted' one and sensor two provides an inverse signal that can be used as a check? I think that sensor two starts high, which is where my theory about the effects of a voltage drop comes from.

Intriguing idea, but I'm not sure how the ECU would feel about it Because of the essential nature of that circuit, I think it switches in the voltages during the initial checks and requires the correct feeds to be in place before it runs a basic setting. There might need to be a constant load on the + feed equivalent to what would happen during normal operation.

Just a thaught could it be the electronic sensor on the pedal itself??? It is a common fault on them could be dirty or worn???

I wondered about that early on and had a look at the various codes related to the whole DBW system; the pedal has a set of its own fault codes that are actually quite similar to the ones for the throttle body. The pedal has two pots just like the TB (as shown in Felicia16v's schematic) but those travel in the same direction albeit with a difference in starting voltage - as long as the gap remains constant, the ECU knows that the pedal is working. Any faults detected at that point relating to implausible or out of range signals are trapped before they're passed to the TB.

I think the key to this is that both pots in the TB have exactly the same power supply, and that threw me for a while.. moreso when I realised quite a few other people have the same issue with sensor 2 alone. As the sensor is fine, the signal wire is new and the fault is intermittent, I can only assume that this error is the result of a wider issue, and the only reason it affects sensor two alone is because that one is wired backwards.

I wondered about that early on and had a look at the various codes related to the whole DBW system; the pedal has a set of its own fault codes that are actually quite similar to the ones for the throttle body. The pedal has two pots just like the TB (as shown in Felicia16v's schematic) but those travel in the same direction albeit with a difference in starting voltage - as long as the gap remains constant, the ECU knows that the pedal is working. Any faults detected at that point relating to implausible or out of range signals are trapped before they're passed to the TB.

I think the key to this is that both pots in the TB have exactly the same power supply, and that threw me for a while.. moreso when I realised quite a few other people have the same issue with sensor 2 alone. As the sensor is fine, the signal wire is new and the fault is intermittent, I can only assume that this error is the result of a wider issue, and the only reason it affects sensor two alone is because that one is wired backwards.

Don't always rely on the codes as it could lead you astray??? There not gospal I have had this before

That's very true, especially with this engine The problem is that there's little else to go on and I've tried all of the more straightforward stuff over the last two years.

The real issue I've hit is that nobody seems to understand how the ECU decides to set many of the codes; I'm sure Bosch have all the info, but for the rest of us it's a case of replacing the entire component because the workshop manual says so. if I knew how this fault was decided upon I could figure out some tests. The only test I'm can think of to start with is to briefly introduce some extra resistance to the positive line feeding the angle sensors - if dropping both the sensors results in a fault on no. 2 alone, then it'll confirm what I thought.

Well, over 2 years on and I still haven't sorted it. A series of exchanges with Skoda's aftersales team left me with a lifetime supply of WTF and a determination not to touch VW based again, and it seems all of the people I've tracked who are having the same problem are in the same boat.

Yesterday I did a Google search just to see if anything new had come up and found this:Information on TB 2018652/3

Seems that fairly new cars with the generation of engine after mine are still suffering with this issue, to the point where VW have put out an advisory for the throttle body connector and wires. The earliest car I found with the problem is a '99 passat, so that's *12 years* of failures to diagnose and repair. I'd assumed that the ECU triangulated the signals from sensor 1, sensor 2 and the valve actuator in order to figure out which was wrong, but it seems now that any fault with the electrics that isn't a short or ground will flag up sensor 2. How many throttle bodies have been replaced on those terms?

Problem is that Skoda are unlikely to accept that it's been their iffy wiring and connectors all along as this fix isn't in the diagnostic process detailed in the workshop manual. Chances are I'll have to get an indie to do the full harness overlay and take it from there.