I'm not sure what's normal but I'm pretty sure this isn't. With a timing light I get around 5 degrees of variation in the ignition timing at idle, making it impossible to set the timing properly. It will go for a few seconds at exactly the same point, but then go off to anywhere up to 5 degrees ahead (but all over the place in that gap), then back to what looks like the proper timing point for a while.

I've replaced pretty much everything on the high tension side (plugs, leads, dizzy (second hand), cap, rotor, coil) at some point. Any ideas what else it could be? I don't really understand how electronic ignition distributors work but I think the ECU controls advance - could it be getting a dodgy signal and firing the spark at the wrong time? I don't know what sensor would influence it so much and so quickly though. I've never seen any of the sensors reading really stupid values on vagcom.

While I'm at it if anyone knows - Haynes says special equipment is required to properly set the timing (i.e more than a timing light). Is this correct for the Mono Motronic (the engine management system on Mk3 SPIs)? If the ECU really does have to be in a special state while you rotate the distributor then I'll never be able to set it properly, but I'd like to know what this equipment is before asking a garage whether they can do it, so I sound like I know what I'm talking about.

Cheers,

Ray

Are you disconnecting the blue temp sensor? You need to do this to make the timing go static so you can adjust it properly.

I'm pretty sure I did, this was when up to temperature (fan had kicked in but wasn't on when observing this) and blue temp sensor unplugged, at idle.

Ah, ok it should stay pretty much static then. did the revs go up at all when you unplugged it?

No, nothing noticeable happens when I unplug. I'll take a look at the coolant temperature reading on vagcom at the weekend and make sure it looks sensible and looks unsensible when I unplug the connector.

5 degrees isn't pretty much static is it

No, but it's not that bad really. And 5° too much probably won't do any harm if it's got a standard chip in as it's set up for 91 ron fuel.

I've had this trouble too, for a while. They best I can do is set the timing such that the 'average' location of the timing notch is somewhere near the recommended place.

I also have trouble with the timing remaining as I set it. It will eventually start pinking under constant light throttle openings. I retard the timing and lose both pinking and torque! I reset the ECU (unplug the battery), re-adjust the timing and everything is fine for a while. Are you having any trouble with pinking Ray?

Are you disconnecting the blue temp sensor? You need to do this to make the timing go static so you can adjust it properly.

The VW manual doesn't suggest the temp sender needs to be unplugged for SPIs'. It does apply to G40 engines though. Instead, just aim the stoboscope and twist the distributor if necessary....job done.

VW say that either V.A.G. 1367 or V.A.G. 1767 (ignition testers) should be used for checking and adjusting the ignition timing. By the look of it, this takes readings from the HT king lead, HT lead for cylinder No.1 and the flywheel (where that bung is located on the bell housing). I don't know if many VW garages will still have these machines to hand....depends if the newer models require it I guess!

Removing the blue temp sender works on my brothers AAU, the revs rise and the timing goes static. without it the timing moves around too much.

I don't get any trouble with pinking, no (as you've guessed I'm no expert on this and have never knowingly heard pinking on any car - I hope I'd recognise it if it was happening but I don't believe it is). It does, however, idle quite lumpily and vibrates between ~1200-2000rpm (including when slowing down through that range, which makes me wonder whether it's an ignition problem at all because the ECU shouldn't be fuelling/sparking at all then). I've also needed a new cat for the last two MOT emissions tests.

Just been out with vagcom, observations:

-Temp sensor reads correctly - between 88 and 95 most of the time

-Unplugging it reads -50 degrees, so that seems fine (no change in revs when plugging or unplugging, fault code gets logged when you unplug it (understandably))

-Intake air temp sensor seems about right - around 20 degrees with ambient temp probably between 5-10 degrees

-Ignition timing does indeed vary at idle between 2 degrees and 6 degrees (2.0, 4.0 and 6.0 are the values it shows at idle). This happens both with the temp sensor plugged in and not. So I guess this rules out all the ignition side for the erratic idle, and probably it's meant to be like this. Seems to make sense then to set it so that the earliest mark shown on the stroboscope lines up with the 6 degree BTDC timing mark - then it will be firing when the ECU thinks it's firing. I'll experiment with this on the weekend

-There are two "load" readouts, one is around 30% at idle and goes all the way up to 100% going up a hill. The other shows around 3.6% at idle and generally varies with the other one but periodically goes for long periods of time (10s of seconds) at 0.0%, with no noticeable pattern for when this happens (doesn't seem to be caused by any noticeable change in driving conditions). I have no idea where this measurement (or calculation) comes from, but it doesn't sound right that it behaves like that. Any thoughts?

- "Lambda" is constantly at 11.8% and never budges. "Lambda factor" varies between ~0.980 and ~1.120, always toward the top end at idle and it drops closer to 1.000 when driving.

I didn't know you could get so much info from a mk3 ECU! :) what software are you using?

The lambda might be worth looking at - The purple wire can be tapped and measured with a voltmeter, it should fluctuate between 0 and 1 volt when the engine is warmed up.

Yeah, you can't adjust anything with it but you can get at all the information the ECU knows as far as I can tell (which isn't a huge amount on this!). I'm using the full version of VCDS (vagcom), not sure what the restrictions on the free/eBay versions are.

I'll look at the lambda on the weekend too, see what "lambda" and "lambda factor" go to when I unplug it too.

Any ideas on the load readings? There's no sensor for load as far as I can tell, so it must be calculating it based on something else, and none of the other input readings seemed to go weird when the load went down to 0.0%.

This is all on an AAV by the way, which is mono motronic (along with AAU I assume) as opposed to Digifant for GT and G40 (and NZ I assume). Never plugged into any other mk3 Polo engine so I don't know what you can do with the Digifants, but I'd imagine at least the same and probably more.

Presumably throttle position and injector duty? I don't know why it would drop to zero though, unless coasting?

Yeah the AAU and AAV are identical systems. They seem to use a single wire for connection to the ECU (yellow with white stripe under the glovebox). My digfant setup is the same as that.

I'd have thought injector duty was based on load, and load would be based on more than just throttle position. Both can be measured and neither look wrong. Injector on time does drop to zero when coasting, but I definitely wasn't coasting when one of the load readings was 0.0%.

There's a throttle opening angle and also a switch (can't remember if it's closed, mid and wide open or just closed/mid), pretty sure the switch has always worked but I'll double check.

Yeah it uses the 2x2 connectors, can't remember how many wires go into each but one of them having only one wire rings a bell.

There's no air flow meter on these, so I'm guessing it works out how much air is going in based on throttle angle and intake temp only. Going to check for vacuum leaks at the weekend which would screw this up.

I'm not sure how else it would calculate load other than throttle position and lambda unless something is measuring manifold vacuum on the throttlebody? There are 12 wires going to the throttlebody and I've yet to figure out what they're all for. 4 of them are the injector and the air intake temp, the rest are either for sensors or actuators. I think the actuator is only there as a choke, it's the only time I hear it do anything on my brother's car.

I'm not sure how else it would calculate load other than throttle position and lambda unless something is measuring manifold vacuum on the throttlebody? There are 12 wires going to the throttlebody and I've yet to figure out what they're all for. 4 of them are the injector and the air intake temp, the rest are either for sensors or actuators. I think the actuator is only there as a choke, it's the only time I hear it do anything on my brother's car.

Here some info you might find useful?

Thanks for the info nige - damn I wish I had an oscilloscope!

Had another fiddle round today, think I actually learned something! Set the timing so the earliest it fires is at the 6 BTDC mark with the temp sender plugged in. I guess that's the best I'll get on the timing. Also did a throttle body alignment/adjustment, mainly because I could - don't know if it's made any difference.

I didn't put the voltmeter on the lambda wire but did plot the lambda factor in vagcom and it fluctuates every few seconds, I think that's the right period and the values were roughly centred around 1.000, but way outside of the 0.097-1.003 (?) that the MOT test requires - I'm not sure if that's directly comparable though.

The "lambda" percentage on a drive started off around 15% and gradually came down to around 12%. From a bit of googling it looks like this is the long term trim/adjustment percentage which the ECU uses to adjust the quantity of fuel it injects to achieve the air/fuel ratio it wants. It learns it over time based on the lambda sensor output - if the ECU thinks it's injecting enough fuel to get lambda 1.000 but the sensor then tells it that it was 1.100, it will inject a bit more next time to account for it. The "bit more" is this percentage. It could therefore be negative too. Positive means it injects more (the engine is running lean), negative is the opposite. I found one forum post saying that 12-15% is the range where a warning light would be shown on cars having one, so maybe this is where my problem is. I think it started so high because I erased the fault codes which might reset it.

So maybe I have an air leak, but I've no idea where it could be from. There's no MAF so it must be calculating the amount of air based on the throttle angle and intake air temperature, both of which add up on the vagcom readouts. So the leak would have to be beneath the throttle plate for it to be "unmetered", which leaves the vacuum pipes and gaskets (throttle body to inlet mani, inlet mani to head). I've checked as best I can for air leaks and obviously not found anything, not sure how I can check any better.

Another thing I thought of is the vent pipe coming from the rocker cover via a few things at the back of the block and into the air filter housing. There's a fair amount of crap pumping out of it constantly, I don't know what a normal amount is. This is not pure air, so might screw the mixture up enough to affect the A/F seen by the lambda. Any ideas?

But also I put a vacuum gauge on the pipe which controls the hot/cold air feed flap and it shows around 19 inches mercury and fluctuates around that by about 1 inch mercury. It's just at the low end of the range the gauge says it should be in, but more concerning is the fluctuation which various websites say could be an ignition problem - argh! The most specific info I found was spark plug gap too small - I've only ever put in the plugs in the Halfords catalogue and never checked or adjusted the gap. I assumed they wouldn't need adjusting, but I guess that's one thing I can definitely check easily. I've changed plugs previously and it's had no affect on the lumpy idle, but again with no gap adjustment on either old or new set.

Sorry this is a bit of a braindump, but maybe it'll help someone else sometime too...

Oh and for years it's had a fault code saying no signal from the hall sender. I've never understood this and always thought it must have been when the engine isn't running, in which case you obviously wouldn't get a signal from it. I don't see how the engine would run at all if it really was faulty (and besides, I've replaced the distributor including hall sender).

The "lambda" percentage on a drive started off around 15% and gradually came down to around 12%. I think it started so high because I erased the fault codes which might reset it.

If you started off with the engine cold (or not at operating temperature) this will be why the reading was high. The lambda probe is heated electricly; it needs to be hot in order to function. During the warm up phase, the ECU does not use data from the lambda probe but a pre-determined map instead. I can sometimes hear a 'click' as the heating elements switch off, usually followed by a slight drop in idle rpm. Untill the engine is fully hot, it will always run slightly rich.

There's no MAF so it must be calculating the amount of air based on the throttle angle and intake air temperature,

Correct, but it uses engine speed also, in this case using the hall sender.

So the leak would have to be beneath the throttle plate for it to be "unmetered", which leaves the vacuum pipes and gaskets (throttle body to inlet mani, inlet mani to head). I've checked as best I can for air leaks and obviously not found anything, not sure how I can check any better.

Yeah, this is difficult to do. You can get leak detector sprays but I don;t know how useful or effective these are. The stout rubber gasket/ block that sits between the injection unit and intake manifold flange can often deteriorate and cause leaks. I've changed mine (to try and solve the pinking) and it's a fiddle. You could spend an afternoon removing the gasket to find that there is nothing wrong with it. As a first stab, you could remove the injection unit and see what condition the top mating face is in. If this is perished and looks like it's blown, might be worth changing it.

Another thing I thought of is the vent pipe coming from the rocker cover via a few things at the back of the block and into the air filter housing. There's a fair amount of crap pumping out of it constantly, I don't know what a normal amount is. This is not pure air, so might screw the mixture up enough to affect the A/F seen by the lambda. Any ideas?

This is the crankcase ventilation or breather. You'll always get come oily mess from it, but it shouldn't be a lot. The breather system can increase CO and HC contents of exhaust gas if the engine oil has become diluted with unburnt/ partially burnt fuel. As the oil heats up the fuel evaporates in to the crankcase, mixing with blow-by gasses. These fumes and gasses are sucked out using intake manifold vacuum (via that pipe you saw) to be drawn back into the engine for combustion.

Changing the oil or getting the engine properly hot will sort this.

The most specific info I found was spark plug gap too small - I've only ever put in the plugs in the Halfords catalogue and never checked or adjusted the gap. I assumed they wouldn't need adjusting, but I guess that's one thing I can definitely check easily. I've changed plugs previously and it's had no affect on the lumpy idle, but again with no gap adjustment on either old or new set.

The proper plug for these is Bosch W8DTC which I don't think is available from Halfords. These are triple electrode plugs so the gap is pre-set and none adjustable. The latest recommendation from Bosch is for WR8DC+ which is only single electrode so maybe you (like me) are using these?

After all of the above, I've done a bit of digging around and apparently, Mototronic systems regulated idle speed by making small adjustments to the ignition timing. If this is true, it makes you wonder how you would ever be able to set it properly without the proper VW kit!

I think unless your having problems, like loss of power/ torque, pinking or emissions woes, I’d set it best you can and leave it alone. If everything seems fine then it probably is.

So maybe the ones in there weren't from Halfords - they were Bosch single electrode so maybe the WR8DC+ you mention, although they didn't have a part number on them that I noticed. They were indeed well under the gap specified in Haynes (they were 0.75mm compared with the specified 1.0mm). I re-gapped them and got maybe a slight improvement in the stability of the vacuum - it was varying by maybe 0.25 Ins Hg by the end. No real noticeable difference in lumpiness though, but certainly not any worse.

I went for a drive and the lambda got down to 9.6% which is somewhat better - I'll keep an eye on it now I know what it means!

Interesting about the TB to manifold block thing - I have a spare TB lying around somewhere but can't remember whether it has that block or not... I may have a look at the top as you suggest, it's about the only thing that could be left leaking vacuum (I've blocked off the hot/cold air flap and charcoal canister pipes just to be sure - no noticeable difference). I think I replaced the inlet manifold gasket when changing the head gasket, so it's unlikely to be that, leaving not much else if indeed there is a leak.

But generally I'm going to just leave it as it is for now apart from maybe checking that block at some point. I've spent so long on it I've forgotten what it was even like before and whether it's just an old car (I also drive a 2002 Passat TDI which is rather a lot quicker, maybe I'm expecting too much performance out of the old Polo now!). It could also be something as silly as an engine mount making it vibrate so much, I don't know. She's probably going to die due to chassis rust within an MOT or two anyway, but I can't give up on the engine because if I can't make such a simple engine run right what chance will I have on future cars..!

Thanks for all the info, I've certainly learned a lot from this thread.

If the idle is fluctuating iirc its a common fault with the hall sender unit On the distributor fails

I Haven't had time to read the entire post, but I had this with my AAU.

I never solved it myself but I had an air leak between both the Injection unit to rubber flange and rubber flange to manifold.. so this could be what's throwing the ECU. I know you may be thinking 'but my rubber flange is virtually new' or something like that.. They never seal!! I've had these leaks with all three polo engines I have dealt with (not knowing it at first). It is the cause of a slight- serious miss noticable most at the tail pipe. This, I worked out later, is because the injection unit has no control over the varying amounts of air getting in.. and the lambda readings must be like.. wtf??? I would say my next move is to buy genuine VW base flanges in the future (from tps for eg).. It must be that the aftermarket ones are crap.

replacing the blue sender helps also, if its as poor as mine was!

I'm having idling issues but it seems fine everywhere else. It will tick over at 900ish then suddenly drop, the idle control valve cures things temporarily but it all happens over again and sometimes cut outs.

I cleaned up the brass bits inside the distributor the other day. Also, i had a look at the plugs and found one to only be hand tight. Put all back together and it was perfect for the whole duration of the drive until i got home and it started playing up again.

The distributor looks like its been played with previously so i'm keen to get the timing set up. Mine does increase revs when unplugging the blue sender so that's a start. With the gaskets, there are two blocks. One between the throttle body/manifold and the other between the throttle body/injector plate bit. The gaskets look to be green paper. I don't see how they can leak. I tightened the four allen bolts today a bit - no different. Had a little play with the timing (no gun) and it seemed worse when advanced (only slight adjustments). After reading all this i'm wondering how to get it set properly. Just to add, i unplugged my Lambda and nothing seemed to changed.

What system should i look to get for diagnostics? Lots on eBay but don't know what is right.

Bought a timing gun yesterday and had a wee play. Didn't set anything as i didn't have the manual with me then but with the blue sender plugged in,t he timing movd around and with it unplugged it stayed still. Revving it up to 3000 RPM advanced the timing to about 30 degrees it think. Anyway, this might be of use to you. I switched the ignition off without plugging the sender back in. I realised what i'd done, plugged it back and started it. I read on here that if you don't plug it in before it's started it will log a fault.

After starting i tried unplugging it to see if it made a difference and it didn't! Bollocks!

Check to see if it was affecting the timing and it wouldn't stay still - just as if it were plugged in. I disconnected the battery for a couple of minutes thinking this would reset the ECU as i read on here also.

Started her up, pulled the plug on the sender and it was back to normal again - picking up revs and giving static timing.

Just for your info - i tested my temp. sender and got a reading of 260 Ohm at the point when the fan cut in which was just over 90 degrees on the gauge.

Having played around with SPI quite a bit lately I've re-affirmed my feelings that the Blue Temp must be unplugged before you set the timing. I tried both methods and doing it with it plugged in wasn't right at all, the idle speed became too low, and the timing was more retarded overall.

Also a good tip is to leave plenty of slack in the throttle cable. As long as you can still achieve full throttle with the pedal pressed to the floor the cable is tight enough. I initially had mine set up like a gt, with zero slack but it messed up the timing and caused cold starting issues, as the choke needed to pull on the throttle cable it wasn't adjusting correctly.