Knock and Inlet Air Temp sensor testing

[S2central.net]

I'm just thinkin out loud here - haven't thought it through, but shouldn't imagine we need to sample anything faster than twice the max operating speed of the engine. Say round up to 8k rpm and double that to 16k samples per minute which is 300Hz rounded up again. Sounds like plenty fast enough and with a simple 8-bit ADC you end up at 2400bps which is perfect for serial RS232 comms... Then you just need some s/w to log the data !

Paul

[Flat on my back]

Crikey...just let me build the circuits first...then we can do the logging.

I'm fine with the 'lecky bits' but my software is only past 'basic'...

[S2central.net]

Theres plenty of aftermarket datalogging widgets and ADCs for PCs - that is much less interesting than a wee circuit board that can provide simple visual indication of knock.

Cheers,

Paul

[Flat on my back]

Picture of output of rpm and no.1 knoch sensor.

Top trace shows the o/p of the VR rpm sensor at idle. Scope was set for a dc i/p, 5v per div. It is an o/p of the teeth on the flywheel.

Remember this is NOT the ignition or TDC sensor.

Bottom trace is the o/p of no.1 k. sensor with the scope set for an ac i/p, 20mV per div.

I'll try and get a picture of the k.sensor o/p in relation to firing cycle signal, maybe hall or TDC.

[S2central.net]

Can only find this from a 3B manual right now - moving house sucks for things hiding in boxes -

99% sure the IAT is common to ABY - spec is for 450-650 ohms at 20 celsius.

I am sure I have seen a graph somewhere - its an 'PTC' device with a linear profile (from memory) so you should see resistance go UP as temperature goes UP.

HTH - Paul

[S2central.net]

Tada - gotta love Scott Mockry

http://www.sjmautotechnik.com/trouble_shooting/20vairt.html

[Rusty]

Is there a specific S2 manual or is it covered in the 80/80/coupe+20V book? OR do you need the S4/S6 manual for the engine side of things?

IAT is curently having a bath in IPA, will check the resistance later today when it's all clean and dry

Knock sensors could have done 155k so guessing that they and a lot of other sensors for that matter are going to be rather tired and in need of replacement

[Rusty]

374 ohms at 15-20 deg C. Seems a little low compared to the graph. Hoping to get a replacement 2nd hand one tonight so will see what that reads. Part number is correct at 034 905 379B and it's made by Hitachi. No obvious cracks to the solder when viewed under a 10x magnification microscope.

Just thinking aloud, if the ECU thinks it's -50 deg C then it might get a bit of a shock when the air is actually +20 deg C! Can the ECU cope woth this mis-read other than using the knock sensors to retard the timing?

[S2central.net]

Unfort - all the knowledge is spread across multiple manuals - you don't get ALL of the info on the ABY engine say in the same document. Some directly refer back to 3B info and some of that refers back to RR for example.

Thats what S2central.net attempts to do - by filling the gaps in all this crap - it has fallen way short in many areas as I just don't have the time anymore. Thankfully sites like Scott's are very useful with info like that tucked away.

I agree that your knock sensors could well be at fault - its only if they go o/c or s/c that the ECU can throw up a fault code - it won't log a fault if the knock sensors are sending false knock indications to the ECU.

When the ECU detects sufficiently high levels of knock it retards ignition by a whopping 6 degrees so its a pretty noticeable retardation. HTH - please note there is a requirement for accurate torque values when fitting new knock sensors - do NOT just whazz them in ! I'll look it up - though I bet its on Scotts site already !



Paul

[S2central.net]

374 ohms at 15-20 deg C. Seems a little low compared to the graph. Hoping to get a replacement 2nd hand one tonight so will see what that reads. Part number is correct at 034 905 379B and it's made by Hitachi. No obvious cracks to the solder when viewed under a 10x magnification microscope.

Just thinking aloud, if the ECU thinks it's -50 deg C then it might get a bit of a shock when the air is actually +20 deg C! Can the ECU cope woth this mis-read other than using the knock sensors to retard the timing?

Andy - if you are careful, you can try to fool the ECU with a fixed resistor in place of the IAT. Technically its a bit naughty as you will be fooling it into thinking its 20 celsius instead of 40 celsius and therefore the timing may well be excessively advanced... But the failsafes will come into effect if knock is detected.

It would be good if you can test the IAT after a run in the car - give it a few bootfuls of max boost then bring it to idle soon after for a test -this would be nearer to 40 celsius.

If your IAT value at 40 celsius is too low then it is trying to tell the ECU that air temps are lower than reality and you will get too much advance and possibly knock thereafter giving you that retarded feeling.

I'll test mine tonight for comparison - for all I know mine is FUBAR also - but at least I can sanity check the data with VAG-COM. It was pretty accurate a few years ago when I was on and off the rollers.


Paul

[Jamo]

Andy - if you are careful, you can try to fool the ECU with a fixed resistor in place of the IAT. Technically its a bit naughty as you will be fooling it into thinking its 20 celsius instead of 40 celsius and therefore the timing may well be excessively advanced... But the failsafes will come into effect if knock is detected.

It would be good if you can test the IAT after a run in the car - give it a few bootfuls of max boost then bring it to idle soon after for a test -this would be nearer to 40 celsius.

If your IAT value at 40 celsius is too low then it is trying to tell the ECU that air temps are lower than reality and you will get too much advance and possibly knock thereafter giving you that retarded feeling.

I'll test mine tonight for comparison - for all I know mine is FUBAR also - but at least I can sanity check the data with VAG-COM. It was pretty accurate a few years ago when I was on and off the rollers.


Paul

S2central in ebay chip mod shocker!

[Rusty]

Cheers Paul. I'll check it after I head over to Bristol this evening. Should be a fair bit of town driving and traffic so expecting high temperatures.

I don't fancy playing games with the ECU, a new IAT is just over £60 so no point messing about.

A combination of false , low temperature reading and knock sensors being a bit over-eager could result in the hot running problems I experienced over the weekend. I had to back off at WOT in 2nd as the car lost all acceleration but still seemed to be producing 1bar / 0.8 bar - I would bet on knock sensors cutting the timing. If the car is running full advance rather than low-level retardation for hot weather then the knock will be fairly large and would cut the timing a lot more brutally than if the knock sensors take it back 6 deg.

All han waving but hope to get it sorted soon enough with a combination sof sensors...

New knock sensors and a 500ohm @25 deg C IAT will hopefully help with the hot running. orst cae is £130 for all new sensors so not the end of the world.

[Rusty]

Seat of the pants lunchtime testing:-

Outside air temp display was 17 deg C. Immediately after a clear run on part throttle, resistance of the IAT was 372 ohms. After negotiating the road into work, parking and letting the fan stop the resistance was 382 ohms. Certainly nowhere near the 500 required!

[S2central.net]

Aye - you are off the scale - the ECU is being told that the air temp is MUCH colder than reality... So I think you are getting too much advance - and maybe too much fuel although I think the coolant temp sender dominates that one...

You should check the water temp sender also - just in case. Its at back of cylinder head... Below the PCV pipework - bit of a ****** to get at to test unless you have a 2-pin connector to hook up (or a breakout box at t'other end). Thats an NTC device - linear - resistance goes UP as temperature goes DOWN.

LOL at Jamos comment - quality bit of observational humour !


Paul

[mcandmar]

I'll test mine tonight for comparison - for all I know mine is FUBAR also - but at least I can sanity check the data with VAG-COM. It was pretty accurate a few years ago when I was on and off the rollers.

Yep vagcom gives you a nice figure in degrees C too. In our AAN i've seen it hit 48c after a few WOT runs and then letting it sit at idle. A cold start should read fairly close to amibent temps.