Leeroy

I currently have an MY97 which suffers a bit of piston slap on cold mornings. Does anyone know if the knock sensor can be inadvertanly triggered due to this, as if it could, ignition timing would also be retarded to a maximum.

From my experience of knock sensors, to test its operation a scope can be connected across its terminals and a hammer used to lightly tap the cylinder block, while the waveform is monitered. This 'hammering' will usually be sufficient to activate the sensor, and hence it can be effectively tested.


Can anyone shed more light on the subject???

Cheers.

ssubaru

hi had this problem on my mk2 rs turbo ,solution was to fit mk1 ecu without knock sensor to stop ignition retard as this is also done by turbo systems it was deemed to be safe, but i dont know if you could change ecu on scoob to sort same problem of ignition retard from engine noise
ssubaru

Leeroy

Not really the answer I was hoping for as this seems an expensive solution.

Anyone else with a 'slapper' have an opinion??

Leeroy

bttt

Leeroy

ttt, any ideas??

Leeroy

TTT

Bob Rawle

Yes, its possible this could happen, the knock sensor is just a sort of microphone and therefore will pick up any "noise". the ecu has filter ccts built in that should then clean up the signal but ... I have known an exhaust vibration retard the ignition so its feasible that piston slap could also do it if it was bad enough.

Leeroy

Thanks for the reply m8. Has anyone got a suggestion of how to solve the retarded ignition (other than disconnecting knock sensor!)? I'm thinking the only 'cure' will be to sort the slap out, which is a tad expensive.

Also, does retarding the ignition alter the A/F ratio?

Cheers.

john banks

If the ECU detects knock (or something sounding like knock) at least on MY99/00 it will also richen the mixture if it is bad enough. Subarus run quite a lot of advance in their maps and rely on the knock sensor. You need to confirm this is what is actually happening on a Select Monitor to decide further I would think.

Bob Rawle

Its possible that re-positioning the sensor will have the effect of "detuning" its sensitivity to this, some Subaru dealers have done that to solve this sort of problem. The only way is to try it and see.

Leeroy

A silly question, but do I need to use a Subaru dealer for the select monitor, or is there another way? I'm assuming the select monitor is not used to pick up faults (codes), as in this case the knock sensor will be working correctly and no fault should have been recorded (I think).

If my knock sensor is under the impression the engine is detting (even though it isn't) it is of no use at the moment as the timing will be fully retarded (although the onset of detting will no doubt be redued!). Its only purpose at present is to retard the ignition and reduce outright performance/efficiency. Also, could this be responsible for the slight light throttle hesitation I keep getting, as I think the ignition is momentarily advanced during acceleration (knock sensor permitting)?

Cheers people.

john banks

The Select monitor gives you a live trace of knock correction as you drive. This is a way of seeing what the ECU is interpreting of all the noises it is getting through the knock sensor. Otherwise you are just guessing that this is happening.

Leeroy

Where can I get access to a select monitor? Are Subaru dealers the only answer (I'm not overly impressed by my local dealers)?

Cheers

john banks

Some Subaru specialists also have them.

submannz

Yes you are exactly right about this. There is a mod you can put on the knock sensor to reduce the sensortivity because most of the older ones especially where too sensitive causing bad performance.

I will try to dig up the documentation for you.

submannz

Here it is.

http://www.geocities.com/hobiegary/hesitation.htm

Leeroy

Cheers for that. Is anyone using this thing then?
What is the likelihood of pre-ignition when using a good grade of fuel (Optimax)? I personally have never had a vehicle that 'pinked', is it common on Scoobies?

I tested the knock sensor last night with the ignition off and found very slight taps with a spanner on the cylinder block had gretat effects on the sensor. The resistance to ground seems to rapidly reduce during 'tapping' and oscillates quite a lot before eventually settling at around 560k ohms. It does appear to be very sensitive and I am absolutely convinced that piston slap will activate the thing (any vibration around the sensor will probably have the same affect).

Is it worth removing the sensor until the piston slap is sorted? Also, does anyone have any tips on No. 4 piston removal, such as tools required, costs etc.?

Cheers in advance.

Jan Shim

You can get yourself an aftermarket Japanese ECU monitoring tool made by TECHTOM and then see interesting activities. Some report I made about KnockLink, light flywheel and the Techtom MDM-100S can be found here. I have a spare MDM100 for a WRX/STi '98/ver 4. Please email me janshim@brunet.bn if you want more information.

http://www.i-club.com/forums/showthread.php?s=&threadid=115948&highlight=techto m

Leeroy

John,
I was also under the impression the ECU monitors more than just the knock sensor with regard to retarding the ignition. It must surely take into account engine RPM/load as well as the crank/cam sensors, and if this is the case the knock sensor signal will only have effect when a particular piston is due to fire. I intend to monitor the ignition timing as a quick check, and then get booked in with Subaru next week.

Cheers.

Andy.F

Leeroy

Not sure about your MY97 but I suspect my own MY95 WRX doesn't even monitor the Knock sensor when it's cold !
I think this because i have it disconnected and the ECU does not report a sensor fault until the engine has heated up !

Bob Rawle

The knock detection is active all the time and the ecu cannot expect det, it has no way of knowing that it could occur, I suspect that your sensor is going "microphonic" and, as such, is becoming much more sensitive to engine generated noise.

Take it to a friendly dealer who will put the Select Monitor on it, then you will actually see the knock retard etc happening if that is the case, of course no faults will show, the ecu is reading an impedance from the sensor and thinks its ok, it may be but ... it could also have become very sensitive.

Since the sensor is a "microphone" then it will react to tapping, thats a bit mis leading as it is being grossly activated in that way.

Bob

Jan Shim

The Subaru WRX/STi ECU employs an air-fuel ratio learning control system that is constantly memorizing the amount of correction it required in relation to the basic amount of fuel to be injected (the basic amount of fuel injected is determined after several cycles of fuel injection). This is a quote from the workshop manual.

IGNITION system control, on the other hand, is based on feedback from various sensors eg AFM, engine coolant temp sensor, crankshaft sensor, camshaft position sensor, etc to determine operating condition of the engine. From this feedback, it then selects an optimum ignition timing stored in the memory and immediately transmits a primary current OFF signal to the ignitor to control the ignition timing.

The ECU constantly provides the optimum ignition timing in relation to output, fuel consumption, exhaust gas, etc according to various operating conditions.

Jan Shim

The Subaru WRX/STi ECU employs an air-fuel ratio learning control system that is constantly memorizing the amount of correction it required in relation to the basic amount of fuel to be injected (the basic amount of fuel injected is determined after several cycles of fuel injection). This is a quote from the workshop manual.

IGNITION system control, on the other hand, is based on feedback from various sensors eg AFM, engine coolant temp sensor, crankshaft sensor, camshaft position sensor, etc to determine operating condition of the engine. From this feedback, it then selects an optimum ignition timing stored in the memory and immediately transmits a primary current OFF signal to the ignitor to control the ignition timing.

The ECU constantly provides the optimum ignition timing in relation to output, fuel consumption, exhaust gas, etc according to various operating conditions.

ssubaru

hi andy f , why dont you wire 560k resistor in line instead of disconnecting sensor and getting fault light on or would this not work
ssubaru

Pavlo

I have had a similar problem with my Legacy. I am convinced that the very noisy hydraulic tappets I had problems with have been putting the knock sensor into overload.

I drove 360 miles to cornwall for holiday, and at times I was reduced to 5 psi boost. When I got there, and was no longer going quite fast 3-8am (avoiding all traffic, 2 stops btw), I notice major tappet noise.

I looked for ages checking the boost soleniod, checking wiring, all to no avail, still 5psi. Did an ECU reset and no problem. Few days later, major tappet noise, 5 psi boost, tappet noise goes, 11psi boost (on the sae journey)

Aha me thinks, remember last owner had it serviced with "fully synthetic oil" oil seems thin as hell. Trip to halfords, 8 litres of Mobil 1 15/50 and an oil filter, and presto, no noise, no boost problems.

So to cut a long story short, weird vibrations DO trigger the knock sensor, and if bad your ECU WILL punish you.

Paul

dowser

Paul - punish, or protect?! Sounds like you were lucky

Richard

Pavlo

Not that lucky, the ECU just protected me from boost when I was getting noisey hydraulic tappets rather than engine knock. The tappets are notorious on the legacy. I am talking about noise that isn't present during idle but came in at about 2000 rpm when light throttle was applied in nuetral.

The worst bit is sometimes the tappet would come unstuck, pump up properly, no noise and whooosh, back with boost, 30 seconds later, no boost. But I think the ECU must remember the boost condition for a limited period of time, then go back to normal, unlike the ignition which (on pre 97) never comes back unless you reset the ECU.

Paul

Leeroy

Pavlo, I may well try going back to Mobil 1, as I don't know what Subaru put in at my last service. However, they did ensure me that my scoob was running perfectly, although they seem like a bunch of muppets to me!

john banks

I would strongly advise against removing the knock sensor since the car relies on its operation and can run a lot of advance and will keep advancing (to a point) if the knock sensor is quiet. It makes sense to confirm it is actually retarding before adjusting the sensor. Just because the knock sensor has a signal does not mean the ECU will retard. I don't think anyone knows exactly what filtering goes on and whether it only listens at intervals when det is expected in each cylinder or not. What you are proposing to do in desensitising the sensor sounds good if it is actually oversensitive, but otherwise a bit daft, so finding out is essential.

[Edited by john banks - 6/7/2002 3:20:48 PM]

Andy.F

May be worth a try ? Not quite sure what the ECU expects to see under running conditions from this sensor though. I was going to mount it on a 'quiet' part of the engine.

Quote ""The knock detection is active all the time"" Bob, is this true for all years of ECU ? Have you proved this ?
I don't understand why my ECU does not report the fault until it is fully warmed up, it reports the fault only after the temp gauge hits middle. (tested on many ocassions)

[Edited by Andy.F - 6/8/2002 10:30:48 AM]