knock detection: earphones vs knock sensors

[sambb]

Just wondering if any of the guys that have had their cars on the dyno could chime in here. I'm curious as to how sensitive our knock sensors are. My mech/engineer/tuner guy has said that he's seen both situations (on other cars) where knock sensors were doing nothing yet he could hear audible pinging, but also had situations where an aftermarket knock sensing unit was picking up pings that he was struggling to hear on the earphones. Just wondering if anyone out there knows what kind of correlation our knock sensors/timing pull have with audible/earphone detection.
sam

[zei20t]

on the dyno where you can have the bonnet open, have someone stick their head in there. its probably a little too hard to do that on the street, unless you can find someone small enough and knows what to listen out for.

my nissans knock sensor stopped working around 4k, so someone had to listen out while it was on the dyno.

[Lucas_R]

What car and engine are we talking about here?

If its a modern car you could use VCDS to monitor the engines a/f ratio and knock.

[Notso Swift]

Knock sensors need to be tuned as well
OEM obviously set their own parameters but that would be based on the OEM tune, knock outside a frequency may be detected but nothing would be done about it because the computer says no.

[sambb]

engine is a 20V 1.8T in a 9N3 Polo. Having another chat with my mech he mentioned that in his experience knock sensors are more accurate on iron blocks vs aluminium blocks. Apparently since aluminium conducts sound better than iron, the origin of the sound and whether or not the frequency the knock generates is getting mangled by other interference is a big factor. As a result you are either prone to false positives or the threshold the ecu picks up at has to be increased which can result in false negatives.
He thought that since my block is iron and bosch kind of knows its stuff, that if I'm getting 0.8-3 degrees of pull then it is very likely due to legitimate solid knock, and not due to an over reactive/over nannied timing pull system, which is what I thought it was likely to be.
I get what you're saying Notso Swift r egarding the fact that the ECU would be deciding on what frequencies constitute a knock intensity worth worrying about with respect to the standard tune. Now that the stock boost has been more than doubled, surely my knock would sound completely different to what it would have at 10psi and also be far riskier for the engine when it does happen at the far greater cylinder pressures. My mech sort of said the same thing and encouraged me to take timing back out of the engine.
I'm on a flash tune. I had been through the usual flash tune process where you advance global timing until you see some timing pull and come back from there. There's a lot of stuff out there saying that the bits of 3, 4 or even 5 degrees of pull that results from this is quite acceptable but that never really sat well with me. I had mine set with little bits of 0.8- 3 degrees here and there. After taking my mechanics advice I pulled 2.5 degrees back out of the global timing advance setting that I'd been running with until I could see no timing pull anywhere. I had never really logged the MAF reading at a timing setting that low previously so was pleased to see that I lost no top end whatsoever and picked up midrange despite the fact that the weather was much warmer.
That's what got me thinking about audible (with headphones) vs knock sensor detected knock. If retarding the timing until no timing pull registered in vcds resulted in gains, I just wonder if there's likely to be still some knock there that headphones could spot.
I think you'd really have to log actual knock sensor voltages and timing pull together to know what the threshold voltages are, but that would probably be dynamic at every load point and hard to work out. You'd then have to do the same between std versus modded engines to know whether the ecu becomes underactive, overactive or stays the same with respect to how it responds to knock.
Only thing is now that I've retarded by quite a bit, the water injection causes flat spots here and there. That makes sense because water injection needs timing to work, but the problem is now that I can't use it in its current guise, I loose a valuable bit of charge cooling. And its hard to get less water into it when you're already on aquamists smallest jet.

[parso_rex]

You need a proper explanation but lack of time kills me atm

The knock retard you see is not an actual retard due to actual knock. Think of it being an acceptable amount leading up to the point where knock will begin to occur as decided by the ECU. Typically you won't hear any knock till you see near 9 or more degrees 'knock retard'. Optimal Torque(hence power) MTBT is achieved with a slightly higher 'knock retard' value than what you have indicated so its on the safe side if anything.
I have my own desired values for acceptable 'knock retard' that I will use on a customers car and this varies with tune providers - some even seem to be happy with audible knock/ping it would seem
It is right for your mechanic to assume knock is occurring just due to the way the parameter is worded but its not quite the case.

got to run........

[sambb]

so the knock sensors are able to pick up proper knock but also pick up on pre knock or hints that it is approaching. Bl%%dy hell Bosch are good at what they do then. So for a 1.8T that's internally stock save for some exhaust manifold porting and wastegate porting , K03s w dump pipe, 1.4bar peak...... what is an acceptable amount of timing pull (ball park of course). Based on what you're saying possibly I was never advanced enough or at least not enough to be running water.
thanks

[parso_rex]

Aim for 4.5 degree timing pull if its an option for you

[sambb]

yep I can go to 4.5 degrees max in spots. If when I'm running the water is it the same deal - 4.5 degrees max pull or does water change things?

[parso_rex]

Can't say too much more as it gives secrets away, do it on a dyno and look at the torque values, not derived but raw or motive force and then you will have your answer