Rover V8, knock sensors and mechanical noise

[cinquecento]

I have a 96 Griff...however I think my question applies to any (TVR) RV8. It's a Megasquirted Griff 500 with S/C Power supercharger..and I have reduced my advance table on boost to control pinging. Aquamist water injection has helped to control the detonation...and I have just successfully installed the MS3 internal knock sensor module into my ECU..basically as an insurance policy.

How mechanically noisy is the RV8?...noise that is likely to affect the knock sensor. Interestingly, I pick up low levels of knock noise under 4000RPM ..noise that substantially increases above 4500 (definitely not pinging). Also, when I back the throttle completely off at high RPM (above 4500) there is a single, sharp knock spike.

There's a fair bit of tuning I can do on the MS, but I'm curious if the RV8 would be considered mechanically noisy?

[350Matt]

where have you mounted the knock sensor?
as the best place should be on the block in the middle, ideally 1 per bank

[MisterT]

I remember reading somewhere, although can't recall whose expert opinion it was or where it was published (MA perhaps?) that the RV8 is mechanically noisy because of the alloy block. Don't know if this is based on scientific principle or if in fact it is the case.

[cinquecento]

Thanks,....At the moment I have a single knock sensor bolted to the protruding boss on the block near cylinder 1. The boss sticks out abt one inche and is abt 2 inches above the cross bolted hole that would have been drilled if the engine had been x-bolted

[anonymous-user]

Robust acoustic knock sensing ALWAYS requires crank angle windowing as well as frequency discrimination in order to improve the signal to noise ratio, and to be able to pick out the knock from the background engine mechanical noise.

The vast majority of after market ecu's do not do this, and hence can really only ever pick up GROSS knock at low engine speeds.


The Rv8 is not particularly noisy, but being a push rod engine does have some extra valve train noise coupled to the block compared to an OHC engine type

[cinquecento]

Thanks..frequency configured for 6.1khz and between sensing between 0-25 deg atdc.
Off for a drive today...will be interested in the datalogs.

[anonymous-user]

cinquecento said:

Thanks..frequency configured for 6.1khz and between sensing between 0-25 deg atdc.
Off for a drive today...will be interested in the datalogs.

You might want to plot out the cam events in your engine an see if you can fit in a window that avoids those events (might not be possible, depending on firing order etc), also , typically, you want the knock window to open some time after tdc to best catch the knocking pressure oscilations, around 10 deg ATDC for example. At 6.1Khz, you can work out how many cycles a given window length at a give rpm can capture (the more cycles the better of course to allow the frequency binning to have the highest descrimination.