Been considering building one of the homemade knock sensors, was wondering if anyone had some tips and tricks to share? Going to keep this to a small budget of around $30-40. I like a cheap radio slack micro phone device modified to use a battery post clamp, this way I can move it around to listen to various parts of the engine, or frame/chassis. Reason why is the GN is picking up a lot of false knock, no amount of tuning will mitigate the issue. I.E. no boost, low boost, low timing, etc. Still gets knock at the same rpm every time, Im thinking it may be the noisy lifters or the down pipe touching the frame when the engine torques up in the engine bay. If I can build one and route the signal to my laptop, I can probably use the EQ to filter out some of the noise too?

Looking around for ideas...

http://www.realhomemadeturb...for...?topic=22257.0

http://www.autospeed.com/cms/article.html?&A=0353

quote Originally posted by sleevePAPA: Still gets knock at the same rpm every time, Im thinking it may be ... the down pipe touching the frame when the engine torques up in the engine bay.

I think you may be onto something. Why not fix the likely problem before trying to engineer a work-around?

[This message has been edited by Marvin McInnis (edited 08-06-2014).]

quote Originally posted by Marvin McInnis: I think you may be onto something. Why not fix the likely problem before trying to engineer a work-around?

I wont be using this as a work around, more of a nifty project that will help me narrow down the noise. Then I could use this device to listen to the engine for added safety, if the knock sensor should fail to report knock. Ive been reading that some folks can hear audible knock(with this device) even when the knock sensor doesnt report any. The idea is to build the electronic det can and port it into the laptop or the head unit and try to use the EQ to filter out noise from the lifters and other normal noise makers. Also I would like to find a software download that can identify the frequency range of certain sounds for added filtering, then I will attach the MIC to various parts of the engine or chassis. Im thinking if I tap on parts of the exhaust and frame and find the average frequency, then I could use the software to quiet down every other sound except for the one im searching for. I think you see where im going with this. If the cheap project works out well then I will spend more time/$$$ to build something better and use a modern wideband AC generating knock sensor, or even a strain gauge sensor. Down the road I may also use this to tune the exhaust using 1/4 wave resonators, just need to figure out what frequency is the most obnoxious and attenuate it.

Where can I find a software download that can identify and filter out certain freq bands? Im no sound engineer so humor me lol

quote Originally posted by sleevePAPA: Where can I find a software download that can identify and filter out certain freq bands? Im no sound engineer so humor me lol

I don't know "where," but probably one of the first functions you'll need is a Fast Fourier Transform (FFT) to remap the transducer signal from time domain to frequency domain. Only in the frequency domain will you be able to identify the unique "acoustic fingerprint" (to mix metaphors) of pre-ignition or detontaion vs. other mechanical noises originating within the engine. FFT can be implemented in software on a general-purpose microprocessor, or it can be implemented in dedicated hardware (e.g. a DSP). The FFT aperture (i.e. the sample time) has to be long enough to capture an entire pre-ignition event, but it has to be fast enough to adjust ignition timing and/or fuel metering before the next cylinder fires. The project you've taken on is not trivial, but learning is always good.

[This message has been edited by Marvin McInnis (edited 08-07-2014).]

Yeah I dont want to get that far into the weeds just yet, besides, I dont write code yet.