So I'm going to try and start logging knock as audio files to go along with data logs. All of this prompted by the sweet action going on with the false-knock work arounds.

I did a little hunting and there are two ways of doing det cans.

First is using actual ear muffs from shooting and routing a tube to the engine connected to the block.





The other way is electronic and uses a microphone.





So I'm thinking to use option 2. But instead of using a standalone mic, amp and hooking this up to head phones I want to use a computer mic and record it on the computer while logging.

So my plan is to take apart a computer mic and bond it to a metal plate. I'll then bolt the metal plate to my block. Then I'll simply record the audio file from the mic and use it to compare to my logs. That way I have the audio data saved instead of trying to listen to it while driving...

Thoughts.

 

Iv'e tried the above method, but was more dirty using a stethescope with an extension and a bit. Engine noise was too loud and hurt my ears. I might try something more refined some time as you plan.

Give it a try, might find something (like off-idle or part throttle knock sum is real)

 

This would work great for determining the actual frequency of the knock event and then maybe setting up a filter table. On some standalones the knock control is frequency based anyway so this will make setting those up easier.

 

I built a microphone. I think the cheap unit that I got didn't respond to the high pitch detonation pings. I could hear detonation, but the stock DSM knock sensor picked it up much better than my eras. I think that the copper tube with stiff nylon pipe into headphones in the best way to do it IMO.

 

I remember that - we used it on my ole turbo crx when tuning it, I will never forget the part throttle detonation I had that sounded like someone shaking a can of marbles....good times.

 

This is what I have done....

http://home.netcom.com/~bsundahl/kno...n.htm#Recorder

Except I hooked mine into the aux ports of the car radio, sure you can't record knock this way, but once you crank up the volume, it sounds miles better than having the ear muffs on, which I have also tried.

Use a lower value resistor if u find that the voulme isn't loud enough.

I have heard the engine start to detonate with knock counts as low as 2 using the above method, and wasn't able to detect it with ear muffs.

 

This is a great idea, I was actually thinking about doing this a few days ago.

 

This is exactly how I would have imagined a way to do it... May be with a different diagram (ampli Op) but definitively using the knock sensor signal...
If you want to get very good recording, get an external audio case connected on Firewire / USB2 port so you will be able to record @ 192kHz...
or find a musician' friend with something like this...
Good luck,
C>

 

The only thing that worries me about that system is if you are tapping into the OEM knock sensor you may be attenuating the signal. I would suggest a secondary knock sensor for your auxiliary system so as to not attenuate the factory knock sensor. Not to mention the OEM knock sensor is shielded and you really shouldn't be tapping into it because you will be introducing noise into the system no matter how careful you are. Good luck.

 

Here's a link with a bunch of sound files I found, so you can hear what it "sounds" like when you get knock.

http://www.theknockbox.cc/content/view/8/8/

I've got my sensor all ready to go, mic epoxied to one of my old prototype seat lowering brackets. Now time to test it out.

 

Yeah I use det cans like the headphone non amplified rubber tube type. Clear as day when you hear light det. Anyone talking about marbles and that is a lot of det I think.

I've often wanted to just throw a mic into the headphones to see what I could grab. To make any sense of how to programatically identify the knock from other noise will take a bit of effort I suspect.

 

FYI I learned how to make my det cans from Mark at MADevelopments in England and he likes to use a copper pipe bolted to the cylinder head off one of the intake manifold bolts. Works a treat.

 

Well I did some testing today. To me it seems the engine is not knocking, but it's being logged as knock.

The microphone works alright. I actually built a set of headphones as well, just haven't tested them yet. I plan on doing that tomorrow. I'll experiment with the headphones as well as maybe putting the mic pickup inside the tube of the headphone setup if it works better.

Anyway here is a log of what I think is false knock based on the audio that goes with it. This is simply revving at idle.



Audio file attached as well. This is done with the metal bracket mounted directly to the engine block.

 

What would knock sound like if there was valid knock? Can you hear it in the cabin without headphones also?