Nistune Forums

Hi ya,

I'm in need of some advice. I'm in the process of upping the boost levels to really start running some decent power and i've been using this kit to listen for det

http://www.justoffbase.co.uk/Electronic ... aley-VS007

But I've not heard anything out of the ordinary so far so before I push on and break something I want to be clear on what i'm listening out for.

Is the kit i'm using suitable? and there a safe way I can simulate det so that I can hear it for myself and recognise it?

I've been using an inlet manifold nut as a point to clip on to but all i've heard is a very noisy engine/timing chain from my rnn14 Pulsar.

Thanks for your time, Ben.

The link does not work. It links directly to the main site after a "not found" error.

soz, try that

Now the link works.
Pretty solution with 6 different inputs.
I tested my own built solution and record it in a mp3-file. I did not hear any unusal noise but you are right: The sound taken from the intake (this is where I put the additional knock-sonsor too) is overlayed from the noise of the valvetrain and other engine components.

Later I can upload my results so we can see if the noise from my engine is comparable to yours.

Thank you. That would be very useful.
Do you have any sound clips where det is present?

this seems to be like a pretty good solution for Knock detection...

where are you putting the clip ? show us a pic if you can...

best to try to get onto the block below the inlet manifold otherwise you will hear mostly cam and valve noise..

However knock ist a click or a tap sound... it is like a crackle as if there was a loose conection on the mic...

chris

Here are for a quick look 2 samples of the sound file. I can upload the whole file but its a little big, even I compress it to mp3 (10 MB).

The first file the engine is running at idle and then I began to drive.
The second file is where I am in the 5. and give some power to the pedal.

I can later link it to the correct thread.

The sensor was bolt onto the intake (look at the red circle in the attached picture).

If anyone knows a better place to simply put the sensor on and to only record the important things I would be appreciated.

PS: I hope there is NO det in my soundfile.

I will later go on search. In the past I found some samples with det somewhere in the Inet...

dont forget the fact that you can look at plugs and also the piston crown at TDC.. you can tell that they are pitted... this is long term damage however..

The plugs normally get pitted or the electrodes break off and go into engine under really bad conditions.. as for the pickup go onto the cylinder bore not the head as you want to listen to the piston vibrate.. not the air above it.. if the engine is pinging it is the piston making the sound.. and it never goes above the head gasket..

From what I have been told my tune does not have radical timing.. as I did before and only had fuel problems.. always had to octane boost cause the fuel in my area does not seem to be very consistent..

The guy who tuned it did the head in the engine bay on the dyno... not for the faint hearted..

Anyhow dont forget the search...

viewtopic.php?t=847&highlight=knock

some samples

viewtopic.php?t=847&highlight=knock

the topic..

micha ... nothing going on in those recordings however it is where your peak torque 3000 - 4500 rpm depending on setup.. is not so much low load / hi / low rpm driving.... find a hill and a tall gear give it a boot and dont break traction and then listen.. best not done from drivers seat.. get a mate to drive one that you trust..

chris2712au wrote: micha ... nothing going on in those recordings however it is where your peak torque 3000 - 4500 rpm depending on setup.. is not so much low load / hi / low rpm driving.... find a hill and a tall gear give it a boot and dont break traction and then listen.. best not done from drivers seat.. get a mate to drive one that you trust..

Thanks for the advise!
Do you know a better place to easy screw the knock sensor on, so there is less noise from the head?

Now I have linked the picture and the sound file to the correct thread too.
If someone can tell me a good place to upload the whole 10MB-mp3-soundfile I will do.

no probs as you can hear in that demo the sound of pinging is noticeable..

as for the thread doesnt matter that much just good if your digging for information like sound files and the like..

this is where the stock one is.. on my s15..



pretty sure it is the same on the s14... See if you can mount yours near the stock one or somewhere on the block itself under the inlet manifold.. normally i just pay for dyno time as full load tests like in my boost tuning its just a little too dangerous on the roads where i live..

some people have done research into using the stock sensors.. I think you can turn the knock sensing off in the flags in the ECU or do a knock copy.. and just interface a system to the stock unit..

Ah, thanks. I will look if I can mount the additional sensor in the near area to the original one.

Here is an engine knock calculator which says that the 86mm pistons from the SR20DET will produce a knock signal at a frequenzy of 6.7 kHz: http://www.phormula.co.uk/KnockCalculator.aspx

Does this means that frequenzies below 6.7 kHz are not important?
Makes it sense to apply a hi-pass-filter to only get the frequenzies above 6.7 kHz in my test file?
Or it is better to amplifiy the 6.7 kHz-Range with an equalizer function?

I wouldnt go too much into filtering the signals through a bandpass filter as once you know what you are looking for you can distinguish it quite well.. however bear in mind that audible pinging no matter how faint is not gaining anything in terms of power as soon as you can hear it some people say 2 degrees others say 5 degrees in timing and thats a lot to back it off... it still isnt as accurate as the torque method... in my opinion anyway.. audible pinging is a good guide for low meduim loads but the high loads are best done using the torque on the dyno..

So the conclusion is that I do not come around to do the last bits on a dyno.