I was wondering what type of signal the Knock sensor sends to the ECU, and how the ECU interprets knock to retard timing.

Does the knock sensor signal to the ECU start with a reference voltage (12v or 5v signal) durning normal (no knock) and as the knock becomes worse, does this signal drop down closer to 0v (ground)? Or is it opposite, starting with nothing and then raising voltage with more knock?

Or am I just way off.

The ECU program I am currently using, disables the knock sensor to gain maximum advance earilier. I would like to know how the ECU would normally interpret this signal. I am going to build a gauge to hopefully tell me how much knock the engine is producing, to make necessary adjustments.

If possible, could someone explain basically what type of signal this is and possible reference voltages and possible ranges of operation for the sensor, and what voltage does the ecu interpret as extreme (maximum) knock.

Thank you in advance.

 

There was a nice big thread about this not too long ago.

I think the knock sensor is basically a microphone, & the signal processing & discrimination happens in the ECU somewhere. Lots of others think differently, but nobody posted back saying they've actually looked at the signal with an oscilloscope...

You can put a DMM onto the sensor wires. But that doesn't prove anything if you don't ALSO look with a scope. I wish I owned an oscilloscope.

 

It is indeed a microphone when you get down to it.

 

Yep, when it comes down to it, it is just a microphone.

PM "BigMoose" as he did a test awhile back on knock sensors using an O'scope. I tried searching under his username, but it doesn't bring up the thread, as I guess its too old.

 

I understand the knock sensor theory of operation, but I was just wondering on the actual signal received by the ECU.

If it's a simple as a "microphone", then would the signal be an AC type signal where amplitude is key, or does the ECU interpret this AC signal and convert it to a digital pulse and then use pulse width as a determination of the intensity of knock?

That is more the question I am having.

Basically I need to know what type of signal the knock sensor outputs. I could use an O-scope and that is probably what I will do but I figured I'd ask to see if anyone has done this to save me the time.

Thanks to all who replied. I'll do another search and see if I can find out anything else.

If/when I do the oscope testing, I'll post results.

IF

 

The KS is a microphone listening for knock, and when an engines knocks it does so at a certain frequency. I'd have to reason that along with amplitude its the frequency that would seem to matter more, or at least to the ECU doing the processing.

But as to what the shape of the waveform looks like and what the ECU does with this signal, is not something I know. However, BigMoose, as he has done a test with KS's, would know these answers more concretely....thats why I steered you his way.

 

Thanks for all the information.

 

Here is a reference to a thread on another board where I shared some scope traces of knock sensors and a spectral density plot of knock during a boosted run:

http://www.theoldone.com/forum...knock

Enjoy.

 

simply put, the ecu is looking for a reference resistance. as the mic diaphram is pushed in/out resistance varies. the ecu knows that resistance X = im knocking .. pull timing. it should be noted that hondas knock code is very weak and troublesome if you're not running a stock bottomend, which I'm sure has been discussed in this forum before.

 

i would have to say the ecu interpets the signal as a impedance.....from there it can measure a preset frequency that the factroy has set as "knock" some cars being more sensitive than others ......

this makes me wonder if the ecu has to see a variance in the signal or can it see a steady signal say from a resistor...therefore never reading "kock"at all ive wondered about this for some time


=edit= opps didnt see your post above mine ...ya what he siad

 

oh and one more thign ......what would the ecu think if the knock sensor was isolated form the motor ....say in an inclosed box that would allow it to still "hear"but kind make it less sensitive ....

 

ttt id like to learn more about this

 

my f22a1 doesnt have a knock sensor and my timing is static. I was hoping on using the h22a1's knock sensor as it just screws into my block but ofcourse my ecu wasnt designed to work with it, so I'm going to use an aftermarket knock detection system because I dont know how to interperet the signals given off by the stock sensor.

 

heh, you're worred about knock on a motor w/8.8:1 compression?

anyway ya.. you could put a knock in a lil box or the like, but the easier solution is to just disable it in the rom. I'd rather tell the ecu to shitcan the knock sensor than to try and put a bandaid on the problem.

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by GimpyAccord &raquo;</TD></TR><TR><TD CLASS="quote">heh, you're worred about knock on a motor w/8.8:1 compression?

anyway ya.. you could put a knock in a lil box or the like, but the easier solution is to just disable it in the rom. I'd rather tell the ecu to shitcan the knock sensor than to try and put a bandaid on the problem.</TD></TR></TABLE>

its not N/A

 

whoopdie ****?

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by GimpyAccord &raquo;</TD></TR><TR><TD CLASS="quote">whoopdie ****? </TD></TR></TABLE>

well you asked, smartass.

 

Ok can someone tell me this

I am running an ls vtec. My wireing harness (gsr) calls for a knock sensor yet the ls block does not a spot for the sensor to screw in. This is throwing a CEL code23. How can i eliminate, bypass, or somehow make this sensor work ???

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by redlinegsr &raquo;</TD></TR><TR><TD CLASS="quote">Ok can someone tell me this

I am running an ls vtec. My wireing harness (gsr) calls for a knock sensor yet the ls block does not a spot for the sensor to screw in. This is throwing a CEL code23. How can i eliminate, bypass, or somehow make this sensor work ???</TD></TR></TABLE>

https://honda-tech.com/zerothread?id=383198

 

what ECU are you running?

 

Hey Big Moose... thanks for the link... I'll take a look at it.

 

putting a knock sensor in a box is ghetto, laff...

here:

http://dolemite.dyndns.org/scr...k.bin

thats a p72 rom, knock disbabled but otherwise stock. if anyone needs a p30 w/just knock disabled or anything else changed (goes for the 72 as well) let me know, i'll post it.

 

ahh hell with it, p30 stock except no knock:

http://dolemite.dyndns.org/scr...k.bin

keep a' knockin but you cant retard

 

A KS is simple a piezo electric device which creates a voltage as "vibration" occurs... On an oscilloscope the output from a KS will look something like alternating current (VAC)... I personally avoid running a KS whenever possible... If you can run a P05/06/07/28, that's just fine with me Latez

 

For anyone interested in reversing this, that might not know. The Knock Sensor is controlled by a duagtherboard inside the ecu case. I'm sure some running of part numbers will get you an answer if it's an A/D setup or what.

also, can someone add an authoritive list of which ECUs have knock sensor? Is it all vtec motors?

-PHiZ