The Central Hub for DSM Community and Information

For 1990-1999 Mitsubishi Eclipse, Eagle Talon, Plymouth Laser, and Galant VR-4 Owners. This is where the DSM platform history is documented and archived. Log in to help us in our mission, and to remove most ads from the browsing experience.

How is 2G knock calculated?

This site may earn a commission from merchant affiliate links, including eBay, Amazon, and others.

HighPSI TSi Guy

20+ Year Contributor
1,212
11
Nov 7, 2002
Fredericksburg, Virginia
i have a few questions. .basicly, what unit does the 2G knock sensor read in and what kind of signal? also, how is the "knock count" calculated? what are the factors in it? anyone know the formula?
 
A knock sensor is basicially a small microphone. When preignition occurs, the "pinging" you hear is a sound generated by the resonance of the cylinder walls, and it tends to be within a certain frequency range. The knock sensor listens to this frequency range, and then sends a raw signal to the ECU. The ECU wil filter this signal (pretty complex, IIRC) and then assigns a "kock sum" value to it.
 
Originally posted by Tevenor
Thread moved.

sorry man i guess i wasn't specific enough. i had it in the other one because i was talking about the "complex" part of it. i know how the sensor works. i want to be able to take the signal and imitate the way the ECU interprets it for purposes of logging, viewing, etc....
 
The answer is that it's hard, and tends to be engine-specific.

If you have the resources, take an audio spectrum analyser (you need a fast one) and trigger it off the timing signal so that you only record around the ignition times.

Hang it off the knock sensor, and run the engine through a range where you experience knock.

Then spend some time staring at the plots looking for abnormalities. You will probably need the ability to look in both the time and frequency domains. What you'll hope to see will be a set of distinctive peaks for which you can then write a discriminator; I would expect that you'll be looking at a FFT followed by some simple peak/delta detection.

90% of the hard work will be in making the recordings and analysing the resultant data; the rest is a SMOP.
 
Add Value - Be Respectful - No Trolling - No Misinformation - Participate Often!
Support Vendors who Support the DSM Community

Build Thread Updates

Latest Classifieds

Back
Top