ok i use uberdata w/ a zeitronix wb and set my user input to datalog the knock sensor on the back of my y7 block (came from a y8). now i tried datalogging the values and get some wierd ****. mostly it will just pick a voltage and stay there, for a while and will then go to another voltage and hang out.
now i figured i would be looking for spikes in the voltage. BUT after doing some research, i think i have found out that the ks is an ac signal and will change amplitude with more noise. right???
anybody do this or have any ideas?
what if i put in an analog to digital converter and then looked for spikes?


Thanks for the help, im just trying to figure this out before i start playing with timing

 

bump

 

Does your ecu even have the circuitry for a knock sensor (I should assume yes?)

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by Burke &raquo;</TD></TR><TR><TD CLASS="quote">... i think i have found out that the ks is an ac signal and will change amplitude with more noise. right???</TD></TR></TABLE>I don't know for sure since I haven't looked at the signal with an oscilloscope. (But you're not getting any other answers...)

I think the KS is like a microphone, so it's a noisy waveform, not really like AC. Knocking is more than just high amplitude, it's a very high dv/dt. The ECU is probably doing a bunch of signal conditioning before the KS signal is available for datalogging.

If you want to datalog the KS signal directly, first watch it with an oscilloscope. Then you'll know what kind of A/D performance you need. Maybe it's higher frequency than you think.

 

thanks for the input guys.
it is not hooked up to an ecu, i am just trying to pull the signal of the ks itself. i am running a p28 w/ uberdata.
jim, if you are saying that the ks output is not only a high amplitude, but knock is present with a high change in the voltage over time, then i should be able to write a simply formula in excel that could somewhat detect the knock when i reveiw the data. i'll have to some more testing with the thing. i am currently not using it because my wideband only allows one other data input which i am currently using for the 3bar.
thanks for the help!!!

 

try the PGMFI forums. lots of stuff going on with R+D concerning knock sensors and such. and isnt the KS either on or off?? maybe just the way the ecu understands. your either knocking or your not.

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by Burke &raquo;</TD></TR><TR><TD CLASS="quote">jim, if you are saying that the ks output is not only a high amplitude, but knock is present with a high change in the voltage over time, then i should be able to write a simply formula in excel that could somewhat detect the knock when i reveiw the data.</TD></TR></TABLE>Not high amplitude, really. High-frequency hash on top of the normal acoustic signal. But what I'm really saying that you better watch it with a good, fast oscilloscope first, because you don't know what A/D you need.

If a scan tool can monitor 'knock' from a running ECU, that's a lot different than watching the raw signal from the sensor. There has to be a lot of signal processing going on somewhere.

I've only seen KS traces dating back to textbooks when I took IC engines in school. But you might be looking for 400khz oscillations, & you're not gonna see them unless you're sampling at 1mhz or better. That's not your typical PC/Excel stuff. I'm guessing even music mixdown software can't go that high. Excel can't handle more than 65536 rows. That's like 65 milliseconds of data...

Definitely check out the stuff b18c1civic95 is talking about.

 

woah, better borrow an oscope!
i'll go did around pgm-fi and see what there is to be seen.
Thanks again for all the help.
also
where are you at in NE ohio Jim?? im in lovely y-town..............................

 

I'm kinda near Akron