Knock detection filter circuitry

SleepyKeys · Post by **SleepyKeys** » Thu Mar 27, 2008 4:18 am

I ran across this IC, looks like it was designed just for knock sensing. Can anyone comment ?
http://www.chipcatalog.com/TI/TPIC8101DW.htm

-sean

Post by **Fred** » Thu Mar 27, 2008 9:39 am

Second to last post, last page, sorry for moving your post, just trying to keep it all tidy :-)

AbeFM · Post by **AbeFM** » Thu Mar 27, 2008 5:48 pm

Holy crap! Here I've been wondering what to do and getting advice from other people about how boxes put out a signal which only cost a few hundred... And there's these. Guess I'll have to look at them a bit more closely.

Still the biggest thing to me is can they put out a voltage proportional to the volume? Or, can we play the neat tricks some other boxes do, and only look in a narrow window around firing time? I guess that could be done in software as well.

ababkin · Post by **ababkin** » Mon Apr 07, 2008 2:53 pm

since this is related to knock, referencing:

http://www.diyefi.org/forum/viewtopic.php?f=16&t=161

toalan · Post by **toalan** » Tue Jun 10, 2008 4:00 pm

Hi,

I have some interest in knock sensing.

To me the ideal method would to use DSP techniques and forget about using too much analog filters. That would probably require a 32 bit uC perhaps maybe even a very high end 32 bit uC.

A good starting point would be to use a USB based DAQ system and do the processing on a high powered laptop/desktop using something like matlab or labview. You could use that setup to figure out how hardcore the DSP algorithms need to be and then make an informed choice on picking the right uC. Even if the uC version does not ever get made, having a tool that lives in windows/linux only is still pretty useful for lots of people.

I would like to try something like that but the problem for me is how do you simulate a knock signal and how would you be able to add enough noise in the simulated signal so that it represents a decent range of knock signals?

Regards,

Alan To

Post by **Fred** » Tue Jun 10, 2008 4:03 pm

Alan, have you seen this thread :

http://www.diyefi.org/forum/viewtopic.php?f=24&t=193

Fred.

AbeFM · Post by **AbeFM** » Tue Jun 10, 2008 8:03 pm

The real question is... Is ion sensing realistic for the home hobbyist, if not, then it's really worth just keeping after looking at the piezos. But a filter that's both RPM/MAP and time dependant would be pretty awesome...

shameem · Post by **shameem** » Wed Jun 11, 2008 8:20 am

Just in case...

MSextra knock sensing circuit-
http://www.msextra.com/manuals/MS_Extra ... .htm#knock

Standalone knock sensing system
http://www.viatrack.ca/

AbeFM · Post by **AbeFM** » Wed Jun 11, 2008 9:41 pm

Yeah - I have the knock sense. It works, it's really not too bad. But you have to have very loud knock to trigger it down low. A real-time tunable filter might help, but windowing would be the best. And an amplitude-vs-rpm input would be HUGE.

The problem is my motor (many motors) make more noise near redline than they make under severe knock at lower RPMs. So either your light is on all the time above some certain RPM, or you miss engine damaging knock at your torque peak.

Hence the need for a better amplifier, something that puts out a bigger signal for louder knock, and then MS or FreeEMS can ignore noise at the higher RPMs by setting a (for instance) 4v threshold at redline, and a 2v threshhold at torque peak.

DIYEFI.org Forum

Knock detection filter circuitry

Knock Sensor

Re: Knock detection filter circuitry

Re: Knock detection filter circuitry

Re: Knock detection filter circuitry

Re: Knock detection filter circuitry

Re: Knock detection filter circuitry

Re: Knock detection filter circuitry

Re: Knock detection filter circuitry

Re: Knock detection filter circuitry