does any one know how i could read my knock sensor. I am using an ecu that doesn't have a pinout for a knock sensor and i really want to use my d16y8 knock sensor to help me tune my turbo'd crx. I know the new vafc II has one but those cost way to much. any affordable solutions

 

help smart people

 

Oscilloscope?

KS is kind of like a microphone. I don't think it gives anything like a proportional signal that says 'now its knocking'. The ECM probably has to look for high dv/dt and high frequency content.

Has anyone watched one on an oscilloscope?

 

ok, so if i hooked it up to some sort of oscilliscope (i am thinking of hooking a data acquition board up to my pc) to my computer and then attached the leads of the knock sensor to that i would be looking for a high dv/dt?

ok, back to basics. a knock is when your timing is of in a manner that causes a detonation or predetonation on or before TDC right? so what does this knock sensor measure? it wouldn't be a pressure spike would it? i am thinking maybe some sort of stress on the head?

ok smart people........i need some input on how to use my knock sensor to tune for detonation.

ps........sorry for the spelling, its allowed, im an engineer.....haha

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by VTgringo &raquo;</TD></TR><TR><TD CLASS="quote">... so what does this knock sensor measure? it wouldn't be a pressure spike would it? i am thinking maybe some sort of stress on the head?</TD></TR></TABLE>
I haven't looked at one with a scope. Maybe someone here has??

I believe the KS is just a microphone. Descrimination happens in the ECU. I'm guessing that some of the higher frequency noise from autoignition is what they're looking for. Maybe they use some kind of FFT on the signal??

I think the big pressure spike in the combustion chamber is when you allow audible knocking to occur. It's supposed to be controlled by backing down the timing before it gets audible... But the KS is screwed into the outside of the block. It can't measure pressure, it can't even really measure stress. That's why I think it has to be a microphone.

You probably have to start with a real oscilloscope; look at the signal before you decide where to go from there. Magnetostrictive or piezo transducers probably have millivolt signals. I think the characteristic noise from spark knocking is maybe 6 to 8 kHz. To sample with a DAS, you want maybe 40 kHz sampling rate to avoid aliasing. I'm guessing that's not your ordinary DAS board for a PC.

 

u guys are getting out of control... it just works on voltage...

i forget the exact voltages that it runs off of, I dont have my helms manuel on me either. I was playing around with it one day and I hooked the wire up to my volt meter (red) and the black to ground, and i took a hammer to the back of the block and lightly tapped and i was watching the voltage jump around. I thought it was a pretty cool concept.

 

nyquist theory. yummy. 44.1sps you can do that with windows. the bit rate can be constrained to save space and cpu time. only thing is it will be making a file while it would be monitoring if you did it that way. holy crap... thats cool.

 

If the KS produce enough voltage when the engine is knocking, is there anyway you can hook up a LED light directly to it. Someone clarify on the voltage.

 

Guys,

Save yourself a ton of work and trouble, and get one of these:

http://www.msdignition.com/1pr...68229

 

can someone put up those voltages that the knock sensor puts out when it detects knock.

 

Yea, I'd like to see them too. I always thought it was a microphone & you have to look for a certain kind of noise.

 

bump

 

Bump cause Im interested in using a homemade KS, would the GM KS be able to detect knock better?

Blaze

 

i am going to veture to say that the MSD kit is extremely basic and almost useless because it is a "universal" kit. The audible frequency of knock depends on the bore of the engine and specifically the size of the combustion chamber, so you can't have a very accurate "knock box" that will work on many cars. Link Engine Systems (I Think) makes one called the knock link that is really basic (too basic) and is probably very similar to the MSD. It is merely a microphone amplifier with the output hooked up the LED's as opposed to a speaker. There isn't even any kind of bandpass filter. I've found it to be border-line useless as it only tells me when I'm going WOT because of normal operational noise, like I didn't know that already

 

if you find out the voltages and what DAQ board you are using i can prolly create a labview VI (a software package from national instruments) and then turn it into a exe that you could open up and it would sampe your KS in realtime and give you any sort of display you wanted. actually if you somehow knew the pinouts on the ecu etc. i may be able to rig something up that could retard the timing too.
ive been ******* with this DAQ **** all summer for research for a prof and it REALLY powerful.
good luck
later

 

Any suggestion for a cheap DAQ board to sample a wideband o2 sensor (0-1v range)

 

i found this start up package that was only like 40$. i forget the name of the company but they were selling there product on ebay. i believe it had 4 output channels and 2 input channels. im not too sure about the sampling rate, but it was decent.

I also had this idea of using an accelerometer along with the package to have some sort of a laptop dyno. i found some equations that the g-tech company was using to convert acceleration to hp and tq using gear ratio, rpm and acceleration as inputs.

burke, yeah, labview is great. I made a simple oscilliscope program this summer. and you probably could use the outputs of the daq to retard or advance distributor timing if you knew what you were doing. but i have a zdyne ecu. what school do you go to? I'll be a senior at Virginia Tech....obviously in mechanical engineering. maybe you could help me write some labview code for my "laptop dyno"

 

<TABLE WIDTH="90%" CELLSPACING=0 CELLPADDING=0 ALIGN=CENTER><TR><TD>Quote, originally posted by VTgringo &raquo;</TD></TR><TR><TD CLASS="quote">i found this start up package that was only like 40$. i forget the name of the company but they were selling there product on ebay. i believe it had 4 output channels and 2 input channels. im not too sure about the sampling rate, but it was decent.

I also had this idea of using an accelerometer along with the package to have some sort of a laptop dyno. i found some equations that the g-tech company was using to convert acceleration to hp and tq using gear ratio, rpm and acceleration as inputs.

burke, yeah, labview is great. I made a simple oscilliscope program this summer. and you probably could use the outputs of the daq to retard or advance distributor timing if you knew what you were doing. but i have a zdyne ecu. what school do you go to? I'll be a senior at Virginia Tech....obviously in mechanical engineering. maybe you could help me write some labview code for my "laptop dyno" </TD></TR></TABLE>

you don't even need an accelerometer to have an accurate dyno. All you need is to log RPM vs. Time and (DELTA)RPM (after some math) gives you your torque. The more quickly RPM changes, the more torque you have, right? Someone offers a gimpy software package (one diskette) and an inductive pick-up (like from a timing light) that you plug into the microphone port on your laptop to get a dyno graph on the road. I forget the company (probably just a guy), but a google search will turn it up. It's VERY consistent within itself and within just a couple % of "actual" dyno horsepower. If you want your actual HP you will need to tak into account your gear ratios and tire size, probably... OR...you can just use http://www.ecucontrol.com, log your rpm/time and then do some math on that ****, I'd think. Speaking of ecucontrol...does someone want to help me out with assembling my own rs232 conversion board thingie? I have a thread here in tech about it...I just bought the chip (but accidently the max232, as opposed to the max233 and don't have more than a clue as to what capacitors I need...

 

hey ionfour......holy ****.....this is the coolest program i have ever seen and its free. I am talking about ecucontrol! since i have modified my pm6 ecu through zdyne will the software work with the cable that i use to program my zdyne. I am thinking that it will work great......omg this is the sweetest thing ever. now i can brew up some pv=nrt formulas knowing air temp at specific rpm's.....this will revloutionize tuning...i can probably tune a great code just knowing how many moles of air i have and putting the right fuel in to reach my desired a/f ratio........i am in heaven........won't have to pay for that stupid tech guy to tune my car and dyno it.

 

the knock sensor works just like a microphone.. it outputs a voltage that corresponds to the vibration it feels. you would deff have to have an o-scope or log the voltage on a computer. the ambient noise gets louder as the rpms increase. 1.5 volts might be knock at 3000 rpms but at 6000 rpms thats going to be normal voltage. you would have to log the voltage and look for spikes. do a part throttle pull all the way to redline and log that, then when you do full throttle pulls compare to the baseline you made at part throttle

 

thanks for the info furry. That sounds like a great way to test out the signal. i swear sometimes i just don't feel like i put my head on in the morning.