| Go back to previous topic |
| Forum name | Turbo/Nitrous Tech |
| Topic subject | knock sensor nice info |
| Topic URL | https://forums.2gnt.com/dcboard.php?az=show_topic&forum=i8&topic_id=37236 |
37236, knock sensor nice info Posted by MotoFool, Jan-11-03 05:41 AM
here check it out i want to try this...
http://www.neons.org/forums/viewtopic.php?t=72568
|
37237, RE: knock sensor nice info Posted by JustOne, Jan-11-03 08:29 AM
MSD makes a LED warning knock sensor. Armond was using one but I think I saw a post in the for sale section w/ him selling it. I never saw him post on if it worked or not. It's like ~140 through Summit.
Armond if you read this post how did you like it???
|
37238, RE: knock sensor nice info Posted by Greg, Jan-11-03 09:00 AM
I also have one installed in my car. It does seem to work pretty well. The only problem is finding the right sensitivity for the unit to actually determine it it's "true" knock and not just to sensitive (it's adjustable). I do know that there are better ways out there for monitoring knock, but this one is not bad. I find this post interesting and I'm curious to find out if it really works. I believe Armond is working on another method of monitoring knock which looks to be very promising. He uses this mothod on his his Stealth VR4 and from what I gather it works very well for him. I'll let him give the rest of the details. ;-)
|
37239, RE: knock sensor nice info Posted by Dirty, Jan-11-03 10:44 AM

I wanna know if there is a way to modify this one to work with our cars. I like this one cause it fits in a gauge cup.

|
37240, RE: knock sensor nice info Posted by MotoFool, Jan-12-03 05:28 AM
MisChiefMaker33: Hey, what was your decision on that msd knock sensor on our cars good or bad? Armond30: I think it is a piece of junk MisChiefMaker33: i see there is a thread on 2gnt people requesting your input on it, Armond30: oh. well, tell them I think it is junk
MisChiefMaker33: what other way is there that is good to monitor knock? Armond30: I hooked up the afc input 2 to the mitsubish knock sensor, and it reads it's voltage MisChiefMaker33: the safc has a knock sensor input? or u used a input for another sensor and jsut monitor the voltage? Armond30: no, mitsubish knock sensors output voltage in proportion to knock, so I just read t like I woudl an o2 sensor with the afc MisChiefMaker33: i see, could we put a mitsubishi knock sensor ina 2gnt? Armond30: not to replace the stock one, but you can put on in just to montor MisChiefMaker33: hmm where would be a good place? Armond30: anywhere on the back of the block
|
37241, RE: knock sensor nice info Posted by MotoFool, Jan-12-03 05:31 AM
now that we have that i want to find a way to monitor knock with an afc =P since i wont be getting one...
|
37242, RE: knock sensor nice info Posted by Dirty, Jan-12-03 10:17 AM
Armond30: no, mitsubish knock sensors output voltage in proportion to knock, so I just read t like I woudl an o2 sensor with the afc
So does that mean we can use an A/F gauge to moniter knock? All I got outta that was the MSD was a piece of shit for monitering knock.

|
37243, RE: knock sensor nice info Posted by MotoFool, Jan-12-03 07:06 PM
he used the 02 sensor inpout for knock monitoring with a safc,
so he has the afc monitoring knock and when it gets some it makes the 02 sensor voltage jump on his afc....
not a bad idea a/f meters run off of 02 sensors so why not use one of those with a modified face plate to keep them straight...
|
37244, RE: knock sensor nice info Posted by mcgyvr, Jan-13-03 12:13 AM
the sample rate at which the safc reads voltage is not nearly fast enough to measure knock, sure the voltage is changing but there is no way in hell I would use that data to tune with.
You need an ocilloscope to truly measure knock on our cars.
<b>Anyone try the Caspro Electronic knock eliminator, it retards timing when it sees knock. </b>
|
37245, RE: knock sensor nice info Posted by MotoFool, Jan-13-03 02:27 AM
he was usging the mitsubishi knock sensor and the o2 readout on the safc not the same as stock chrysler knock sensor id assume?
i was told by armond our stock knock sensor is like a microphone that listens and the ecu interpets the signals...
the mitsubishi knock sensor is just a voltage regulator when it feels the right vibration...
|
37246, RE: knock sensor nice info Posted by jamesman, Jan-13-03 02:52 AM
With regards to knock sensors, from what i tested with John at J&S it puts out an AC voltage with an amplitude of about 3 volts. i do not know if this is a maximum voltage because on a piezoelectric sensor there is what i would call a diaphram that when resonated produces an output voltage. For our knock sensor, i do not know the true resonant frequency that outputs the largest amplitude, therefore i do not know for sure what voltage level output from our sensor constitutes a knock condition. that is the important thing. see the knck sensor will output a voltage at any vibration that is within its sensing range. but the ecu is programmed to pull out timing only when the output from the knock sensor reaches a certain value. it would be nice to see what voltage that is. I am hypothesizing that it is in the ~3volt range from the data i have gathered. I have devised a circuit that will switch a transistor when knock sensor output reaches 3 volts and that will in turn light up an LED on my dash. I have researched this and i am 99% sure it will work. actually. if you just built a rectifier and hooked an led directly up to the knock sensor output and knock reaches 3 volts then the led by itself will blink. i do not know if this willaffect the ecu's ability to monitor knock, but i guess i will find out when i finish building this circuit. lates james
|
37247, RE: knock sensor nice info Posted by MotoFool, Jan-13-03 04:26 AM
hmm if j&s can figure out what voltage it is maximum and what makes the ecu pull timing we could rig up something to apply that amount of voltage as boost rises hmm sounds like a good idea =P
|
37248, RE: knock sensor nice info Posted by The1Bill, Jan-13-03 06:27 PM
I would recommend using an operational amplifier (op-amp) to isolate the signal travelling from the knock sensor to the ecu from the tap to moniter the sensor. That way, the line can be monitered without the voltage drop that would result from the act of monitering the signal. -=B-=
|
37249, RE: knock sensor nice info Posted by MotoFool, Jan-14-03 05:57 AM
okay bill u lost me there ^
|
37250, RE: knock sensor nice info Posted by mcgyvr, Jan-14-03 07:45 AM
yes good point bill, and max voltage is 6vdc according to chiltons
|
37251, RE: knock sensor nice info Posted by jamesman, Jan-14-03 08:18 AM
okay our knock sensor does not output 6 VDC it is an ac voltage, this voltage switches a transistor that switches a 5 volt supply inside our ecu. There is however a peak to peak votage difference of 6 volts, but is is still not DC. lates james
|
37252, RE: knock sensor nice info Posted by MetalJim, Jan-14-03 05:35 PM
So you need an AC volt meter to monitor the voltage output from a sensor?
 ride__________95 Eclipse RS : 5 speed power________Field SFC Hyper-R : Gude cams : Jeep TB : AFX UDP : ram air : test pipe : GST catback suspention___Hurricane F/R lower strut ties : BOMZ F/R upper strut bars : ES inserts sound________Kenwood KDC 5000 : Audiovox 8" sub toob : Xplod 222 shizzle_______AutoCommand remote start : dual air horns : Caspers o2 simulator(rear) incomming___Bosal header(maybe), 14b turbosystem(eventually) Jeep TB writeup - http://www.dimensia.com:81/jimbo/JeepTBfor2gnt.html
|
37253, RE: knock sensor nice info Posted by jamesman, Jan-15-03 05:38 AM
i removed the knock sensor and hooked it up to an oscilliscope. if you build a rectifier with some diodes you can make it output 3 volts DC, but other than that i dunno, also the ring time is only 20 mS so like if you tap the sensor it wil output 3 volts and then die down. it takes only 20 mS for ths to happen, i don't know if a multi meter is fast enough to read this. A good DMM will read like 6 to 8 readings per second. so you may be able to get a reading, but i think it will be iffy lates james
|
37443, RE: knock sensor nice info Posted by WheatKing, Jan-16-03 07:02 AM
>With regards to knock sensors, from what i tested with John >at J&S it puts out an AC voltage with an amplitude of about 3 >volts. i do not know if this is a maximum voltage because on >a piezoelectric sensor there is what i would call a diaphram >that when resonated produces an output voltage. For our knock >sensor, i do not know the true resonant frequency that outputs >the largest amplitude, therefore i do not know for sure what >voltage level output from our sensor constitutes a knock >condition. that is the important thing. see the knck sensor >will output a voltage at any vibration that is within its >sensing range. but the ecu is programmed to pull out timing >only when the output from the knock sensor reaches a certain >value. it would be nice to see what voltage that is. I am >hypothesizing that it is in the ~3volt range from the data i >have gathered. I have devised a circuit that will switch a >transistor when knock sensor output reaches 3 volts and that >will in turn light up an LED on my dash. I have researched >this and i am 99% sure it will work. actually. if you just >built a rectifier and hooked an led directly up to the knock >sensor output and knock reaches 3 volts then the led by itself >will blink. i do not know if this willaffect the ecu's ability >to monitor knock, but i guess i will find out when i finish >building this circuit. >lates >james
Just an idea.. why not go the other way..
Hook up a clean 3v AC powersource and slowly up the voltage on the knock sensor until you see the ECU pulling timing.
Reverse engineer the computer, not the sensor. The sensor is meaningless in the overall scheme of things.. we just need to know the AC voltage that the ecu starts to reduce timing at.
-- WheatKing
|
37445, RE: knock sensor nice info Posted by jamesman, Jan-16-03 07:58 AM
Ahhhh Yes that could be done, i'll try it, i gotta go look for an ac powersource that i can borrow here at work. lates james
|
37571, RE: knock sensor nice info Posted by babybeclipse, Jan-18-03 05:09 AM
>Ahhhh Yes that could be done, i'll try it, i gotta go look >for an ac powersource that i can borrow here at work. >lates >james
So let us know if it works!!! :thumbsup
|
37446, RE: knock sensor nice info Posted by dougie2, Jan-16-03 09:09 AM
Damn... great idea! I would really love to see this work out! Great job guys, keep up the good work! :)
|
37447, RE: knock sensor nice info Posted by MotoFool, Jan-16-03 09:16 AM
ya now were getting down to business =P nice to see ya around wheatking =P
|
37450, RE: knock sensor nice info Posted by thedawg, Jan-16-03 10:15 AM
From the formula on http://www.autospeed.com/A_0348/P_3/article.html, here is the detonation resonant frequency for our engine.
.020 over (3.464) = 82.7 Hz Standard (3.444) = 83.2 Hz
Totally guessing, but if this equation is accurate I bet the factory knock sensor will detect knock as any tone around 81-84Hz at sufficient amplitude over ambient. It must work over at least a small range just because there are .040 over factory pistons and service specifications in the factory manual.
Hmm... I just ran it thru a tone generator and it's an awfully low tone... dunno... email me and I'll send you the .wav if you're interested.
|
37459, RE: knock sensor nice info Posted by jamesman, Jan-16-03 01:35 PM
dan that formula is 900/cyl radius in M *3.14 so our bore is 3.465 (thats .020 over) this means that 3.465*2.54=8.8CM which is .088 meters. you divide that by 2 and get radius .044 then you multiply .044*3.14= .138 then you divide 900 by .138 and get 6521.7Hz i think i'm right? i think 80 HZ is way too low, that is like sub sub bass, and remember when we get detonation it is a high pitch sound. Lates james
|
37506, RE: knock sensor nice info Posted by thedawg, Jan-17-03 06:58 AM
Damn metrics.... ;) You'd think a Mopar guy would convert to inches.
Ok, yeah that sounds a lot more like a pinging noise... almost exactly what you get when you drop a 12mm socket on concrete. Believe me, I've heard that alot lately hehe... but it does sound different when said socket hits you in the forehead.
|
37525, RE: knock sensor nice info Posted by jamesman, Jan-17-03 09:24 AM
heheh or a front sway bar hehe so what you doing this weekend? lates james
|
37526, RE: knock sensor nice info Posted by 98RsT, Jan-17-03 09:34 AM
Going over to James house and watch him play with his new eManage unit!
|
37535, RE: knock sensor nice info Posted by thedawg, Jan-17-03 10:56 AM
YEAH! I just got tickets to the "Watch James Play With His Emanage" tour on Sunday!!! I'm gonna bring a picnic basket with an EGT sandwich and a manifold on rye. Oh, and a really big bag of peanuts to throw at the monkeys, and some rice. TURBOCOOKER RICE!
Oh, and if the car's functional tomorrow, I'm going to do a quick burn to Vegas Saturday nite and meet up with Vegasdsmr's crew. 50% off Blue Man tickets :P
|
37654, RE: knock sensor nice info Posted by lazerus, Jan-20-03 02:54 PM
I manufacture an audible knock sensor, you have an alligator clip, you clip onto a bolt near the top of the block, and then an amplifier and speakers, and you personally "listen" for knock, more details coming soon.
the price will be 40$
|
37732, RE: knock sensor nice info Posted by babybeclipse, Jan-22-03 03:10 AM
Jamesman-Well? Did you try it?
|