Knock sensor gauge idea!

[turbovanman²]

Based on Rob's thread on knock sensors, the voltage has to be at a certain voltage to be real knock, I think it's 4.5 volts but whatever at this point, doesn't matter. I am having issues with my SES light always flickering or coming on, so I was thinking, why not have a small LED display showing the voltage, so then you know instantly if its real knock or phantom.

So would simply wiring a display from the knock sensor wire and ground work or would that mess up its performance?

[roachjuice]

http://img.tapatalk.com/d/12/09/09/nazunyqa.jpg

[turbovanman²]

????????????

[shackwrrr]

Sure a gauge like the Dawes that has a few colored leds would work.

[turbovanman²]

Sure a gauge like the Dawes that has a few colored leds would work.

Hmm, that's not a bad idea, but we don't really need all those LED's, just one would do that job, and easier to see at a glance too.

[cordes]

Why not properly set your knock table so that the CEL comes on during actual knock?

[turbovanman²]

Why not properly set your knock table so that the CEL comes on during actual knock?

Did you not read what I wrote?

[cordes]

I did, and I don't see a problem. Why is your gauge flickering? You can change all of the knock parameters so far as I'm aware. I don't see why your light should come on unless you want it to.

[turbovanman²]

Its coming on all by itself, has nothing to with knock, so I can't trust it anymore, hence my need for another way to monitor knock.

[GLHNSLHT2]

The voltage output that says you're knocking or not is not a linear curve.

[Force Fed Mopar]

You can log the knock sensor voltage with MPScan. The voltage that actually indicates knock can vary from sensor to sensor and engine to engine. I actually have mine set to 3v at 2500 and 3.7v at 6k, IIRC. But my knock sensor is new and very sensitive.

[turbovanman²]

Can't remember but does the sensor generate A/C or DC voltage?


The voltage output that says you're knocking or not is not a linear curve.

True, but I can set the LED or whatever I choose to start to come on at the known voltage of knock.


You can log the knock sensor voltage with MPScan. The voltage that actually indicates knock can vary from sensor to sensor and engine to engine. I actually have mine set to 3v at 2500 and 3.7v at 6k, IIRC. But my knock sensor is new and very sensitive.

That would help figure out where I start to knock as right now, my SES light is useless, :(

[Force Fed Mopar]

DC volts. I think anyway lol.

[RoadWarrior222]

There's a bargraph module that will work for that, just got some the other week...
http://www.turbo-mopar.com/forums/showthread.php?67638-bag-o-bargraphs-%28geekout-warning%29&highlight=volts

Basically, you'd put in the resistors for 5v, and it would flicker low at low knock, and nearly full scale for real knock... or there's a way of setting a bottom threshold with a variable resistor so you don't have it flickering all the time, but I think that would change the scaling, so say threshold was 3v the scale would be 3v to 5v across 10 LEDs, so when it's flickering at the left it's just hitting over 3v

[turbovanman²]

Yeah, my friend is going to make me an adjustable LED setup, :nod:

[ShelGame]

True, but I can set the LED or whatever I choose to start to come on at the known voltage of knock.

If you KNOW the voltage at every RPM that causes actual knock, why not just re-calibrate the knock threshold line? Then the light would only come on when you get actual knock.

The knock retard system works like this:

The knock sensor outputs an AC-type voltage the amplitude of which is proportional to the noise that the engine makes.

Many things in the engine make noise. Mostly the combustion itself. Knock (ie, poor combustion) is just 'noisier'.

There's some circuity in the ECU that converts the knock sensor AC to DC that can be read by the A2D converter in the processor.

Someone, somewhere, within Chrysler, long ago, sat down with a scope and a voltmeter and listened to the engine under various load conditions sometimes intentionally inducing knock while recording the voltage. They used that data to generate a knock voltage curve. Then, they took some voltage off of that for when the chambers get carbon built up, and for when you buy cheap gas, and for when the rings get bad and you start burning oil, etc.

That line gets programmed into the computer. Any voltage over it the computer 'thinks' is knock and timing is retarded until the voltage goes down. The computer actually has no way of knowing what is 'real' knock and what is just normal engine noise - outside of that calibrated line.

The CEL light in the Turbonator (and BB60, LB60) cals comes on only when there is active knock retard - IE when the knock voltage is over the threshold.

So, if you know what the threshold is for every RPM, why not just fix the threshold line instead of making a new gadget?

Finally, the 'actual' knock threshold is going to be different for each and every engine. Even 2 completely stock engines will have slightly different 'actual' knock voltage thresholds due to wear, machining tolerance, plug indexing, etc.

If you're running a stock 2.5 knock threshold on a 2.5 TIII, then it's not going to be right. The TIII knock threshold should be closer, actually. But, even that is very conservative and probably not what you want to run...

[RoadWarrior222]

instead of making a new gadget?

LUL whut? But a new gadget is a new gadget, and it would have blinky lights! :D


But Rob maketh sense that you don't want retard when you don't need it, so haxx0r teh cal already.

... although he also wanted a new gadget when his stage limiter didn't work.

[turbovanman²]

If you KNOW the voltage at every RPM that causes actual knock, why not just re-calibrate the knock threshold line? Then the light would only come on when you get actual knock.

The knock retard system works like this:

The knock sensor outputs an AC-type voltage the amplitude of which is proportional to the noise that the engine makes.

Many things in the engine make noise. Mostly the combustion itself. Knock (ie, poor combustion) is just 'noisier'.

There's some circuity in the ECU that converts the knock sensor AC to DC that can be read by the A2D converter in the processor.

Someone, somewhere, within Chrysler, long ago, sat down with a scope and a voltmeter and listened to the engine under various load conditions sometimes intentionally inducing knock while recording the voltage. They used that data to generate a knock voltage curve. Then, they took some voltage off of that for when the chambers get carbon built up, and for when you buy cheap gas, and for when the rings get bad and you start burning oil, etc.

That line gets programmed into the computer. Any voltage over it the computer 'thinks' is knock and timing is retarded until the voltage goes down. The computer actually has no way of knowing what is 'real' knock and what is just normal engine noise - outside of that calibrated line.

The CEL light in the Turbonator (and BB60, LB60) cals comes on only when there is active knock retard - IE when the knock voltage is over the threshold.

So, if you know what the threshold is for every RPM, why not just fix the threshold line instead of making a new gadget?

Finally, the 'actual' knock threshold is going to be different for each and every engine. Even 2 completely stock engines will have slightly different 'actual' knock voltage thresholds due to wear, machining tolerance, plug indexing, etc.

If you're running a stock 2.5 knock threshold on a 2.5 TIII, then it's not going to be right. The TIII knock threshold should be closer, actually. But, even that is very conservative and probably not what you want to run...

Your not getting it Rob, my SES light comes on by its own whenever it wants, already said this twice, so It can't be used right now to monitor knock as its always flashing on and off. I can't figure out why and tired of working on the damn thing, so this is my fix for now.

[RoadWarrior222]

Didya check for codes? :D :p

[ShelGame]

Your not getting it Rob, my SES light comes on by its own whenever it wants, already said this twice, so It can't be used right now to monitor knock as its always flashing on and off. I can't figure out why and tired of working on the damn thing, so this is my fix for now.

Ohhhh.... Short in the dash? Or caused by codes?

[turbovanman²]

Ohhhh.... Short in the dash? Or caused by codes?

No idea and right now, fed up of working on it, :(

I'll dig into more in the off season but right now, just need a valid way of monitoring knock while I am driving, this is easiest route for me right now.

[turbovanman²]

Rob, check out this please-

http://www.turbo-mopar.com/forums/showthread.php?67847-Charging-issues!

[csxtra]

Simon,

Just create an electronic Det can like DJ did:

http://www.turbo-mopar.com/forums/showthread.php?39999-93-2.4l-DOHC-Shadow&p=521051&viewfull=1#post521051

(scroll about halfway down through the mega-long post)

It works perfectly to calibrate your knock volts vs RPM table so you know that the light is always coming on when knock is detected, and that it is only coming on when real knock is detected (not just mechanical noise).

I plug mine into the aux input on my stereo, and you can definitely hear the difference between mechanical noise and knock.

[turbovanman²]

Yeah, I am going to have to make something like that, thanks for reminding me. :thumb:

[turbovanman²]

Simon,

Just create an electronic Det can like DJ did:

http://www.turbo-mopar.com/forums/showthread.php?39999-93-2.4l-DOHC-Shadow&p=521051&viewfull=1#post521051

(scroll about halfway down through the mega-long post)

It works perfectly to calibrate your knock volts vs RPM table so you know that the light is always coming on when knock is detected, and that it is only coming on when real knock is detected (not just mechanical noise).

I plug mine into the aux input on my stereo, and you can definitely hear the difference between mechanical noise and knock.


Found one, but not sure if it will work, added a post to DJ's build.

[wallace]

Simon,

Just create an electronic Det can like DJ did:

http://www.turbo-mopar.com/forums/showthread.php?39999-93-2.4l-DOHC-Shadow&p=521051&viewfull=1#post521051

(scroll about halfway down through the mega-long post)

It works perfectly to calibrate your knock volts vs RPM table so you know that the light is always coming on when knock is detected, and that it is only coming on when real knock is detected (not just mechanical noise).

I plug mine into the aux input on my stereo, and you can definitely hear the difference between mechanical noise and knock.

Is it as simple as connecting the two leads of the sensor to the stereo plug? I'd like to try this out myself to see how well it works.

[csxtra]

Is it as simple as connecting the two leads of the sensor to the stereo plug? I'd like to try this out myself to see how well it works.

For me it was, I went to the junkyard and got a few VW wideband sensors like in DJ's post, mounted the sensor to my intake manifold (used a longer bolt to mount it to one of the fuel rail mounting points), went to radio shack and got a mini headphone extension cord, cut the female end off of it, wired two of the wires from the sensor (the signal and ground - the third doesn't get used) to the left, right, and common leads on the extension cord, plugged it into my aux input on my stereo, and it worked fine.

I didn't need the resistor on my setup, but your mileage may vary.

Warren

[ShelGame]

I'd love to hear an audio recording of what that sounds like...

I think I'm going to do this for my Daytona. I've never had the chance to really tune the timing on my car...

[csxtra]

And if you don't have a stereo with an aux input, one of these works pretty well:

41419


http://www.radioshack.com/product/index.jsp?productId=2062620

(plus it comes in handy on SDAC awards night) :eyebrows: :bolt:

[turbovanman²]

Cool, I'll have this done tomorrow and my deck has an AUX input, :nod:

My sensor doesn't match DJ's, is that ok?

[RoadWarrior222]

OR a cassette adapter would probably work, or an FM radio adapter....

You might think the radio adapter wouldn't need a long wire through the dash, but I would, because for one, spark wires and such close by might play hell with it and for two the signal might not be strong enough to make it through the firewall... If you wanted to just loop the wires over the back of the hood and ziptie it to the back of the scuttle grille, that might work... but waterproof it.

---------- Post added at 04:52 PM ---------- Previous post was at 04:51 PM ----------

Oh, radio adapter might be a biatch to tune in... might sound like off station static unless you're knocking real bad at the time... may be less of a problem with a digital tuner.

[csxtra]

My sensor doesn't match DJ's, is that ok?

Simon,

From what I understand, as long as the knock sensor has a through bolt (donut shaped), it is a wideband knock sensor and will work for this.

However, if yours is a KS1 054 905 377, that is the same one I have on my car...so you should be just fine. :thumb:

If you want to be sure, before you mount it, wire it up and plug it into your stereo. Tap the bottom metal part with a wrench, you should be able to hear it clearly (you will have to experiment to find the right volume though so you can hear it and don't blow any speakers).

[RoadWarrior222]

Also the radio adapter you can pick up with cheap FM headphones if you've stripped out your radio or it doesn't work.

[RoadWarrior222]

Oh dunno if anyone else has noticed this but a 3.5mm jack connected to an amplifier, i.e. not the one on a pair of normal headphones, is sort of an insensitive microphone, if it's not a very high quality jack. Tap the tip of it and you'll hear speakers pop. This may mean that you could tap a thread on the end of a jack cord, then JB weld a captive nut on the block somewhere and screw it on, then pick up knock with active amplification.... which would be independant of the real knock sensor...

The reason it can work is because the capacitance at the tip changes when you tap it because the plastic ring between it and the next ring gets compressed slightly.... however, this may be a bit experimental, and you might have to screw around a bit with a few cords and jacks. It may be dependant on what static charge the thing has picked up, or partially reactive to your body as a ground, so not sure if it would work reliably, but just thought I'd mention it in case anyone felt like a science fair project :D

[turbovanman²]

Simon,

From what I understand, as long as the knock sensor has a through bolt (donut shaped), it is a wideband knock sensor and will work for this.

However, if yours is a KS1 054 905 377, that is the same one I have on my car...so you should be just fine. :thumb:

If you want to be sure, before you mount it, wire it up and plug it into your stereo. Tap the bottom metal part with a wrench, you should be able to hear it clearly (you will have to experiment to find the right volume though so you can hear it and don't blow any speakers).

Thanks Warren. Need to find the harness side connector and off to my electrical store for a radio jack and some wire.

[turbovanman²]

Ok, can't find a spot on the front or back of the block, is the side ok or the intake?

[RoadWarrior222]

I had another idea, a knockoscope... takea cheap bluetooth headset and get into it, hook up the knock sensor to where the microphone was, with resistor, most audio inputs expect about 1V peak to peak, not 5V. Then you pair your iPhone or Android to it, and use an audio recording app that gives you a waveform display... then you can see any spikes, and also record a run up the track or whatever.

I did start thinking about stripping the guts out of old nokia brick phones and using a LM3914 LED driver/voltmeter chip to drive the pixel rows and a counter to scan the columns, making a mini-oscilloscope..

this kind of thing http://www.transkommunikation.ch/dateien/schaltungen/diverse_schaltungen/messgeraete_circuits/Solid State Oscilloscope.pdf but with maybe another one or two of each chip daisy chained, for more resolution and some adaption for driving the LCD because it isn't quite teh same as driving LEDs...

But... that seemed to be getting a bit complimicated.

Edits...

BTW, there is also a way to get a dyno analysis out of spark events from an audio recording, you rig a (ratehr insensitive) spark pickup to an audio device, do a WOT run and how fast your motor speeds up with known weight and gearing can be analysed from spark data over time which is what the audio recording is. Anyway, bluetooth headset and smartphone can be rigged for that also.

For a stereo recording input, you could get spark against knock! even measure how many milliseconds, then convert that to a degree adjustment for certain rpm in the cal.

BTWII: brick phone displays got popular for the microcontroller crowd.. https://www.sparkfun.com/products/10168 that's one mounted up, not sure if it's the same controller as you may pull right out of the phone. however, for the simple analog input purpose I was thinking above, the controller might "get in the way" as it were. Some data and hints there. Instead of the bargraph drivers you'd need analog to digital conversion. However, that led me to thinking, a brick phone (5190, 5110, 6190 and similar) has a 5V serial interface on the bottom, along with audio input, and an ARM CPU, and a display... which MAY be suitable for a real time monitoring device from teh data bus, and code reader. i.e. you could get text display of map value knock value rpm etc. like the palmOS digital dash but far less fancy. However, it would take "mad h4x" to figure out how to reflash it to such a purpose. I dunno if it would be a waste of time, but you can pick those things up for next to nothing. If doing it, I guess I'd try and write a portable C code "chrysler diagnostics module" or library that was independant of the hardware interfacing/drivers and then it could be ported to use the maintenance console serial input on a hacked 'droid or something later. But anyhoo, don't hold your breath waiting for that. Though I'm just gonna go off and see how much the bricks are hackable.

[wallace]

Simon I have a question for you on the knock light problem you are having. Are you running solid core plug wires?

[turbovanman²]

Simon I have a question for you on the knock light problem you are having. Are you running solid core plug wires?

Nope, Aurora's, had them for years, :nod:

[wallace]

Nope, Aurora's, had them for years, :nod:

The same set? Just thinking out loud but that would be one thing I would investigate if the wires are old and the insulation they are using for suppression has begun to break down it could be triggering your ks.

---------- Post added at 01:23 PM ---------- Previous post was at 01:21 PM ----------


Simon,

From what I understand, as long as the knock sensor has a through bolt (donut shaped), it is a wideband knock sensor and will work for this.

However, if yours is a KS1 054 905 377, that is the same one I have on my car...so you should be just fine. :thumb:

Would that be for any manufacturer? I've got a GM 3500 V6 in here with dual knock sensors on it and they are the donut style that are through bolted.

[turbovanman²]

The same set? Just thinking out loud but that would be one thing I would investigate if the wires are old and the insulation they are using for suppression has begun to break down it could be triggering your ks.

Well so far, changing out the regulator portion of the SMEC seems to have fixed it.

[turbovanman²]

Bump for my location question.

[Reeves]

Which side of your intake manifold? Front or back or left or right? I would think left to right would be OK, but pass side or drivers side would give some skewed readings. Like if #1 was knocking but you had it mounted on the #4 side.

[turbovanman²]

I can put it on one of the middle mounting bolts.

[csxtra]

Simon,

I'm not sure of the best place to mount it, I went with the most convenient place to mount mine, using the forward fuel rail mounting bolt on the driver's side of the engine (between cyls 3 & 4). I know it is not an optimal location , but I am able to hear knock pretty well, although I have not verified how well I can hear knock on each cylinder.

But this is on an 8 valve, I'm guessing that if you can find a place to mount it on your T3 intake manifold that is close to between cyls 2 and 3 that it should be able to hear knock, but you might have to experiment with different locations.

[turbovanman²]

Thanks Warren, going to put it on the intake right in the middle at the head, :nod:

[turbovanman²]

Picked up my wire, plug and a 0-50K ohm adjustable pot, hope to get it installed tonight.

[Reeves]

As DJ stated, it's easy to drill out for a 10mm bolt if need be. I just finished doing mine on our Bridgeport :thumb:

[wallace]

Picked up my wire, plug and a 0-50K ohm adjustable pot, hope to get it installed tonight.

How did it work out?

[turbovanman²]

Sorry, got slammed at work then had to go racing and my stocker actually worked Fri, but will be doing it this week as I want to turn the boost up on Friday, my last kick at the can for the year.

[Force Fed Mopar]

Simon,

I'm not sure of the best place to mount it, I went with the most convenient place to mount mine, using the forward fuel rail mounting bolt on the driver's side of the engine (between cyls 3 & 4). I know it is not an optimal location , but I am able to hear knock pretty well, although I have not verified how well I can hear knock on each cylinder.

But this is on an 8 valve, I'm guessing that if you can find a place to mount it on your T3 intake manifold that is close to between cyls 2 and 3 that it should be able to hear knock, but you might have to experiment with different locations.

On your 8v, why didn't you use the stock knock sensor location?

Also, it looks like you said you wired 2 leads off the knock sensor to 3 leads off the headphone cord? Can you explain better, or snap some pics? I need to do this myself.

[RoadWarrior222]

On a headphone jack, the tip is supposed to be ground, and the sleeve is signal, left and right if stereo but dunno which way round they are, and feeding in a mono signal you can connect them together.

[Force Fed Mopar]

Ohh, I see :)

[csxtra]

On your 8v, why didn't you use the stock knock sensor location?

I didn't use the stock location because I didn't want to risk changing the sensitivity of the stock knock sensor and having to modify my Knock Volts vs RPM table because I moved the sensor. So I just looked for the next most convenient mounting location, which was a fuel rail mounting bolt.


Also, it looks like you said you wired 2 leads off the knock sensor to 3 leads off the headphone cord? Can you explain better, or snap some pics? I need to do this myself.

And I found the following picture that shows which parts of the stereo plug are which:
42000

I connected the ground to one of the output wires on the knock sensor, and tied the other two wires (left and right) from the stereo mini plug together and connected them to the other output wire. That way, I can hear the sound from the knock sensor on both channels of my stereo.

I hope this helps.

Warren

[Austrian Dodge]

if it just has to show knock why not go with one of these: http://www.phormula.co.uk/KnockDetection-KS-3.aspx

ran both, the headphone setup like DJ and because i'm anal the KS3 ... both sourced from the same sensor and the KS3 was showing me exactly what i was hearing in my headphones. so that thing works for sure!

[Force Fed Mopar]

I didn't use the stock location because I didn't want to risk changing the sensitivity of the stock knock sensor and having to modify my Knock Volts vs RPM table because I moved the sensor. So I just looked for the next most convenient mounting location, which was a fuel rail mounting bolt.



And I found the following picture that shows which parts of the stereo plug are which:
42000

I connected the ground to one of the output wires on the knock sensor, and tied the other two wires (left and right) from the stereo mini plug together and connected them to the other output wire. That way, I can hear the sound from the knock sensor on both channels of my stereo.

I hope this helps.

Warren

Ah, didn't realize you were still using the stock sensor. Makes sense.