Jump to content
HybridZ
supernova_6969

hooking up stock 280zx knock sensor to megasquirt

Recommended Posts

Hi. Jester
Thanks for the reply. By low load, do you mean low electrical load, i.e. very faint signal, or do you mean that the sensor would only work when the car is being handled gently?
I read somewhere (megamanual?) ms2 can use knock sensor outpout and that you need to give it an upper rpm limit ir else it's going to think parts of the mechanical noises is pinging..
Just not certain how i would wire out the sensor and make it work... ( if the stock sensor coukd even soeak the same language)

Edited by supernova_6969

Share this post


Link to post
Share on other sites

The sensor is just a piezo electric crystal that outputs a small voltage when it "hears" a knock. Hook it up to an analog input on the MS and you can monitor the knock level. 

 

MS will pull back ignition timing based on the setting you enter. Look under Ignition setting for TunerStudio.

Share this post


Link to post
Share on other sites

Low load as in very low rpm not much road resistance. A knock under high load high rpm will be catastrophic, a knock under low load low rpm will not be anywhere near as bad. 

 

What jacky says sounds correct, it is just a "hearing" circuit that sends a voltage once it "hears" a sound. You just have to provide the sound via a low load knock, then set the threshold for the resulting voltage, then go into megasquirt and have it wired to the correct input, (one of the spares or the IACV if you are not using it or anything else in the drop down menu) and have it control ignition. I don't think MSII is advanced enough to have bias control, but it should have some control mechanisms in the ignition/spark settings. 

 

This is as far as I know, I haven't done it, I've barely looked into it. I would suggest the megasquirt forum, they are going to want as much information as you can find, sensor ID, your MSQ etc and they will set you straight.

Share this post


Link to post
Share on other sites

Oh!

sorry, I misunderstood what you meant.  yeah, wouldn't want to have blow the engine just to set up a knock sensor, no.  

 

one question that I have though is about the physicality of the wiring;   the gm sensor that the megamanual recomends has only one wire, which goes to a gismo, which then only sends one wire to the MSII.    the stock nissan sensor has two poles.

 

anyone know what goes in and out of it?  I'm guessing one might be positive and the other one the signal, although from what I know, a piezo electric crystal wouldn't need a positive, unless the sensor has electronics to boost the signal.  

 

anyone know how to test the sensor?  like, how to make it talk without inducing knock on purpose in the actual engine?  I could then test one pole or the other to see what is the output...  

Share this post


Link to post
Share on other sites

Causing knock at low load shouldn't hurt the engine.  You really need to do it to find out the actual frequency that your engine produces when knock is present so that you can properly set up the bandpass filter to eliminate noise.  However, I believe you can calculate a rough estimate using bore size.  Knock sensors are essentially a microphone attached to the block, so tap on the block with something metal near the sensor.  You might be able to measure the output using a multimeter, but if you have access to an oscilloscope, that would be better.

Share this post


Link to post
Share on other sites

Actually, I tested mine today because I was curious and will probably head down this road in the future.  Tested with a DIY soundcard Oscilloscope, hooked the two leads up and tapped the block with a hammer and this is what the response was.  So then I hooked the leads up to a multimeter and tried tapping with a hammer to see the response.  It wouldn't register at all, the signal is just too small.  

 

post-6369-0-75516900-1461191162_thumb.png

Edited by Namor

Share this post


Link to post
Share on other sites

Lol I just tried to click the X in your picture to close the picture. 

 

I think that is why my tuner was suggesting the low load knock, an explosion is probably a bit more acute of a sound no?

 

Also curious as supernova was saying, how did you wire the knock sensor? Did you run a positive and a signal return?

Share this post


Link to post
Share on other sites

For the test, I just hooked the test leads directly to it, no external voltage necessary.  Piezoelectric knock sensors are their own source, the vibration of the crystal generates a voltage.  You should only have to ground 1 of the leads (probably to sensor return) and hook the other up to the signal wire and then it would all be setup and testing inside of tunerstudio.  Actual knock would probably be louder, but also you need to figure out where the peak is so you can ignore everything else.  This was done with the engine off, but I just did a little further testing and here is what I found.  

 

 

post-6369-0-45899900-1461196220_thumb.png

This is what it looks like idling around 900-950 rpm on the same scale I used before

 

 

 

post-6369-0-04540700-1461196225_thumb.png

Idle with scale change of x10

 

 

post-6369-0-36661500-1461196229_thumb.png

Rev to probably around 2500 RPM

 

While idling and revving it is noisy enough that you can indeed use a multimeter to test.  Just hook up to the two connections, set to AC Voltage and rev the engine a bit.  You should see the value jump and follow engine speed/noise.  I saw mine go from like 30-40 mV up to 90mV when revving.  

 

Edited by Namor

Share this post


Link to post
Share on other sites

hi!

 

seattle jester you beat me to the question...

 

so, Namor, one of the lead to the engine block, the other one to..  and oscilloscope or the megasquirt, I guess..   and If I don't get anything, I switch them up?  

 

Namor, quick, unrelated questions. the oscilloscope, did you follow instructions to set it up?  if so, which ones and how much trouble was it?

 

thanks!

Share this post


Link to post
Share on other sites

There is no positive or negative on the knock sensor, at least there shouldn't be.  I hooked up the O-scope ground to one terminal and signal lead to the other terminal, for megasquirt you will likely be grounding one terminal to sensor return and the other to whatever input you decide to use.  You can do the same for testing with a multimeter.  Red lead to one terminal, black to the other.

 

 The O-scope I use is a DIY one.  instructions here http://makezine.com/projects/sound-card-oscilloscope/  It's not the greatest, but as you can see, it gets the job done for what I need and it was cheap.  

Share this post


Link to post
Share on other sites

Yeah, the 'sensor return' is essentially an isolated ground within megasquirt solely for the purpose of acting as ground for different sensors.  It keeps noise from large loads like fans, fuel pumps, etc from affecting sensor readings.  I had a constantly fluctuating TPS reading until I realized that my o2 sensor needed to be grounded to the chassis, not the sensor return.

Share this post


Link to post
Share on other sites

Thanks a million guys.  

 

I'm not certain as to when I'll get to do this (I have to find some significant issue with fuel/ignition with my 5-6 cylinders, then change the oil (ideally take the pan off and re-gasket it and more) but i'll definitely post results when i do..

 

oh, and to induce knock at low RPM, I guess all I have to do it to advance timing to unreasonable levels in one of the boxes on my table, and it'll just register as a ping?

Share this post


Link to post
Share on other sites

As I understand it.

 

There is a combined signal return it basically is an internal ground so that all the sensors are zeroed to the same as I understand. Your o2 sensor and coolant sensor should both share this signal ground.

 

The other wire will go to the input as to whatever you set it.

 

I had a discussion with a friend last week about this, but I believe yes advance timing (bringing it closer to BTDC) very slowly, once you get to a certain point  you will register a ping as the flame front from the spark expands the excess pressure causes a remote detonation event elsewhere in the cylinder. 

Share this post


Link to post
Share on other sites

You really should have some way of monitoring the amplitude and frequency of knock as you induce it.  

 

As for grounding, I would be careful with grounding your O2 sensor to the same location as all other sensors unless you can wire the internal heater up completely separate.  My Innovate LC-2 only has a single ground connection, and when wired to the sensor return it would cause all the other sensors to fluctuate even with the car off.  

Share this post


Link to post
Share on other sites

Wouldn'it work to simply ground these potentially troublesome sensors on the block at the same place you ground your megasquirt?

While the potential would be the same, at least the current wouldn't go throught the computer itself (shortest route is right to the block)...

I know improper grounding has an impact on O2 sensors; when i hooked mine up, i had to play with the output values to get a reading in the megasquirt that was similar to that of the physical gauge. Went away when i re-grounded the sensor better and i was able to use proper stock signal out of my lc-2. (Not sure if the O2 sensors ground wire is involved into the circuit that pulls power to warm the sensor up.. If so that might wreak havok on everything by pulling all sorts of power backwards...)

Edited by supernova_6969

Share this post


Link to post
Share on other sites

Hmm, that may be something to look into.

 

The IAT and the coolant definitely share the signal ground, I think TPS might as well. 

On mine I have IAT, Coolant, TPS, MAP, and Fuel Pressure sensor.  Essentially you want every sensor you can to use the sensor return, unless there is a problem with doing that.  

 

As for why to not ground to the block, I can't get a real technical answer because I didn't build my ECU and I haven't analyzed the circuit diagrams but if the 5vref is referenced to the 'sensor return' and not the megasquirt ground, there could be variations in supply voltage if referenced to the engine block.  The real reason I have done mine this way is because that is the way that megasquirt says it should be done.   

Share this post


Link to post
Share on other sites

Join the conversation

You can post now and register later. If you have an account, sign in now to post with your account.

Guest
Reply to this topic...

×   Pasted as rich text.   Paste as plain text instead

  Only 75 emoji are allowed.

×   Your link has been automatically embedded.   Display as a link instead

×   Your previous content has been restored.   Clear editor

×   You cannot paste images directly. Upload or insert images from URL.


×
×
  • Create New...