Jump to content
SAU Community

Recommended Posts

For those that like to solder!

OK Jaycar sell a knock sensor kit that is designed to interface with their ignition system (but the system starts to get expensive especially with the programmer etc etc)...but the output is luckily linear in relation to knock intensity; 0 -5v range. You can take the output of their knock sensor kit and wire it into their bar graph voltage display kit (set at 0-5v input) and you've got a nice visual knock warning with graduations!!!

All for under $40

productLarge_10364.jpg

productLarge_9569.jpg

Link to comment
https://www.sau.com.au/forums/topic/230430-cheap-knock-sensor-kit/
Share on other sites

looks promising, bu i wonder how accurate it is... nobody really seems to use there stuff though, so its hard to get a decent review. i've only heard of one or two people using the ignition system, one said it was quite succesful (LPG setup, but others said when you add up all the costs of various components, you can afford more of a known thing.

dont suppose it has any form of adjustment for frequency?

it will revolutionize the tuning world when someone markets a an actual monitor for knock that has capabilities to display a set figure/sign for knock, but this really doesnt do that.

ionization current ftw!

PS - for that money i'd make a set of det-cans, wayyyy more accurate adn reliable (providing you are)

Edited by VB-

This kit listens between 4.8-6.4khz...it has sensitivity adjustment for eliminating background noise.

Yep I agree the full jaycar ignition system would get fairly costly, that's why I used the output straight into the voltmeter so I've got a cheaper interface.

Anyway I'll give it a go and see how it is.

  • 2 weeks later...

VB here's something of interest!

I've ditched the bar graph volatage display in favour of the >$8 multimeter Jaycar have. When you wire this up you now an output in Volts so can use that as your degree of knock; anything under 1.00V is safe, anything over needs attention :happy:

You basically free-rev the car, set the pot so free revving = 0.00v and it eliminates background noise, even a slight knock will spike well over 1.00v instantly as it's actively listening in that frequency range.

It works great and can all fit in a jiffy box with 3 inputs (knock sensor, 12V + & -) just have a rotary dial to set sensitivity and you've got an LCD display from the cheapo voltmeter, an accurate and sensitive det-box for $30!!!

but you'd still need to set it up using something more reliable wouldnt you? and if you have access to it why not just use it instead of building your own? unless your talking more for long-term install rather then straight tuning. in which case it sounds awesome

For those that like to solder!

OK Jaycar sell a knock sensor kit that is designed to interface with their ignition system (but the system starts to get expensive especially with the programmer etc etc)...but the output is luckily linear in relation to knock intensity; 0 -5v range. You can take the output of their knock sensor kit and wire it into their bar graph voltage display kit (set at 0-5v input) and you've got a nice visual knock warning with graduations!!!

All for under $40

Looks interesting but how does it sense knock? What exactly is it measuring? Does it have a microphone listening for knock or is it measuring the vibrations causd by the knock or what? If it works I would be glad to buy one.

productLarge_10364.jpg

productLarge_9569.jpg

VB here's something of interest!

I've ditched the bar graph volatage display in favour of the >$8 multimeter Jaycar have. When you wire this up you now an output in Volts so can use that as your degree of knock; anything under 1.00V is safe, anything over needs attention :whistling:

Does the knock sensor do peak hold and for how long? If it peak holds for less than a second then you have no hope of seeing the result on your multimeter because multimeters are quite slow to react. For tuning a multi might be OK but for general road use a bar graph will catch your attention better and can be seen via peripheral vision. You have to look directly at a multimeter to read it properly.

Not a bad idea though. I would be tempted to use this kit with a PFC to cut ignition/fuel if knock was detected rather than flashing a stupid warning light and hoping you see it in time to save your engine.

but you'd still need to set it up using something more reliable wouldnt you? and if you have access to it why not just use it instead of building your own? unless your talking more for long-term install rather then straight tuning. in which case it sounds awesome

Not really as once you've eliminated engine noise (background) over the rev range any slight knock will immediately shoot you straight over 1.25v anyway (that's the trigger frequency its looking hard for) So sticking to under 1.00v knock you know you're safe. I'm getting a max of 0.65v and there is not a hint of detonation, even that I could use as a yard-stick to tune against.

Would be good to put it on a car with a PFC and compare knock values though :D

Does the knock sensor do peak hold and for how long? If it peak holds for less than a second then you have no hope of seeing the result on your multimeter because multimeters are quite slow to react. For tuning a multi might be OK but for general road use a bar graph will catch your attention better and can be seen via peripheral vision. You have to look directly at a multimeter to read it properly.

Not a bad idea though. I would be tempted to use this kit with a PFC to cut ignition/fuel if knock was detected rather than flashing a stupid warning light and hoping you see it in time to save your engine.

A capacitor and a diode wired on the output gives you peak hold ability :D

  • 6 months later...

Well if anyone cares,

I've found the frequency for knock in a 86mm bore engine is about 6.7khz, which is a hair above the jaycars low pass filter. So out of the box you're not

getting the whole story. I've bought 2 kits to see if I could some way increase that low pass filter to about 7.5khz.

I am very keen to find an accurate way of detecting knock as I want to do my own tuning and have got a wideband kit but there seems to be no cheap solution. Guilt-Toy reckons that the PFC is too conservative with its knock detection but I don't know if that is a problem with the PFC hardware or software or the nissan sensors. He uses a K-mon knock detector with two special sensors (total around $1500) so I will try to get the answers from him.

thinking more into it, i rkn the knock sensor itself will be more of an issue than the display. what sort of mods have you got maximajim?

I have the regular mods, greddy im, 760cc injectors and a custom 62-1 turbnetics turbo. I also have an AEM EMS standalone. It has knock correction with high and low filters. It is possible to add fuel and/or subtract timing in the event of knock detection. We can't always use det cans so in theory this system should work well.

Unfortunately it needs to be in a 0-5v scale and the nissan knock sensors don't suit it well enough. By using the jaycar kit with proper freq filters it should work very well.

kiwi: det cans

maximajim: would be useful but plenty of people dont use it and dont have an issue, just be a little conservative in the tune. in theory it all works well but i just cant see it coming together as your expecting it to. that said, you try it and let us know how it goes :sweat:

i'm keen to find out, just not to be the guinea pig :P

  • 2 weeks later...

I've built one of these and hooked the output of it up to one of the inputs on my Lambda sensors data logging software.

Gone for a few runs and compared it to the PFC's knock level. The knock boards output seems to rise in volume directly in line with revs. Small spikes occur but you can see pretty clearly what is normal noise and what is abnormal.

Havn't had a chance to actually go out and make the engine knock to test it properly though. I'll report back when I do.

I'll also grab some screen shots of the logs off the laptop when I get a chance.

kiwi: det cans

maximajim: would be useful but plenty of people dont use it and dont have an issue, just be a little conservative in the tune. in theory it all works well but i just cant see it coming together as your expecting it to. that said, you try it and let us know how it goes :(

i'm keen to find out, just not to be the guinea pig :/

Well the boards should be here on Tuesday. But I have disagree, I'm sure with the a PFC or other [non-knock] standalone it would be pretty useless. But the EMS I have is fairly powerful, not only will it react to knock, it has the ability to ignore mechanical noise. All you have to do is pull timing way back and go for a test run, that will be your base line which will be in volts. Any voltage above that will be treated as knock, and the higher the voltage, the higher the timing retard and fuel addition. If knock sensor 1 sees knock, it will add fuel to cyl 1,2 and 3. Knock sensor 2 does the others.

That said, it's a really safety device, there are conditions on the road which can't be replicated on a dyno where the car is tuned.

Create an account or sign in to comment

You need to be a member in order to leave a comment

Create an account

Sign up for a new account in our community. It's easy!

Register a new account

Sign in

Already have an account? Sign in here.

Sign In Now


  • Similar Content

  • Latest Posts

    • I had 3 counts over the last couple of weeks once where i got stranded at a jdm paint yard booking in some work. 2nd time was moving the car into the drive way for the inspection and the 3rd was during the inspection for the co2 leak test. Fix: 1st, car off for a hour and half disconnected battery 10mins 4th try car started 2nd, 5th try started 3rd, countless time starting disconnected battery dude was under the hood listening to the starting sequence fuel pump ect.   
    • This. As for your options - I suggest remote mounting the Nissan sensor further away on a length of steel tube. That tube to have a loop in it to handle vibration, etc etc. You will need to either put a tee and a bleed fitting near the sensor, or crack the fitting at the sensor to bleed it full of oil when you first set it up, otherwise you won't get the line filled. But this is a small problem. Just needs enough access to get it done.
    • The time is always correct. Only the date is wrong. It currently thinks it is January 19. Tomorrow it will say it is January 20. The date and time are ( should be ! ) retrieved from the GPS navigation system.
    • Buy yourself a set of easy outs. See if they will get a good bite in and unthread it.   Very very lucky the whole sender didn't let go while on the track and cost you a motor!
    • Well GTSBoy, prepare yourself further. I did a track day with 1/2 a day prep on Friday, inpromptu. The good news is that I got home, and didn't drive the car into a wall. Everything seemed mostly okay. The car was even a little faster than it was last time. I also got to get some good datalog data too. I also noticed a tiny bit of knock which was (luckily?) recorded. All I know is the knock sensors got recalibrated.... and are notorious for false knock. So I don't know if they are too sensitive, not sensitive enough... or some other third option. But I reduced timing anyway. It wasn't every pull through the session either. Think along the lines of -1 degree of timing for say, three instances while at the top of 4th in a 20 minute all-hot-lap session. Unfortunately at the end of session 2... I noticed a little oil. I borrowed some jack stands and a jack and took a look under there, but as is often the case, messing around with it kinda half cleaned it up, it was not conclusive where it was coming from. I decided to give it another go and see how it was. The amount of oil was maybe one/two small drops. I did another 20 minute session and car went well, and I was just starting to get into it and not be terrified of driving on track. I pulled over and checked in the pits and saw this: This is where I called it, packed up and went home as I live ~20 min from the track with a VERY VERY CLOSE EYE on Oil Pressure on the way home. The volume wasn't much but you never know. I checked it today when I had my own space/tools/time to find out what was going on, wanted to clean it up, run the car and see if any of the fittings from around the oil filter were causing it. I have like.. 5 fittings there, so I suspected one was (hopefully?) the culprit. It became immediately apparent as soon as I looked around more closely. 795d266d-a034-4b8c-89c9-d83860f5d00a.mp4       This is the R34 GTT oil sender connected via an adapter to an oil cooler block I have installed which runs AN lines to my cooler (and back). There's also an oil temp sensor on top.  Just after that video, I attempted to unthread the sensor to see if it's loose/worn and it disintegrated in my hand. So yes. I am glad I noticed that oil because it would appear that complete and utter catastrophic engine failure was about 1 second of engine runtime away. I did try to drill the fitting out, and only succeeded in drilling the middle hole much larger and now there's a... smooth hole in there with what looks like a damn sleeve still incredibly tight in there. Not really sure how to proceed from here. My options: 1) Find someone who can remove the stuck fitting, and use a steel adapter so it won't fatigue? (Female BSPT for the R34 sender to 1/8NPT male - HARD to find). IF it isn't possible to remove - Buy a new block ($320) and have someone tap a new 1/8NPT in the top of it ($????) and hope the steel adapter works better. 2) Buy a new block and give up on the OEM pressure sender for the dash entirely, and use the supplied 1/8 NPT for the oil temp sender. Having the oil pressure read 0 in the dash with the warning lamp will give me a lot of anxiety driving around. I do have the actual GM sensor/sender working, but it needs OBD2 as a gauge. If I'm datalogging I don't actually have a readout of what the gauge is currently displaying. 3) Other? Find a new location for the OEM sender? Though I don't know of anywhere that will work. I also don't know if a steel adapter is actually functionally smart here. It's clearly leveraged itself through vibration of the motor and snapped in half. This doesn't seem like a setup a smart person would replicate given the weight of the OEM sender. Still pretty happy being lucky for once and seeing this at the absolute last moment before bye bye motor in a big way, even if an adapter is apparently 6 weeks+ delivery and I have no way to free the current stuck/potentially destroyed threads in the current oil block.
×
×
  • Create New...