knock sensors

[slighty_sykotic]

And just to agree with Sergei and the others,

There are several 4agze knock sensors. Try a map one on a afm one, and you will get a check engine light. I have tried, the knock sensor came from a WORKING car, put on my engine, check light on, put sensor back to original car, works fine.

Get knock sensor from EXACTLY the right motor, no check light, works fine.

[MAGN1T]

You can get to the moon and back on 48K. If you want to digitise a 4KHz signal you need a clock frequency of 8KHz.

It's all the other rubbish that takes up processing power.

All older ECUs use a separate board for knock filtering. If you do it (1980s style) it only takes at the most 2 op amps. DSPs didn't exist then.


BTW I've got the knock sensor in my Mitsi hooked up to the knocklink and also hooked up to a preamp which feeds into the car stereo so I can hear it coming out of the speakers, oh and the datalogger too...... try telling me sonething I don't know. Try googling "42"

Sergei, you seem to have problems getting most things to work.

Steve

[sergei]

Sergei, you seem to have problems getting most things to work.
Steve

I don't know but most of my things work .
Unlike you I have ability to learn (and I can admit that I made mistakes).

You seem to argue like a religious nut, when the evidence is opposite to your believes.
Now about 3 or 4 people stated when they used wrong working sensor they got CEL on, and you say "nah, they are wrong"?
You have attitude problem (finally some one with bigger attitude problem than me on TS )

You are confusing clock frequency, instructions per clock and sampling frequency.
Although some CPUs will do 1 instruction per clock (like AVR) most of them don't.

Getting to the moon and back is not a big feat, as it is all pre staged and preprogrammed, very little data input (mostly from astronauts) and certainly no FFT.

Of course you could use analogue filters etc. But it is far better to create specialised sensor. It is very difficult to create a opamp based band pass filter that will be as discriminative as physical piezoelectric sensor that resonates at certain frequency. If you pulled apart a Toyota ecu you will see that they don't have a lot of components, certainly not enough to account for elaborate bandpass filter.

Do you deny that if I use 3SGE sensor in my 3SGTE it will not work?
Or if I use 4AGZE sensor in my 4AGE?

In fact I will go as far as saying that using knock sensor from a AE101 GZE will not work with AE92 ECU (CEL).

You need to have reality check.
I have tried different sensors from different engines, and most of the time I got CEL on, while the sensor would function correctly with original engine.

Toyota knock sensors are generally very narrow band, that "ring" at specific frequency (knock frequency of the engine, and guess what every engine got this frequency slightly different).

I even have an idea how ECU can detect that there is wrong sensor,
you know how we elaborated that they ring at certain frequency, well if that frequency is not present then ECU faults the sensor, and it is a lot simpler to do a frequency counter than elaborate DSP as it involves only setting timer and interrupts for timer event and one of the multiplexed states responsible for knock signal. Well that is my theory anyway, implementation might be different...

[MAGN1T]

The sensor has only 2 wires.

If they are the wrong way round , it won't work. Reversing the wires will probably fix it.

It is an output only,one way, you can't send a signal into it from the ecu.

Maybe you'd like to explain how bore size comes into it when knock occurs close to TDC and the only measurement of relevence is the combustion chamber volume, or the shortest distance across it.

Anybody who relies on google ends up filling their heads with rubbish.

Steve

[slighty_sykotic]

Anybody who relies on google ends up filling their heads with rubbish.

How about relying on REAL LIFE STUFF.

You have never taken a 4agze map knock sensor, and put it on a afm 4agze. Nor a 4age knock on a 4agze.

If you had, you would know that IT DOES NOT WORK.

Its not much simplier than that buddy. Whatever theory your trying to explain, is not correct when it comes to 4age toyota knock sensors (and probably more).

[sergei]

The sensor has only 2 wires.

If they are the wrong way round , it won't work. Reversing the wires will probably fix it.

It is an output only,one way, you can't send a signal into it from the ecu.

Maybe you'd like to explain how bore size comes into it when knock occurs close to TDC and the only measurement of relevence is the combustion chamber volume, or the shortest distance across it.

Anybody who relies on google ends up filling their heads with rubbish.

Steve

Hah insults
The knock sensor has got only one wire .
It uses block as second, no way to reverse it.
Shows how much you know about Toyota knock sensors.
Google didn't tell me that there was a CEL when I put in a 4AGE knock sensor on the 4AGZE motor. It actually happen.

You are indeed a troll.
No more talking to you.

[fivebob]

Anybody who relies on google ends up filling their heads with rubbish.

Man, pull your head out of your arse, you'll find that (in most cases) the world smells better that way.

Google is a tool, and like any tool if you don't use it properly it doesn't do the job it's designed to do

[fivebob]

Maybe you'd like to explain how bore size comes into it when knock occurs close to TDC and the only measurement of relevence is the combustion chamber volume, or the shortest distance across it

Perhaps you should learn to use google properly and you might find the answers you seek, that is if you're actually open to new ideas

If, instead of pulling a random figure out of your arse (maybe that's why your head is stuck up there), you'd like to calculate the fundamental knock frequency of an engine there's this useful webpage.

[sergei]

I bet he is a tinfoil wearer hence against google conspiracy.
Or maybe he is so bored that he likes to stir shit.
Every single post he made had a slight (or full of) hint of disinformation (perhaps on purpose). Maybe moderators will need to look into this.
I don't like on-line shit stirrers as I can't see their faces and can't tell their intent.

[fivebob]

You can get to the moon and back on 48K. If you want to digitise a 4KHz signal you need a clock frequency of 8KHz.

No you need a minimum 8KHz sampling frequency for a 4KHz Nyquist frequency. However unless you have 145mm pistons 4KHz is too low. e.g. for a 4AG bore of 81mm the fundamental knock frequency is approx 7.1KHz so you need a minimum 14.2KHz sampling frequency.

As to processing power required to do FFT it's obvious to me that you do not understand what is involved. FFTs use complex mathematics involving imaginary numbers. If you have the the necessary understanding of the mathematics involved the this webpage explains the process involved, and even provides some "C" code to do the FFT.

From this code it's obvious (at least to me) that it will (depending on frequency resolution required) take several thousand, if not several hundred thousand, instruction cycles to complete one transform,. If you are to take a sample every 20° of crank rotation at 6000 rpm, that's 1800 samples per second, even if you only sample the knock window, say 20° either side of ignition, that's 400 samples/sec for a 4 cylinder engine.

It's all the other rubbish that takes up processing power.

Nope, most of that is relatively straight forward when compared to FFT

And in case you think I just Googled all this information and I don't really know what's involved in FFT calculations take a look at the single cylinder audio Tachometer program that I wrote in my spare time

[Mr Revhead]

now now ppl, play nice

laying out your theories and ideas with actual figures and examples is good, and lets others work out who is right or not for them selves

sergei and fivebob are very good with electronics (not saying magn1t isnt), and while it is good that they are challenged, and not assumed to be right 100% of the time, there comes a point when its clear they are correct this time and shouldnt be challenged further without some good info as back up.
however, often the rest of us learn a lot more if they are challanged!

some very good info here, and i have to admit im struggling to keep up, electronics is black magic to me!
i think i will FAQ this thread

[fivebob]

Do you deny that if I use 3SGE sensor in my 3SGTE it will not work?

Actually Sergei for the same model year they are the same sensor, and they even share that sensor with a lot of other engines that have around the same bore size (+-3mm)

Or if I use 4AGZE sensor in my 4AGE?

There you appear be correct. However it is probably due to the sensors being from different manufacturers, 4AGZE is Toyota(Denso?), 4AGE is Matsushita or NipponDenso, not because of any fundamental difference in the way the sensors work i.e. narrowband

Are The 4AGE & 4AGZE ECUs made by the same company?

[sergei]

Well, I tried to use 3SGE (and 3SFE) sensor out of ST16X in ST165 and that did not work.
As ECU for GZE and GE appears to be made by same company (denso) although it has been awhile since I saw one.

EDIT: it appears that ST162 don't have knock sensors, so I have been mistaken on that part. I might of pulled it out of NZ new ST165 or facelift....

Also looking at EPC there are two types of knock sensors listed for ST165 so that confirms the different models.

EDIT2: Oh here is the problem:
http://www.toyodiy.com/parts/p_J_1991_T ... l?hl=89615

I have also later used ST85 ECU which added to the confusion

[fivebob]

now now ppl, play nice

I am playing nice, I just don't toeraate fools gladly

laying out your theories and ideas with actual figures and examples is good, and lets others work out who is right or not for them selves

ok in that vein I'll provide a few quotes from a 1995 Texas instruments White paper on Engine Knock Detection Using Spectral Analysis With TMS320C25 or TMS320C30 DSPs

Two types of remote sensor are being used today: tuned and broadband. Tuned or resonant sensors are used in many low-end knock detection systems. Either mechanically or electronically, the sensor amplifies the magnitude of the signal in the frequency range of the knock-excited resonance (sometimes called the fundamental frequency). A limitation to this approach is that a different sensor can be required for each engine type, due to variations in the characteristic frequency. The resulting part number proliferation increases overall system costs for the manufacturer. To eliminate the cost penalty, sensor bandwidth can be made wide enough to encompass all expected variations in the fundamental frequency. However, doing so can possibly decrease system performance.

When engine knock occurs, a shock wave is generated inside the combustion chamber. The shock wave excites a characteristic frequency in the engine, which is typically in the 5 kHz–7 kHz range. Cylinder bore diameter and combustion chamber temperature are the main variables that affect this fundamental frequency. Variations in the fundamental frequency for a given engine configuration can be as much as ± 400 Hz. Larger diameters and/or lower temperatures result in a lower fundamental frequency.

sergei and fivebob are very good with electronics (not saying magn1t isnt)

Thanks for the compliment, but I'm crap at electronics, though I do have a reasonable understanding of them. I'm just an old unemployed/unemployable computer programmer/consultant

[fivebob]

EDIT2: Oh here is the problem:
http://www.toyodiy.com/parts/p_J_1991_T ... l?hl=89615

Funnily enough all those sensors listed will work on any of the SW20 ECUs, so I would assume it's the same for ST185s. I know because, after changing the connector, I've used a Gen III sensor on a early Gen II ECU and had no problems.

[sergei]

EDIT2: Oh here is the problem:
http://www.toyodiy.com/parts/p_J_1991_T ... l?hl=89615

Funnily enough all those sensors listed will work on any of the SW20 ECUs, so I would assume it's the same for ST185s. I know because, after changing the connector, I've used a Gen III sensor on a early Gen II ECU and had no problems.

Really strange.

I guess I will retract my statement about 3S case as in my case I used too many different parts in my ST165 to clearly remember.
But the case still stands for 4AGE vs 4AGZE. as it is confirmed by other people.

[AE82 FXGT]

so the knock sensor i have now is a metal one, the connector fits but im not sure about the thread as i dont have the other half of my old one...

the old one i had had a plastic housing.

anyone know if thats normal or the part has been superceded etc?

[slighty_sykotic]

Chances are your old one is different from the new one, if one is plastic and one is metal all over.

So chances are it won't work. However, try it, its not going to kill anything, it will just get a CEL and won't work.

I know at least the early 4agze (aw11) had a plastic bit on them.

[frost]

ok so i get that the bore of the engine/size/density determines the knock frequency. so what if i have over bored my 4age block to 82mm.

1. would i need to recalibrate a new knock sensor for the new frequency?
seeing as there so narrow band.

2. if i do then i'd have to use an aftermarket knock sensor system correct?

3. i would still have to have the factory one plugged in to keep the factory computer happy? because the ecu will pull timing if there is no signal correct?

[fivebob]

ok so i get that the bore of the engine/size/density determines the knock frequency. so what if i have over bored my 4age block to 82mm.

1. would i need to recalibrate a new knock sensor for the new frequency?
seeing as there so narrow band.

From this useful webpage the fundamental knock frequency goes from 7.1kHz to 7.0KHz. Also you will note that I said that 3S sensors share that sensor with a lot of other engines that have around the same bore size (+-3mm) so in the case of the 4AG, provided the sensors are the same design as the 3S,then 1mm shouldn't make enough difference to render the sensor obsolete.