fatboy7

Would anybody be interested in some electronics that would tell your e6k, e11 to retard timing when it detects a certain level of knock from the stock piezo-electric knock sensor?

I'm just trying to gauge interest, but I'll probably make it regardless for my own sake.

Basically the circuit would monitor the knock sensor, filtering out all but the frequency range of interest, measure the signal strength, convert that to a digital signal. The strength of knock would be output on a display (likely a LED bar graph, Although I'm open to suggestions). The knock level would also be used to control a digital potentiometer that would plug into the trim control set for timing control. When the knock signal reaches a user-prescribed level, the unit would change the resistance of the poteniometer, that in turn tells the EMS to retard the timing. With the right programming it could be set up to progressively retard timing as the knock signal strength increases.

The nice part about this type of system is that it would retard leading and trailing ignition at the same rate, as well as allowing the haltech to datalog the retard. The only downside, is that I'm unsure how high a sample rate the A/D converter is on the haltech.

I have a working prototype for the analog portion of the ciruit , but I still need to find a microcomputer to control the digital potentiometer, the gain of the analog circuit, the A/D conversion, and the output display.

As to the price, I haven't done enough research to set it positively, but It would likely be between $150-$250.

fatboy7

Don't all chime in at once

little rotor

Just curious if you know what the fundamental knock frequency on a 13B is. I've seen information as far as piston engines go but not for rotaries.

Cosmo Donk

I voted choice C. I had a J&S when I first bought my car but have since pulled it. I am not even sure if those kind of devices can save our cars because it happens so quickly that even if it manages to pull the timing after the first ping, your engine may already be dead. I think it is just safer to tune on the safe side of things so you don't run into that problem.

Just a few cents of opinion.

fatboy7

I understand what you are saying, and respect your opinion. And maybe you're right, especially if you expect this type of device to save your engine when tuning was poorly done.

I do know however that there are various levels of knock, and there are light levels of knock that are deemed "harmless" (certainly not audible from inside the car), but often light knock progresses into heavier knock as it causes temps to rise. If a knock sensor can detect even this light of knock, and let you know that something's amiss by warning, and retarding the timing before it progresses into heavier knock, would that not be worth the engine it saves?

The natural resonance of my S5 13B when subjected to a quick impact is around 7kHz +/- 300 Hz. At least thats what I measured by tapping on the engine with a hammer in various locations, and looking at the output waveform on an oscilliscope. I assume that knock will excite the engine block at a frequency near this resonance.

little rotor

I'm not sure if if the assumption that knock will exite the same frequency holds. But give it a try and let us know how it goes.

Even if everything (including all of the correction maps) is tuned correctly there is still the possibility of getting a bad batch of fuel. For this reason I think closed loop knock has its place.

Cosmo Donk

Agreed. I really don't know the levels that knock sensors can pick up on but if they do indeed detect things quickly enough for you to realize something is amiss in that area it would be a great tool. Save you a lot of time on the dyno when you need to work out the gremlin.

93rotorhead

Yes, I think it would be worth $150-250 if you could tell it to retard at the knock level you want.

fatboy7

I appreiciate the feedback guys. Like I said in my first post, I'll likely produce a unit for myself regardless, and if there is enough interest I can make more.

I'll build it to be sensitive to frequencies from 5kHz to 8kHz. This is the area that "most" engines that I have read about knock in. My assumption is somewhat reasonable since the natural resonance of the engine is the easiest frequency that the sudden impact of knock can excite. I guess I'll find out though.

The lowest signal that the knock sensor can pick up depends on a number of things. Noise is always the biggest issue. Most of the noise that exists in the engine is much lower in frequency than 5kHz(and can be filtered out), but there will likely be some weak higher order harmonic noise that will persist. Still, the ablity to filter out most of the noise that you normally hear from the engine greatly enhances its ablity to pick up knock. Its likely to be able to pick up knock 20x to 1000x weaker than the knock you can hear in the passenger compartment.

The initial level of knock that the unit retards would be user setable, as well as the amount it retards.

fatboy7

Oh, and one more thing. The unit right now has a rise time in the 100-200 microsecond range. I suspect with the A/D conversion, processing, and the updating of the digital potentiometer, the reaction time would be in the .3-.8 millisecond range, still thats only 20-30* of eshaft rotation at 8k rpm.

So assuming the haltech samples its A/D's at a reasonable rate(one or more times per ignition event), the retard should kick in before the next time a plug fires.

Cosmo Donk

Wow. Sounds sweet. Good luck on the whole project and I hope things work out with it. Maybe by the time it is all said and done and you are selling them I will be in the market for one

Wankel7

Wee, from teh dead.

Did it work?

LinuxRacr

I used the J&S on my car for over 2 years, and it worked flawlessly! Saved my engine more than once.

Claudio RX-7

iTrader: (1)

The J&S is a pretty good buy, but kinda pricy if im not mistaken. And then theres the MSD Knock Sensor Alert but that does nothing to control the ignition.

The MSD unit cost around less than 200 bucks, anybody upto the task of modifying it so that it would put out a 0-5v signal that the ECU could read and change timing from that?

Hmm.. Haltech needs to add more compensation options, like the AEM, it can compensate the maps a certain % depending on the reading from that particular input.

RETed

Are you talking about a single channel unit?
You're living a lie...


-Ted

pistonsuk

This sound like a great idea. I just wen to the pick-n-pull (Ogden)and got a knock control module and sensor from a Blazer ($10). This unit should read 8-10V when there is no knock and pull low when there is knock. It has a design frequency of 5200 hertz, and they produce a signal between 4800 rpm and 5600 rpm.

I was planning on just datalogging it at first and maybe building a circuit to tie into the trim signal.

This is where I got the idea: http://megasquirt.info/

Scroll down on the left and click on knock sensing.

Have you done any more with this lately? Are you still in Utah, I'm in Logan until May.

Justin

GUITARJUNKIE28

this is a good idea, but **** the stock knock sensor. use a new bosche or something that isn't 10 years old.

pistonsuk

Are you refering to the stock Blazer one mentioned in my post? If so I will test the one I have to see if it works. I read that the module should match the sensor. If it turns out to be bad AZ stocks replacements for the Blazer for $30.

Justin

89t295k

Did it work?

pistonsuk

The GM module seems to work fine. I put it on an o-scope and when hit with a hammer the signal wire drops from 10V to 0V for 2-10ms. I am still working on the circuit to make this signal useful to the trim function of the E6K.

I will post current info here:

https://www.rx7club.com/showthread.p...=1#post5422215

Justin