EFI University

[Pantera EFI]

Which crankshaft sensor is best for Knock Detection:

The Hall sensor or Variable Reluctance sensor ?
One sensor produces a "sine vave", the other a "square" wave.
Does the type even matter ?

Which type is more sensitive ?

My preference is the late GM (58x) Gray LS-7, L92, LS-9 sensor.
Low cost, availability around the world, robust.

What is your experince?

Lance

[HOODEY]

knock detection or measurement of speed?

[Pantera EFI]

Knock Detection IS measurement of speed, the speed of the crankshaft.

I know that with the GM sensor this method is improved when used on an engine with eight cylinders AND a greater crank rotational mass.

I know that the Long 60-2 target is fitted to the crankshaft.
I know that the GM LS-x engine has the crank Target Wheel fitted inside the engine.

Could the "placement" AND mounting method of the TW be a factor ?

Lance

[Pantera EFI]

Sad, OK.
We, good tuners in general, have very little experince with knock in running engines.
Allow me to make a a current example, a true report from David at Water Thunder:
Air boat, GM LS-2 (58x), good summer tune.
The owner reported a "stumble" at 4700, then again for a longer time at 5300-5400 RPM, then ran fine to prop redline 5700.
After six months of good performance, the owner reported that he "now runs the boat harder", in cooler winter air, maybe "winter" fuel blend.
Data Logger found that "timing retard" of 55 degrees happened
after tooth 32 on a WDBug of 59.
We inspected 4 & 7, no sign on the sparkplugs, my suspect was 7 as it is in the rear of the stock LS-2 intake.

The engine advance was reduced by two degrees, the "miss" stopped.

No broked pistons, sparkplugs, wallet.

Lance

[Turboivo]

Hi Lance,
Do you mean that lean knock on 4 or 7 cyl created wrong wheel decoding and wrong rpm estimation?

[Pantera EFI]

Yes, you are correct.
Fine work "Turbo", if you are aware of this than your system must be able to decode the
crankshaft acceleration correctly.

The requirement of "soft" rev control must also exist and be enabled with an amount of timing retard.

My normal value I use is 55 degrees.

Lance

[mrx]

interesting topic, i've worked on missfire detection that works exactly that way with measureing of segment times.

but i've never heared that knock can be detected that way... i can only imagine that you can detect massive knock with it, but slight knock, i think will not have a big, repeatable effect on segment times. Therefore combustion instabilities from cycle to cycle are just too big...

but if you have some data concerning this topic, it's for sure very interesting...

Best regards

[HOODEY]

Please enlighten us on thi subject.

How does the system know it is knock and not rapid acceleration?
How is speed use to determine miss fire?

Hall effect sensor technology has improved significantly.Some even have on board controllers. The differential gear tooth sensors are extremely accurate and repeatable.

I learn this while developing a small inertia dyno which has to measure the speed every rotation accuaratley otherwise the data becomes useless.

ATs625 by allegro is a good example of this technology.

[mrx]

For Missfire i can explain...
first an algorithm calculates the teeth time for each segment (cylinder)... when then suddenly the teeth time during the combustion phase of the segment of a certain cylinder is significant longer then normal or the other cylinders a misfire is detected...

this is a very simple description of the process and in real it's far more complicated, because you have to take wheel speed sensors into account to detect rough road conditions to avoid false detections and also the dengine design has a big impact on misfire detection thresholds.

Also the formula for the segment time calculation is different for the amount of cylinders you have.

This is how it works in general and is standard in nearly every current production car.

But to use this kind of data for knock detection is new to me... but maybe he found something out...

[Pantera EFI]

We know that a "missing tooth" will be seen as a decrease in crank acceleration, not normal on the why "up".

This is the method also able to detect engine "phase" without the cam sensor,
though the rule is not "cast-in stone".

My interest is that the 2006 GM LS-7 (58x) crank sensor seems better at "detection" than the older VR sensor ?

Lance

[Seth Francis]

So, is GM using crankspeed to determine knock instead of a standard knock sensor?
The reason I ask is: I just got done playing with an 08 z06 on the dyno. During pre-mod runs , we saw peak power of 414 wheel, then a set of long tube headers, high flow air filter and tube, and a little EFI Live tweaking, we were only able to get 429.5 wheel. We noticed that knock limit was being hit at about 3400 rpm with nothing audible from the engine, but it removed 1.5 degrees immediately. I feel that the timing was conservative at 19 degrees in the map. I know that LS1 engines have 2 knock sensors below the intake manifold, but the guy working on the car said that this block was different and the knock sensors were mounted on both sides of the block under the exhaust manifolds.

[Pantera EFI]

NO, there are also two piezo's on the block in the area you stated.

My belief is that they use both methods and other methods not mentioned.

My opinion is that what you found is correct and "tooth acceleration" measurement accounted for the spark retard.

The EGR valve is also missing on most all of the 2006+ GM engines.
This removal was made possible by fitting the Long 60-2 (58x) target to the crankshaft.
Now with "degree based" spark instant calculations major timing lag is eliminated.

Lance

[Turboivo]

Hello Lance, what type(construction) of flywheel does these engines have? Is it two part/mass flywheel?