kevincnc

I would like to be able to see knock counts that the KLR sends from pin 15 to the DME on my Vitesse Piggyback.

I looked at the signal a couple of years ago on a scope, and IIRC it was about a 7-8 volt spike. The SMT6 input needs 5V, and I think John said the pulse duration is too short for the SMT6 to capture.

So...I want to design a circuit or better yet find an existing one that will convert the signal. Here is what I think it needs to do:
1) Clip the voltage to 5VDC.
2) Extend the pulse width enough that the SMT6 can read it (I don't know the official pulse width that the SMT6 can reliably see yet)
3) Possibly include a buffer? If the pulse has to be widened so much that it "runs over" the next one, several knock counts could look like one. Maybe this isn't really a problem?

If anyone has done this or knows how, please post it here. If anyone else wants to know and no solution is presented, also let me know. I'll figure it out one way or another and let you know.

fast951

There is a much simpler solution. The only issue I see is making the car knock in order to verify the knock Y/N and capturing it. Causing knock on purpose is not healthy to the engine! We looked into creating a device that simulates engine knock to be captured by the knock sensor (while the engine is running). It ended up a time consuming and expensive development project.

Look for a new product release soon

Josh B

I am interetested in figuring out how to capture knock count as well. I've read over the threads about making a counter but it would be so nice to be able to log it on the SMT6 or the Zeitronix..

kevincnc

So John your new product release will incorporate this new solution? Will it be built into the new software (should I wait to upgrade to V-FLEX?)

The problem is that I saw a couple of knock counts per gear even at 15 PSI with both the 26/8 and 27/6. Adjusting the FQS helped but to make the counts go away completely the car felt slow. Maybe with the Stage II turbo I won't see any. I thought the consensus was that a few counts are ok, but who knows? It would be really nice if someone could give a clear explanation of how knock counts relate to timing being pulled. John I think you told me a long time ago that the software looks at an internal register to know when/how much timing to pull rather than that pin. My motor only has 36k miles and less than 2% leakdown on all cylinders, and I'd like to keep it that way.

reno808

Take a screwdriver and hammer and bang it near the knock sensor

Bri Bro

You need a Retriggerable One-Shot like a 74LS123. The input voltage should be clamped to 5V using a resistor followed by two diodes clamped to 5V and ground.

kevincnc

Thanks, that looks like it would work. Do you know what the pulse width needs to be for the SMT6? Funny, I found this comment about it searching for a datasheet-

"I call these devices one shot in the foot.

I will, if needed, use 5 chips to avoid using a one shot in the foot. The one shot in the foot chips are the most noise sensitive chips you can find. The mixture of analog and digital circuitry makes for a device that acts different from all other chips of the same family."

fast951

Let me guess you are running 91oct with 10% Oxygenated fuel? Keep the FQS set to pull timing.

The software in the KLR pulls the timing if it sees knock. The Y/N flag is just a indication. The DME has nothing to do with retarding ignition to handle knock.

The solution we will be releasing will work with the existing V-FLEX. The V-FLEX was designed to be the infrastructure platform for many things we want to do. Existing V-FLEX customers, will be able to add whatever subsystem they need, there will be no need for a V-FLEX software upgrade.

Bri Bro

It depends on the application. I have used them with and have never had any problems. A typical application is were a non time critical pulse needs to be generated such as a pulse stretcher. Some of the slower one-shots like the 555 timer had slow rise time and could cause false triggers to high speed digital logic. The 74HCT123 has very good rise/fall time and is very dependable. If you try and use one-shots chained together for precision timing of control circuits, it just will not be reliable.

As far as the sample rate of the SMT6, not sure but the pulse should be 2x the sample rate. If its 10Hz, the pulse should be 0.2 seconds.

As John stated above, it is just a tool to see if pinging has occured.

kevincnc

Yep, crappy Arizona gas. I am keeping timing pulled, just not as far as it will go. We'll see how it does with the new turbo. I am serioisly considering buying a 55 gallon drum of race fuel to mix in when I fill up. Thanks for the explanation.

Ok, I might try it if I still see any counts after putting in the new turbo and V-Flex.

Of course I will have to finish wiring the spec car first