Tom M'Guinn

I'm finally getting around to making up one of those Curtis knock counters, and have a few questions for the crowd:

1. In the factory schematics, the knock signal line (Pin 15) and the black rectangual 12-pin diagnostic connector is not shown in the 86 schematics, but are shown in the 87 and later schematics. Is that just a documentation issue, or does the 86 harness actually lack that connector?

2. In previous threads, people say that Pin 15 tells the DME when knock is detected. In the factory schematics, however, Pin 15 seems to be connected to nothing but the 12-pin diagnostic connector, while KLR Pin 24 appears connected to the knock line on the DME Pin 31. Anyone ever look into the signal on KLR Pin 24?

samluke

Have you searched the existing threads for knock counting? There is a lot of info out there, with different approaches.

toddk911

So if you tie in to pin 15, what will the output signal be? a flat line until knock is detected and then a spike? Or will it spike for each count of knock?

I would just like some type of "idiot light" for when any knock is detected.

Tom M'Guinn

Yes, that's why I posted the above-- the old threads seem to say KLR pin 15 is used by the dme, which appears incorrect based on the schematic (and the spare harness I inspected). If someone corrected or clarified that, I missed it. Also found it interesting the 86 schematic does not show this pin or the diagnostic connector (despite it clearly being present on my 86 harness).

Fluidplay

Tom - I have an 86 and I spliced into the #15. Mounted mine on the console.

blown 944

I am pretty sure that is where I tapped into on my 86. That or at the diagnostic plug. You may want to ask josh he has all this right off the top of his head.

Tom M'Guinn

Thanks, it appears to be just an issue with the manual -- just confirmed my 86 has the diagnostic connector. Guess it was undocumented originally...

JohnKoaWood

Pin 15 is the knock detection impulse sent not to the DME, but to the diagnostic port.

It is a small, narrow pulse (IIRC less than 5 VDC, and less than half a second in durration, most multimeters WON'T detect it), the curtis counter advances ONE for each time it detects the pulse (the pulse is sent EACH time the KLR detects a knock) hence the counter accumulates the counts. FWIW the curtis counter will count anything over 1.5 VDC 30ms in durration, but I could be wrong, it has been a while since I looked at the tech specs on it...

I posted a DIY on how to add a backlight and reset button to the curtis knock counter, the circuit is IN the counter, just not used by curtis... there is ONE component missing to enable the reset function (resistor switched to positive 12VDC, added in the DIY), and a light on the power pins supplies backlight.

I have modded two counters to date, and never got around to modding one for my own car, now that I am going stand alone I do not need to do one for my self...

I will see if I can dig up the mod thread, as it never got migrated to the DIY section...

JohnKoaWood

https://rennlist.com/forums/944-turb...t=modification

there ya go!

Van

I don't think an "idiot light" would work. The blink might be too fast. I have my pin 15 datalogged by my traqmate, and I get a digital signal spike for just 1 sample rate at a time. (I think I have it set up for 20 Htz.)

Here's a picture that shows the results:

JohnKoaWood

I think I remember you posting this before... IMHO you should up the sample rate if you can, as the pulses on pin 15 are VERY short in durration... (under 50ms) and you COULD be missing some in between your 20hz rate (you are sampling 20 times per second(every .05 seconds), but the knock signal is per knock and less than .005 in duration... but I could be wrong as I am working off notional data, and not scoped results...

I would love to see what would happen if you turned it up to 100Hz or even 1Khz... if that is possible...

what you are seeing now PROVES you have knock events, but you could be missing 3 in between samples....

And yes a simple light would NOT work, unless it was tied to a capacitor to hold the light on long enough for it to register in your brain... (that was the next itteration of my mods to the counter.. adding a red warning light that would stay on longer the more detection events were being counted.... but I got dragged off center mass on the effort, I might be able to resume it one day....) just wont be able to test it in my own car...

Van

I checked - yes, I sample at 20 Hz, I can go as high as 40 Hz. I'll do that for my next track event and see if it makes a difference.

JohnKoaWood

I would expect you will double the counts you are seeing, as it looks like the 3 blue peaks are diplaying detection at your 20HZ frequency... I am willing to bet you will see simmilar groupings, but twice as many events...

toddk911

Ok pin 15 is supposed to be solid blue however I see two solid blue on klr. So which one? Lol

Fluidplay

Side the cover off the plug, the pins are numbered.