Carpokes Porsche Forum

Do you know where you sampled the ref sensor signal? From the connector or from the 8051 pin?

whalenlg wrote: Sat Mar 21, 2026 5:47 pm Do you know where you sampled the ref sensor signal? From the connector or from the 8051 pin?

Definitely not any external connector, since the digitized version can only be found on the DME board. I forget which pin of which chip I probed for that one but this should provide enough info:

Note that the digitized reference sensor signal is pulled low when the raw signal meets the threshold on the way up. The S100 actually pulls the output high for this event, but the signal is then inverted via one of the NOR gates of a 74LS02 (S705) before reaching the 8051. The speed sensor output pulses are not inverted.

So I was probably probing it at either the input of the 8051 or the output of the 74LS02 to get the logic that the software sees.

For some reason the S100 ref sensor output latches and has to be reset by the software. I have no idea why it works that way - the state of the output is never read anywhere that I can see, it's only used to trigger the interrupt, and the interrupt flag stays on until it's dealt with.

Thanks for that @johnb !

The S100 is clearly lengthening the ref sensor pulse - it's a small spike on the input to the S100, but it's more like a 40% duty cycle square wave coming out. Guessing it counts speed sensor pulses to form that output signal.

I fixed a very esoteric bug in my 8051 model (yes, claude does make mistakes) and that injector pulse is now 2ms. Phase is somewhat different, so I need to check on that.

I think things are very well sorted - I confirmed that my very short KLR trigger pulse (20usec) looks valid according to your KLR page:
https://jhnbyrn.github.io/951-KLR-PAGES ... iming.html

whalenlg wrote: Sun Mar 22, 2026 1:24 pm Thanks for that @johnb !

The S100 is clearly lengthening the ref sensor pulse - it's a small spike on the input to the S100, but it's more like a 40% duty cycle square wave coming out. Guessing it counts speed sensor pulses to form that output signal.

I fixed a very esoteric bug in my 8051 model (yes, claude does make mistakes) and that injector pulse is now 2ms. Phase is somewhat different, so I need to check on that.

I think things are very well sorted - I confirmed that my very short KLR trigger pulse (20usec) looks valid according to your KLR page:
https://jhnbyrn.github.io/951-KLR-PAGES ... iming.html

The S100 doesn't count anything - the ref output just stays on once triggered, until it's turned off by the 8051 (like a latch).

Great to hear about the progress!

Looking at the schematic for the S100 to 8051, I only see a -WR and A12 signal connection - that doesn't look sufficient for the software to reset it.
On the other hand, the S100 has speed sensor inputs, so I still surmise some counting is happening from speed sensor pulses.