Megasquirt Support Forum (MSEXTRA)

Hi,

It seems to be "launch/flat shift" that makes Strange things. Is someone could tell me how I still see 1023 on ADC7 but PW to 0 without any status going in "Flatshift or launch control" ?

Here's another log from today.


Thanks guys

How do you have JS4 (ADC7) set up? I assume there is some kind of "protection" circuit between the car wiring and that CPU input. Is there a pull-up in that input circuit? If that is a standard mod described in the manuals, just point me to the proper area of the manual and I will read up on it...

http://www.msextra.com/doc/pdf/html/MS2 ... .4-32.html

I did low input circuit, the one on the top of the page. As for the ground, clutch goes to relay and relay goes to ground, witch ground the input (SPR4).

Thanks

What voltage (to ground) are you seeing at the junction of the two diodes and 1K resistor; both "switch on" and "off"? How about posting pictures of the mod, maybe we can spot some problem, like a diode connected backwards.

I didnt installed diodes. Just the resistor.

Why were the diodes left out? I think their purpose is to protect the CPU input; let's hope that hasn't been damaged.

Voltage to ground was 0V all the time. Without grounding, 4,98V.

how I could still read 1023 on datalog when I get the "problem" and clutch isnt pushed in ?

What do you recommend me to do now to fix that issue? Thanks again.

"Voltage to ground was 0V all the time. Without grounding, 4,98V."

Clarify that. If it was 0V "all the time", where does the 4.98V come in?

Let's move your voltage probing out to the 40-pin socket pin 30. What is the voltage there (again, relative to ground) when the clutch switch is both open and closed? If voltage at pin 30 doesn't change state, I would remove the MS2 daughter and test again at pin 30.

billr wrote:"Voltage to ground was 0V all the time. Without grounding, 4,98V."

Clarify that. If it was 0V "all the time", where does the 4.98V come in?

Let's move your voltage probing out to the 40-pin socket pin 30. What is the voltage there (again, relative to ground) when the clutch switch is both open and closed? If voltage at pin 30 doesn't change state, I would remove the MS2 daughter and test again at pin 30.

As requested. On daughterboard

Pin 30 to GRN with clutch switch open = +4,94V
Pin 30 to GRN with clutch switch close= -5,2mV

With daughterboard removed

Pin 30 to GRN with clutch switch open = +4,98V
Pin 30 to GRN with clutch switch close= +2mV

With the MS2 installed the ADC7 count stays stuck at 1023, even as pin 30 toggles from ~0-5V? That sure sounds like the MS2 is damaged, but we need for somebody with more experience to chime in. I'll ask again... why didn't you include the protection diodes for that input?

Wait, was there a typo in your most recent post? The only test condition shown is "clutch switch open"; no "clutch switch closed".

I've edited my previous post to avoid confusion.

Here's a datalog showing ADC7 is working fine, goes to 0 when clutch is pushed in to change gear and once released, goes to 1023. The question is, why I'm getting fuel cut without a status or ADC7 remains 1023.

Thanks by the way for all your help.

Ah, I finally got it. The ADC count is responding fine, and you feel the 1023 indicates the clutch switch is open (pedal not pushed) so that shouldn't be triggering the fuel cut.

With the "launch/flat-shift" off, are you still not getting the strange fuel cut?

billr wrote:With the "launch/flat-shift" off, are you still not getting the strange fuel cut?

Since disabled, the "strange" fuel cut didnt show up! I just dont know what should I look at next!!!!

So, at least we have some confidence the problem is related to that feature, and can keep focusing there. I'll stare at everything again later today, but don't really have any new ideas for now.

You may need to run a 0.1 uF capacitor from the input to ground to filter out noise.

It is possible to get ADC7 to 1023 and status unchanged while fuel cut happens?

By your experience, noise is coming from where?

It is likely that the reading changed for a brief instant that was faster than the log could capture.

Not sure where the noise would be from.

Matt Cramer wrote:You may need to run a 0.1 uF capacitor from the input to ground to filter out noise.

Would you please show me on my diagram where I should put the capacitor? Thank you.

JS4 to ground.