Data Logs:
oh1.zip
Tune:
CurrentTune.zip

History: Car ran great, spent some time tuning *clearly not done yet on the ve table*, drove a couple hundred miles on the tune, all happy fun times.

Woke up one morning
Car cranked rough, flooded.
Deflooded the car, watched what happened at the end of the attatched data log
flickering of RPM signals on and off @cranking, sync losses 11 and 17
let it sit
hour later car cranks right to life.
drove it around town to be sure and then home from college (2 hours)
wrote it off as a fluke


next night car wouldn't start, ecu showed sync losses again
towed, v3.0 board dead
upgraded to v3.57, rewired whole car
new cas wires, new daughter board, whole 9 yards, started first time I turned the key car started, took above data log, turned it off, tried to start it again and sync losses reappeared. (felt like it was missing)

Shielded cas wires
3 12ga ground wires (read the megasquirt grounding pdf, multiple wires act as parallel resistors, figured 3 would be plenty)
unplugged the coils, same results

where should I start looking?
solders on the cas wires are good.
checking the daughter card solders, but I solder all day for the engineering lab I work for.
and I am going to adjust the VR circuit pots as well

I have another cas coming my way since that is the only thing that hasn't been replaced and the issues persist with a different megasquirt.

help me get my FC running again.

 

How do your grounds run? Where is the MS grounded?

How is your 2nd trigger circuit set up?

What condition is your starter in? (May not seem relevant, but I've seen it cause problems).

Where and how are the CAS wires' shielding hooked up?

Ken

 

everything Megasquirt related:
Wideband
Megasquirt
Cas shield
prosport gauges and boost sensor
main & fuel pump relay trigger grounds
(sensors are grounded to one of the ground pins on the db37, this is the only change other than the extra 12ga ground wires from the prior harness)

all go to a grounding block/pad at the fuse panel
then from there, there are the 3 12ga to the engine block



my 2nd trigger setup is the zeal daughter board VR2 circuit
into JS10, using SPR3 and SPR4 on the DB37 to bring in the cas connection to the daughter board
pulling 5v from the proto, SG is the ground.


Two cas wires, I used the sheilded pair wires from DIYautotune,
connected ONLY on the megasquirt end, braids were soldered together and then an 18ga wire takes it to the gounding block.



Will test the starter idea by pulling the cas out and spinning the gear by hand without the starter. taking it out of the equation.

 

Typically I only ground the MS and related stuff to the engine, and nowhere else... then the engine to the chassis and the battery to the chassis (and starter).

Ken

 

update:

Took the starter out of the picture and spun the cas by hand, no dice.
also tested the Cas plug at G1/G2 and Ne1/Ne2 as per the FSM
166ohm and 168ohm respectively
within the 110-220ohm the FSM calls for

double checking my solders and then I'm going to take both R56 and R52 from fully counter-clockwise to see if the trigger voltage or (mainly) the hysteresis will change anything

with a multi meter, at Vr2out on the daughter board, and at cpu pin 17 and resistance was present. so resoldered the lead at JS10 and it is gone. fingers are crossed.


I read somewhere that since these surface mount Potentiometers don't "click" you just need to turn six times counterclockwise so I'm going to do that again aswell.

:edit:
adjusting The potentiometers and correcting the solder didn't help

 

As far as my grounds go, they are doing the same thing as in Aaroncake's writeup.
http://www.aaroncake.net/rx-7/megasq...nel_in_car.jpg
Basically collecting at the panel (bottom right in picture) and then finally grounding via the larger cables to the engine.
In my setup that panel is not grounded to anything but the engine block.


I have the engine grounded to the chassis via a pair of 4ga wires, battery and starter using the factory stuff (albeit in good condition)

 

If you're not getting RPM spinning by hand, I'd do the following:

1) Check continuity from the CAS back to the MS for each wire
2) Check continuity from the pins on the back of the MS to the places they are supposed to go internally (inside the MS).
3) Verify function of the primary tach input circuit using a stim or by putting the MS in "distributor" mode, and spinning the CAS (disconnect injectors and coils of course)

It sounds to me like there is an intermittent connection somewhere.

Ken

 

inspect the grommet in the CAS where the wiring pokes through
,, i expect the wires here are frayed and shorting each other or to the CAS body

very common to see the CAS like this ,,requires you to splice around the issue

 

Pulled the cas, checked the wires in the grommet, they were stiff but not shorting.
double checked my cas wires, making sure I saw the same resistance at the cas plug, cas wires at the DB37, and at SPR3 SPR4 and pin 24 and 2.

finally, thought I would update my code from 2.1.0d to 3.1.0

spin the cas, and now I have a smooth rpm signal.

What could have caused the prior 2.1 code to not work right randomly?

 

drove it around the block, smooth rpm for a minute and then I started loosing sync again.

 

Is this happening when you're hooked up to the MS and tuning it with tunerstudio?

Have you tried without connecting tunerstudio?

Ken

 

Swapped out my cas with another one, and my problems seem to have been solved.

Although my old cas tested out with the multi-meter as per the FSM, turning the gear at the bottom felt smooth, the indents almost totally un-pronounced. As you spin gear on the one that is in the car now, the indents feel significant

Drives fine so far, drove for a few hours and saw only a handful of sync losses ~20, maybe messing with the potentiometers will clean all that up.