Greetings from Greece

[vosk]

Hello everyone,

I am currently messing with a French TU3JP engine and I realized that somehow this car has no cam sensor, at all. It has a skipped tooth flywheel, (not sure the ratio but it doesnt matter). I was thinking of building a rusefi in the long run, I am a software developer by profession, I have worked in various robotics platforms that require precise event timing, and I am have worked with AVR microcontrollers quite a lot. I was hopping to contribute.
So, my first question, as I get around to reading and compiling the code and the wiki and the forums. How on earth does the crappy Sagem ECU figure out the cam angle without a primary trigger?? I read somewhere that the software detects the primary coil current, and since this a waste spark coil pack engine, (I think it groups 1+2, 3+4) but then again how does it figure the injection cycle? I am really confused.

Thanks
v

[puff]

no exact answer, but probably by angle acceleration/deceleration?
let's suppose that skipped tooth on the crank allows to figure tdc for cylinders no1 and no3.
During starting, it might inject fuel simultaneously to all cylinders. Then, it switches to batched injection (1+3, then 2+4), and then, to define the cam angle, it doubles the fuel to cyl1 and gives no fuel on inj3. if the crank continues to rotate at the same speed, it detected cam phase correctly. if the crank deccelerates - then it missed the phase. could be a cheap alternative to installing a cam sensor.. %-)

[vosk]

What I can tell you that the engine starts fine with injector #1 or #2 disconnected (one at a time). It doesnt like it, but it runs. I had a clogged injector and I listened to the injectors by putting a screw driver on top of them and pressing my ear (you would be surprised how much you can hear). The injectors were not firing at 800rpm (my cold-ish idle). The were firing at 400rpm. So its not doing batch injection.

[AndreyB]

Saturn cars were somehow detecting engine cycle by reading ignition wire as well? I could be wrong, do not know much about it. Also ion sense from Saab - so, there are options but rusEfi has nothing to do with them

[puff]

They shoudn't be making 800rpm - normally they're injecting only once per two crank rotations - aren't they?

[kb1gtt]

Lots of possibilities, might just fudge it using wasted spark, and crank which has 1 and 3 both TDC at the same time, then it might not be sequential fuel, etc. or it might do some exotic advance / retarding to find which is cyl 1 and then switch to sequential once it knows what is going on.

Do you know if TDC for 1 and 3 are 180 from each other? I understand that for balancing purposes they are typically something like 175 degrees or something like that.

[stefanst]

You don't need cam-signal for wasted spark ignition w. batch injection.
One fuel squirt per engine revolution and one spark. The ECU doesn't know which cylinder is firing and doesn't care.

[vosk]

Hello everyone,

I pulled the fuel rail and cranked the engine, and I think its NOT doing batch injection, and I found this error code while surfing: "P0341 'phase detection integ ignition coherence'"
It really seems that the coil pack is detecting some difference between cylinders when firing the spark plugs pairs. The coil plug pack has a 4 wire plug, one is +12, other 2 are control signals to the pack, and one seems to be some kind of sensing signal.
One interesting thing, Peugeot has been installing some kind of capacitor-filter between one of the 4 wires and the engine ground to avoid ECU damage when the coil pack goes bad. Mine was not installed at the factory, it was done sometime later on, and its a bit sloppy wiring. I have seen this factory installed component on another car.
I am replacing all four injectors, waiting for them to arrive, I will try and record the fuel rail injecting with a good camera and see if its doing batch injection.

PS. In case anyone is wondering why I am replacing them: they are all from different cars, (the originals were toast), they from different factories and the coil reads different resistance between them: one is 11.9 Ohm, another is 12.7. They all should be around 12.5

[stefanst]

I am completely speculating here, so please take this with a few grains of salt:

It may be possible for the coil-driver to detect which spark plug is actually firing by measuring current in the individual spark wires. However, this would only be possible once the engine has started and we actually have a combustible mixture in a cylinder that is being ignited. So during cranking and especially with the fuel-rail removed and therefore no fuel getting into the cylinders, I don't see how we could get the information which cylinder is in need of injection other than from a cam sensor.

[stefanst]

Do you know if TDC for 1 and 3 are 180 from each other? I understand that for balancing purposes they are typically something like 175 degrees or something like that.

It's a standard 4 cyl engine. So 1&4 and 2&3 are on the same crank angle and the planes are exactly 180º off.

Looking at this again, I doubt there's a lot of sophistication going on. I'd still assume wasted spark and batch injection until there's some evidence to the contrary. In any case, if the goal is to make it run on rusefi, just run it batch-injection for now and deal with sophistication later.