Adamw

Moved to the G4+ forum. You dont need to set any source in G4+, it should just work as soon as the CAN is assigned. Just ensure no analog input is assigned to lambda 1 function. To confirm its status, go to the runtimes screen (F12), ECU status tab, and you will find lambda 1 status in there.

Its not going to make a lot of difference either way, but the way you have it is a decent option. You will have cyl 1,2,5&6 all squirt at the same time, then 3&4 will squirt together either 180 or 360 deg later depending on injection rate setting. So on the first squirt you have 2 injectors on each side firing together, on the 2nd squirt you have 1 injector on each side firing. The main thing I like to avoid is having all injectors on the same side squirting together as this can cause large fuel pressure oscillation - so your set up is good from that point of view. Use an injection rate of 1/2 cycle provided your injectors arent bigger than say about 600cc.

Adding fuel is typically not very effective knock supression on 4 stroke piston engines - unless IAT is very high and you are using a high HOV fuel like methanol or ethanol. Pull 5 deg timing out to confirm if it is knock. Yes, cyl 4 is usually the first to knock on the subarus.

Yes the V88 was released in 2009 and discontinued in 2013, a time when very few ECU's had enough memory for a scope function. For the cam sensor, if you "back probe" the trigger 2 pin at the ecu connector by poking a paperclip or similar down beside the wire so it touches the terminal, connect a voltmeter to that then watch voltage as you put a piece of iron (wrench or similar) in front of the sensor like your pic shows. If sensor is working and wiring is ok then you should see high voltage (>4V) when there is nothing in front of the sensor, and close to 0V with iron in front of the sensor.

At the ECU end set up as per below. "Transmit Generic Dash", set the CAN ID to 1000 and CAN ID format to "normal". At the ADU end go to >project tree>add>import CANX/DBC file>LinkECU_GenericDash

For the Knock sensor, wire it to the same pins the factory sensor used. Knock ground and shield both to B136-29, knock signal to B136-26

Yep, looks good so far.

In multipoint group mode the injector drives are fired in 2 groups, all even drives fire together and all odd drives fire together. Whats the injector resistance? If low impedance, does it have a ballast resistor?

The odo and speedo arent driven by the ecu at all, the gearbox speed sensor connects directly to the speedo.

Density doesnt look right, the other 2 are in the ball park. Density for petrol is generally around 0.745, pure ethanol is around 0.789, so you should be 70% of the way between those two. So about 0.776 by my count.

This would suggest it is mis-syncing - the cam sensor isnt working or isnt being seen so it will sync on the correct crank rotation approx 50% of start attempts and on the wrong phase the other 50% of attempts (roughly). You really need to get an oscilloscope on the crank and cam sensor to see what is going on.

No, it hasnt been implemented in G4+ firmware.

Hmm, thats a very odd one. Im suspicious there maybe a hardware issue here. Can you try removing the ecu from the car and powering up on the bench, then trying to connect - just to eliminate power supply issues. You can power the ecu up on the bench by clipping power on to the big TVS diode like so:

The switching thresholds on the DI's are fixed by hardware, not user adjustable. From memory it needs to rise above about 1.8V to be considered active, and fall below about 1.0V to be considered inactive. A pull-down resistor connected to ground might be enough to solve it. Another option you could possibly use an analog input (can measure up to 5.0V), use this voltage as a condition in a GP output/virtual aux, which you can then assign as your Brake switch source.

I would try to trace the original wiring first rather than just replacing with new ground wires. The reason for this suggestion is say for example the poor ground is due to something like a corroded ground strap on the engine, then just replacing the wires to the ecu isnt going to solve the problem - and the bad engine ground is likely going to other problems. So in my opinion it would be best to try to determine where the issue is first, if that becomes difficult to diagnose only then try running new grounds.

0-5V = 0-1250C

Sorry, I kept forgetting to come back to this. Im going to have to talk to some of the smarter engineers to see if they can suggest something. Its got a very odd circuit on this input and I dont really understand how it is meant to work.

I dont really want to even guess with the difficulties in the electronics world at the moment. My feeling is less than a year but thats about as specific as I could be.

Ok, it sounds like you may have a bad connection somewhere. A scope would certainly make your life easier.

This is not a subject I that I have ever researched much so I dont have much to offer. Only that I have seen cars with a sensor or switch on both the low and high side - or only on the high side. But I have never come across one with only a sensor on the low side.

So it appears you have a broken/disconnected or high resistance main ground. Pin 1 should be very near 0V at all times.

Yes, it would need 720 sync for semi sequential.

16/33/34 are nearly always related to power supply. Your battery voltage is very erratic - I think that is the first thing I would look at trying to improve.

If you are sure you are getting no pulse from pin 110 when injector 5 is in test mode then contact tech support to get the ecu inspected. Make sure you are using an LED test light or similar so it is fast enough to see the very short pulse.

Set to semi-sequential, Inj drive 1 to Cyl 1&7 Inj 2 to Cyl 2&8 Inj 3 to Cyl 3&5 Inj 4 to Cyl 4&6 It will then effectively be sequential/timed injection, 1 cyl of the pairs will get its fuel 90deg earlier than ideal, the other cyl will get it fuel 90deg later than ideal.