Adamw

This doc here gives some general info about connecting and setting up the basic lamp signals: https://www.aim-sportline.com/download/faqs/eng/hardware/sensors/mx_series/FAQ_Sensors_MXx12_StradaSpieIcone_100_eng.pdf The lamp signals connected to dash inputs dont need a dedicated ground, they will work via chassis ground. Fuel level sensor would benefit from a dedicated ground connected to the dash analog ground provided the sensor does actually have a ground wire - many dont.

Yeah your cam timing is wrong. I dont remember if it is physically possible, but I would say since one of your earlier pics shows the exhaust cam dowel is correctly pointing vertical, then they are possibly swapped - ie the intake cam is fitted on the exhaust side. At TDC, your number 1 Intake lobes should be pointing back towards the firewall, say about 10'oclock if looking from the cam sprocket, and the exhaust lobe should be pointing forwards, say about 2'oclock. Pic below is looking from the opposite end but my pink lines indicate roughly where the lobes should be.

I wanted a pic with the cam cover off (not the belt cover) so I can see the cam lobes...

The Evo 6 has a 2 speed fuel pump control. The default fuel pump settings are shown in the screenshot below. In your diagram the relay on the left is the main power supply relay, this turns the pump on or off, it is controlled by Inj 5 output. The relay No2 on the right side of your diagram switches the pump between low speed and high speed mode. This relay is a "normally closed" type, when it is not energised the contacts are closed so the current from the first relay flows directly to the fuel pump (full battery voltage). For low speed mode the ecu energises Relay No2 so its contacts open, the current then has to travel through the resistor to reach the fuel pump - the voltage that reaches the pump will be about 9V instead of about 14V. Typically low speed mode is just used to lower pump noise at idle. With the settings below the fuel pump is in lowspeed mode when RPM is below 2000 and injector duty cycle is below 5%. When RPM OR duty cycle is exceeded then the pump switches to high speed mode.

I dont really see how it matters where the starter relay gets its power from in relation to the PDM if it is not controlled by the PDM.

No, knock level wouldnt usually change noticably with fuel, most influence comes from cyl pressure - so MAP and ign advance. Usually you would have less advance with less ethanol so I would expect knock level to drop with petrol provided boost was similar. I notice your threshold table uses TP on the axis so that could potentially be some of the issue if it is reaching a higher boost for a given TP (you are only at 30%TP in the log).

Can you give me a photo of the cams with the cam cover off with the engine at TDC.

Yes, as well as A/C, Fans, wheel speeds, and the imobaliser which all work over CAN

There are 4 different tests that the fault status relates to, but usually if it is showing a fault when the aux is inactive it means the voltage on the aux pin is less than 2/3 of battery voltage - ie there is some resistance to ground in the circuit attached.

One pin connected to sensor ground, the other to the AN Volt with pull-up.

Yes we will have some DI's with adjustable thresholds in a future ECU, but all our current ECU's have the thresholds set by hardware.

There wont be any manuals, but you could start with what specific model Forester you have, and what specific Link ecu you have.

If you have a spare analog input you could just add your own pull-up to make it work as a temp input. But to answer your question, yes the 30-2226 should work for what you want.

Can you tell me the firmware version that is listed in >help >ecu information when you are connected. If it doesnt work in test mode it would suggest there is a wiring issue. If you unplug the ecu and switch on ignition, measure voltage on the aux 4 pin to confirm you have 12V there, provided you do, then poke a paperclip or something in and short it to ground to confirm the pump relay then clicks in.

The log and scope look fine. If your injectors dont click in test mode then you have a wiring issue or injector issue so you need to diagnose that. They should click in test mode regardless of ecu settings.

Yeah I dont think that will work well without the main TPS assigned to the sensor that is on the ITB's. Accel enrichment, lockouts and delta's for many functions like knock control etc all require TPS. There are plans to add a generic H-bridge control function to the ecu and I have just started on documenting it so you could consider it in-progress now and I think it shouldn't be a major job to add from what I have written so far. I dont want to make any promises on behalf of our firmware team but my initial feeling based on similar jobs and the workload I know of at present, it might be available in a month or so. Otherwise you could probably find a robotics/hobby motor controller that will take a PWM aux and convert it into a CL motor position control.

It looks like it should work in that case provided that wire is not spliced to anything else. You also have AN temp 3 on C12 so you may just be able to move the pin from C31 to C12 to make temp 3 available at the maf plug.

Does the fuel pump run if you set the aux to test mode? What ecu and what firmware?

No, the trigger looks fine in the most recent log/scope. This is normal. The RPM value that is shown in a log when the engine is running normally is updated every crank tooth, but averaged over 1 TDC event (This is user adjustable but 1 is default). When the engine first starts cranking however (before it has seen the cam signal to sync), it doesnt know where TDC is so the RPM is not filtered/averaged. The RPM for up to the first 2 crank revolutions will be more erratic as it is updated every tooth with no filtering. Everything looks fine from the ecu perspective in the log, it has all the info it needs for fuel and spark. There is no injector PW, but I assume you had fuel injection mode set to off when you done that log? If it's backfiring then either a spark is occurring at the wrong time or a valve is open at the wrong time. Since you say only the trigger has changed since it was running I guess we can assume the valve timing is ok for now. You say you have a OEM beams trigger wheel and sensor, so your offset of -35 or 335 doesnt seem right - assuming the sensor is mounted in the stock location. The stock offset on these are normally around 145, but in this case since you have a non OEM cam trigger that could be 360 different - so I would expect the offset to be either 145 or -215. Oddly that is exactly 180deg difference from what you have set, so that may indicate you possibly have a coil mis-wired or something like that.

Allocate the cylinders to a knock table and it will work.

I only have a very low res pic of the wiring diagram for the 04 pinout, but I think you will find the maf ground wire doesnt actually go direct to the ecu on its own, it will go to a slice somewhere, im not sure if this is near the ecu or somewhere else in the loom.

Try the values on the help file page that I said to try.

Go into the GP input folder.

Have you already used AN Volt 3 and AN Volt 6 on the main header?

It will be fine, the CANTEE is designed for exactly what you are doing. You are correct in that the main trunk should be twisted pair, but the stubs dont need to be. The "maximum stub length" of 200mm which is in some of our documentation, im not sure where that originally came from but I've seen a few variations of that "rule". For example the common CAN standard ISO11898 quotes 0.3M max stub for a 1Mbit bus. Some standards instead use a calculation based on propagation delay and when you work through the calculation they often end up giving max stub length results of >1.5M for a 1Mbit bus, so I suspect in many cases it is not all that critical. The reason for the stub length restriction is stubs are not terminated so you can get signal reflections that will bounce the voltage above/below the thresholds causing "bits" to be seen that arent actually there. However, CAN has a very robust error checking and validation routine so these errors will not make any difference to the received data - just some devices will report the error on the bus so all devices will reject the message with the error and make the sender resend it. I quite often end up with stubs longer than 500mm for things like curly cords to steering wheel controls etc and have never yet seen any obvious negative effect.