Hi All,

Any help greatly appreciated as this is my daily driver.

56 plate – hawkeye - 2.5 wrx – wagon – EJ255 motor.

P0335 Crankshaft Position Sensor A Circuit.

Driving home on Monday night the car started spluttering whist going up a hill. The car stated losing revs and jerking forward and back – like kangaroo petrol. When I pressed down on the throttle pedal it didn’t respond. The car just hunted up and down the revs (roughly 400 – 1200 rpm), engine and gearbox started vibrating excessively, finally stalling. Jumped out had a quick visual around engine bay, seen nothing untoward, tried to start it – it struggled and finally started. Drove a little more trying to input some throttle to the engine, wouldn’t take it. Started spluttering, hunting up and down the revs, vibrating excessively, finally stalling again. This time it wouldn’t start again. Ended up on the back of breakdown truck and was recovered home . Prior to loading I asked the recovery driver to plug in his diagnostics scanner. Out he came with his snap on , plugged in to OBDE 2 port, P0335 Crankshaft Position Sensor A Circuit. When I arrived home I plugged in my scanner and also came up with P0335.

Was pretty sure this was the problem so didn’t even bother to investigate further, just ordered up patent CPS from Luke at import car parts, installed it, same problem. Checked the resistance in both sensors, both within range, OEM was 1.93 k ohms and patent was 1.96 k ohms. My understanding on this CPS: This is magnetic speed sensor which produces its own AC voltage when the metal tabs on the crankshaft pulley (the relluctor) pass the sensor face. This action interrupts the magnetic field inducing AC voltage. I set up a bench test to check if the sensors were producing AC voltage. I connected up my meter and passed a hammer (the relluctor in this case) past the face of the sensor. It produced a small AC voltage reading, I tested both, and both worked as they should. With the 2 pole connector disconnected from the sensor I did a continuity test to check if I had ground – I had a good ground.

Sensor pole 1: White is the live/signal wire. Connection position A13 on engine control module.

Sensor pole 2: White with red stripe is ground. Connection position A14 on ECM.

On the wiring diagram there is also a cylinder with broken lines drawn encapsulating both sensor wires described as SB (guess seat belt) with a connection position of A24 to ECM. Guessing this has to do with the seatbelt reminder alarm coming on after you reach a certain low speed.

Possible problem but don’t know .

With connector disconnected from sensor, ignition off, I have 0 voltage on both wires.

With connector disconnected from sensor, ignition on, I have 0 voltage on ground but I have 1.33 v on white wire live/signal side . Why do I have 1.33 v when I am guessing there should be 0 ? It is the sensor that creates the voltage through disturbance of the magnetic field by the crankshaft pulley tabs.

I also just happen to also have a classic uk turbo EJ205, which runs fine. For a comparison I disconnected the 2 pole connector form the CPS and checked for voltage on both wires with the ignition on – none present! This is why I suspect there is a problem with the wagon i.e. 1.33 v on connector whilst not even connected to sensor???? I am just not sure what to do next to troubleshoot this situation . Again any help/thoughts greatly appreciated as not having daily driver becoming a problem.



Cheers Paul .

 

SB == shielding braid

 

Have you actually checked the physical cam timing?

On other cars, I've had the belt was two teeth out and car will basically refuse to run or barely run and log cam/crank plausibility codes.

Few cases the woodruff key had sheared so pulley marks still looked aligned

Other diagnostic routes would be scoping the cam and crank sensors at the ECU end and making sure the waveforms are clean and time up ok.

 

Hi Don thanks for the info, very much appreciated Sir .

Excellent info, step 1 & 2 already completed, should get a chance tomorrow to check as per number 3. Will report back as soon as.

Shielding braid - seat belt - so much for my electrical abbreviations list. Cheers for going easy on me on that one Don. Was a lot of mileage in that one.

Cheers Paul .

 

Hi ALi,

Thank you for your excellent input Sir .

Changed the timing belt, water pump, idler pullies and hydraulic tensioner myself. After replacement I always give it a whirl round at the crank and re-check timing marks prior to starting - so all is good there, or was good there!

Also after deleting the CPS code I can get the car to start and run at 800 rpm for about 4-5 seconds before it starts hunting between 400-1800 rpm finally stalling after jerking violently for about 7-8 seconds. If the timing was out I suppose I wouldn't get the short period constant 800 rpm idle.

None the less worth a double check, engine and box jerk pretty violently before she cuts out Re-your possible woodruff key shear scenario , that is a nasty one! Thanks for sharing that one - a new one on me

My mates scooby was parked on a hill with hand brake on and left in gear. That caused his car to jump timing . Have heard of that scenario more than once too. Also hand brake on in gear strapped down on back of a tow truck too .

Wanted to get into it today as this is my daily driver, and it not running is starting to cause problems, unfortunately was waylaid with family matters so will get to this tomorrow.

Will start with step 3 as per the info Don kindly supplied in his previous post. The live signal wire on the CPS is reading 1.33 v when disconnected - ground is 0 V, i think this this could be the problem . If step 3 is good then need to think about scoping.

These magnetic sensors produce their own AC voltage, so to me it should read 0 V until the relluctor induces a current from the sensor. My 1999 classic turbo reads 0 v and it runs perfect - pity I need to fit new sills to get it through the mot.

Cheers Paul .

 

This is with a totally different ECU on a totally different car, but I did once have a Mercedes E320Cdi (Bosch EDC15 Ecu) that was hard to start or would cut out once started. Logging a crank sensor fault when doing so (typical two wire sensor, not excited by ECU etc). The only thing I could find wrong was 1volt on the signal wire when it should be 0v. After many hours swapping sensor and running new wires to no avail. I removed the signal wires at ECU connector and measured the voltage at the ECU. Turns out the 1volt was coming from the ECU itself. Sent it off for repair, they confirmed fault. Fixed it.

Not saying this is the case with you as JECs and Bosch are totally different, but then this is the only EDC15 I have ever come across with this fault (most european diesels of that era were EDC15 based) maybe moisture inside the ECU or duff cap leaking electrolyte or some other component fault could 'leak' voltage to the pin for the crank sensor, it wouldn't take much.

 

Thank you for the Info ALi - much appreciated.
Very interesting indeed. You have a vast array of knowledge, and I am very glad you answered my post. I hope to get to this later tonight after work. I will update you on my progress.

Cheers Paul .

 

Hi All – Update.

Disconnected connector B134 at the ECM. Tested both the live/signal wire and the ground wire from the CPS to their opposite end positions on the ECM connector being A13 and A14. The outcome was a good test on both. The live/signal wire and the ground showed full continuity with the resistance on both wires stabilizing at 0.6 ohms. Reconnected ECM and battery – stared vehicle – usual jerkiness then stalled – checked CPS signal/live wire again – still 1.33 V present.

Bear in mind with my somewhat limited knowledge I have in this specific area of electrics my immediate thoughts moving forward with this problem are as follows. Please comment or, advise if there are better options to diagnose this fault, thanks in advance.

1. Is there supposed to be voltage present there or not? My 1999 classic has no voltage and starts and runs prefect every time, obviously there is a 7 year age gap between these ECM’s on my classic turbo and Hawkeye– has changes been made to the ECM’s that reflect the newer module do have voltage present? I have read that some ECM’s produce a bias voltage to monitor for short circuiting. One way I suppose to confirm the situation pertaining to the voltage would be to kindly ask someone with a running ej255 Hawkeye wagon to double check their signal wire on the CPS with the ignition on.

2. Shield braid and has this something to do with it. Came across this info when googling: Unlike unshielded cables, shielded cables are used to protect power and data paths from degradation caused by EMI exposure during power and data transfer. Electrical noise and EMI affects electrical circuits via electrostatic coupling, conduction, or electromagnetic induction. That disturbance interferes with the performance of circuits, and in some cases can prevent its functioning properly or at all. Wherever a data path exists, these effects can vary from increased error rates to a complete loss of data. Then you also have the interference from RF sources to consider. This CPS uses electromagnetic induction to operate?? This sensor has shielding braid installed for this very reason. Could the shielding braid be compromised?? Any one came across problems with the shield braid integrity??

3.ALi: To check if the ECM is producing the voltage would I proceed as follows. Remove ECM connector B134, leaving the other connectors attached to the ECM, reconnect my battery and switch ignition on. Then test pin A13 on the back of the ECM for voltage. I am guessing that this test would determine whether the voltage was coming directly from the ECM or, from some other source further up the line on the wiring between the ECM and the CPS. What you think, does this sound plausible? Worth a try?

Cheers Paul .

 

 

Yes, you can try that, but it depends what else is on B134 (supply, ground etc.). What I usually do though is disassemble the connector and unclip and remove the wire and pin from terminal A13, then backprobe through where the wire once was. Not the easyiest; Need a few lock-pick/connector disassembly tools.

Easier way is just cut the wire to A13 test it to see where the voltage is coming from then join it back up afterwards.

I am no expert on JECs and Denso (usually work with Bosch and Siemens), but the basic technology is similar. There will be a very small test current to check circuit integrity but unless I'm mistaken it should be millivolts not volts; it's the same way a multimeter measures resistance.

I'd expect a two wire sensor to be a traditional passive sensor and will produce a sine wave on a scope. A ECU excited/powered sensor will provide a square wave; they can be two wire but I've only ever noticed this on ABS sensors not on cranks. Either way if it is a powered sensor it will have 5volt reference (maybe 12v on other stuff) from the ECU, not 1.6v.

 

Thank you very much for your time and all the information you have supplied ALi, much appreciated.

Apologies for not getting back to you sooner, been a traumatic time for myself and the good lady as we had to put down our dog of 11 ˝ years yesterday.

After reading your post I proceeded as per number 3. on my last post. Removed ECM connector B134 and left the other 3 connected. With the ignition on I checked for voltage on the ECM at pin A13 which showed 1.25 v present. This obviously 100% proves that the voltage I was reading at the CPS connector is coming directly from the ECM.

In summary: This completely removes the theory of EMC or RF interference compromising the shielding braid, in turn, interfering with the circuit.

I have resistance/continuity tested the integrity of the wiring (both poles) from the CPS to the ECM. The wiring is good.

I tested the CPS connector to ensure I had a good ground – successful test, ground present.

I have bench tested two (an OEM one and new Patent one from ICP) CPS sensors for resistance – good test and, passed a relluctor (hammer in this case) across the sensor faces – good test as they both produced an AC voltage. This proves both CPS sensors are 100%. I have also tried both of these sensors in the block to no avail.

It is increasingly looking like the only viable answer is that the ECM had developed a fault on the CPS circuit. Especially when my 1999 classic turbo runs and starts every time on the button, and there is 0 v on the live/signal wire with the ignition on. I think the only thing left to check, and the sensible thing to do before I start looking at sending my ECM away for repair or, looking to purchase a replacement second hand one. Is to double check that there is no difference/advancements on the CPS circuits pertaining to the ECM’s on my classic (no voltage) and my hawkeye (voltage). That would remove the theory of some ECM’s produce a bias voltage to continually monitor the circuit for shorts. If a hawkeye that runs as it should has no voltage with the ignition on then that is all she wrote. I don’t know anyone local with an EJ255 hawkeye so I was thinking about a new post kindly asking for someone who has a running hawkeye to pop off their CPS connector and check the live/signal wire for voltage with their ignition on. What do you think? Does that sound like the most logical course of action to troubleshoot this problem further or, please advise if you think there could possibly be something else at fault here, something else that needs to be checked ? Much appreciated.

Cheers Paul.

 

It's looking towards a ECU issue.

First I would double check all positive and earths that supply the ECU. Temporarily running new ones from the battery if needed (connector back probes are a godsend for this). As a low voltage or floating earth can throw up issues. If you can get live data at ECU from the OBD showing voltage, that helps.

Also one headache is a short on wire for another sensor or actuator can throw off something else and not necessarily log a relevant fault code; I once had shorted throttle pedal wiring throw off a fuel pressure sensor (that blew my mind). Doubt this is the case here, but can't rule it out. ECU swapping between cars is the way to see if the fault moves with ECU onto the doner or vice versa.

 

Sorry, forgot to ask: You said you did the timing belt in the past. But I think I have missed you saying you re-checking it, in particular the woodruff key in the crank sprocket.

At this stage it's worth double checking before forking out on other ECUs/testing/etc.

 

For sure ALi. Thank you for your last 2 posts - much appreciated. No, you are right, I haven't double checked the key yet but I am doing it tomorrow as could be a strong possibility. If the key is broken obviously the runout deflection would play havoc with the gap at the tip of the CPS sensor.
Hoping that this is the problem. Fingers crossed in all that. Will update you with my findings.

Cheers Paul.

 

Hi ALi,

Hope all is well with you.

Finally managed to get time to check the timing out. Got into this tonight. The woodruff key between the crankshaft and the crankshaft pulley is perfect. Also all the timing marks are in the correct position, everything is **** on.

Thinking out loud here. In summary, bearing in mind all the previous checks I have carried out (refer to my post on 26.08.21 @ 7.22 pm) and now that I have confirmed all is well with the timing, and the key is good. Do you think I would be correct to assume that the problem is exactly this: With the ignition on, and without the engine turning over, there is a steady 1.25 volts present on the actual ECM at pin A13 – the crankshaft live/signal pin? Note I am not even trying to second guess what is causing this to happen at this stage, basically I am just trying to establish what is this the initial problem? Sorry for my layman’s understanding in advance, but I need to firstly establish is this the likely culprit or, not as the case maybe?

MY OBDE scanner is telling me it is a CPS problem. The sensor works 100% as it should. My old classic UK turbo has no voltage on this same pin with the ignition on, it fires up every time on the button. The following makes sense to me as (don’t know if I’m right – like I say just thinking out loud here) this is an electromagnetic speed sensor that should only create a very small AC voltage once the crankshaft pulley (the relluctor) pick-ups start turning past it, in turn inducing a very small AC current. This should only happen when it’s running i.e. the pulley is moving. There should be no voltage present on that pin when the car isn’t running. I am thinking that the 1.23 volts is drowning out the very small amount of AC voltage produced by the sensor when the car initially starts. The ECM cannot interpret the data so cannot supply the required timing data for the car to run - hence the erratic vibration and no run after the initial 3-4 seconds. Also it will only start for that initial time after an ECM reset. I think it is impossible for the ECM to read the data from the CPS sensor.

You obviously have a far greater knowledge than me on this, am I correct or, am I away on the completely wrong tangent here?

You thoughts would be very much appreciated.

 

Swapping over the ECU would be my first point of call. Must be someone local with a similar car?

 

Thank you for your advice Technopug, Much appreciated. I'm afraid that is the problem with living out in the sticks, no local car.

If i was to get another ECM would I be able to get past the immobilizer and turn the ignition on? Would i not need to firstly have my original keys coded/paired to the new ECM before i could get ignition to firstly check the situation with regards to voltage, and then try for a start and run time of more than 3 to 4 seconds.

I was thinking about putting up a post kindly asking for anyone with a EJ255 WRX Hawkeye if they would quickly slip of their CPS connector and check if they had voltage or, not on the live/signal wire with their ignition on. That way i would be able to confirm that this is my problem. Then i can move on with trying to fix it.

 

I was forgetting about the immobiliser issue.

 

Yeah put up a post in a general.

For another ECU to work on that car I think it'll also need the matching immobiliser ecu and key from the same doner car. Or a modified/aftermarket ECU without a immobiliser.

Asking someone to check what volts they they have at their CPS connector would be great if you can find someone.

Only other red herring I can think of at this stage is Cam sensor issues (should be a P0340, but you never know, an erroneous cam signal could confuse the ECU and make if think the crank signal is wrong). Scoping cam and crank outputs is what I'd do here. I can't remember if these will run with the cam sensor(s) disconnected, it's only there to identify no.1 cylinder (for sequential injection )and run the AVCS so most cars will operate a limp home strategy with a disconnected cam sensor. Someone else will have to answer that.

Other red herring; Bad earths. Running a few jump leads from the negative terminal to the engine block and chassis. And electrical interference (usually alternator), again that's found by using a scope.

And finally just to ruin your day the P0335 code is totally irrelevant and just logging because you have a slow cranking starter and what is causing it to cut out and stall is something else entirely. Doubt it, but have to put that out there.

 

Thank you for your comments Ali, in this, and the previous posts – as always very much appreciated.

Moving forward may I clarify a few points with you, just want to make sure I am on the right track. Apologies in advance as we are now going into new territory for me. This is a more advanced level of electrical troubleshooting that my current knowledge base. I have a can do attitude, and am willing and love to learn, not much option out here in the sticks. Not much in the way of Auto electricians going about.



1. Asking someone to check what volts they have at their CPS connector would be great if you can find someone. Yeah put up a post in a general.

When you say put up in a post in General do you mean General Technical, the same place I posted this one or, Scoobynet General? Just want to make sure I post in best place to hopefully receive some help.

2. For another ECU to work on that car I think it'll also need the matching immobiliser ecu and key from the same doner car. Or a modified/aftermarket ECU without an immobiliser.

Will leave this route at the moment until I have exhausted all other possibilities.

3. Only other red herring I can think of at this stage is Cam sensor issues (should be a P0340, but you never know, an erroneous cam signal could confuse the ECU and make if think the crank signal is wrong). Scoping cam and crank outputs is what I'd do here.

Will scoping work if the car will not run? It will run for 3 to 4 seconds after reboot only. Then it won’t run full stop unless I reboot the ECM again.

I am thinking it may be worth checking the camshaft sensor i.e. proper ground at connector, resistance, continuity of the wiring from the sensor to the ECM, producing AC voltage ect, basically the same checks i carried out on the CPS? Do you think this would be worthwhile prior to scoping? Where is the cam position sensor?I see nothing at cylinder 1 and 2? Must be at the back on 3 or4?



4. Other red herring; Bad earths. Running a few jump leads from the negative terminal to the engine block and chassis.

Can you recommend a kit that can be purchased online for back probing, running jump leads, opening up block connectors ect.

What would I actually do here? This is new for me.

What connector’s negatives would I run the new grounds to? I am guessing all the engine sensors i.e. cams, throttle position sensors etc. I am also guessing that after running the new earths one at a time, I would then either check if my car is running properly or, to see if there is still voltage present at the ECM on pin A13. That way I would know that a bad earth on that specific sensor was causing a problem with my CPS system. Am I on the right track here or, the completely wrong one? Apologies again for my ignorance in this area.

 

1. Seeing the forum isn't that busy these days both Scoobynet General and General tech.

3. Scoping is better with engine running at constant speed. Be it on the starter or idle, otherwise you have to keep messing with the time base.

Checks are same as crank; Correct number of pulses per revelation and a clean pulse without noise or random spikes.

Sensor should be just to the front left of the oil filler neck.

If it has AVCS, it'll have another one on the opposite side under the air intake hose for the air filter.

4. I just use battery jump leads from battery to good earth points on engine and chassis.

The only other thing I'd pay attention to is Earths on ECU itself using multimeter/banana plug leads, croc clips to earth points and back probe pins (something like this that's just a quick google, may be able to get a better/cheaper kit elsewhere ). As it's these that grounds the sensors snd actuators. Need ecu pinouts diagram to know which pins on which connectors are earths.

 

Attached

 

Thank you ALi.

 

Thank you for the information again Don.

 

Hi guys, did anyone find a solution to this? Was it the ECU at fault? My car has the same/very similar performance issues and same error code but replacing the CPS did not help and my local subaru guy is out of ideas. Thank you!

1999 classic STi, EJ207