[Eth207]

If it’s a good price, it’s money in the parts bank… Easy to cash out if not required.

That’s the only way parts swapping is viable, if cheap parts are on hand and the fault mode won’t blow up swapped parts

Yes, I did mean to wire up both ocv’s at the same time, could get fancy and extend the wires back to the cabin so you can test it solo?

Okay Good point. I was setting the car up to log and I noticed something that may leave us with a conclusion, I will try my best to word this correctly.

Both OCV connectors were conected, I put pins in each wire in the connector and attached Alligator clips to each side. (R/L red and black matching on each side, making sure to keep them from contacting eachother)
Unaware of the correct polarity, I start on the left OCV and attach the 2 wires to 9v, I hear a loud click (OCV Actuated).
I then swap the wires around to reverse polarity, I hear the same loud click.
Then I grab the right side OCV wires, I put them on the battery, I hear a weak click (Initially this made me believe the OCV was bad)
I swap the wires around, nothing actuates.

To summarize this part of the test:
The Left OCV strongly actuates regardless of polarity and the right only actuates one way, and it is weaker sounding.

I then bridge both wires together and put them to the battery, nothing actuates.
I switch them around, nothing actuates.

The OCV's would not actuate in any way, when bridged.

Then I unplug the OCV wire connector from the ECU (the circuits are now only connected to my 9v). I do the same tests...
I start with the left side: just like before, the OCV strongly actuates for both polarities.
I then do the right side: to my surprise, both polarities result in a strong actuation.
I then bridge the wires: Both OCV's actuate at the same time for both polarities.

This, as well as my wiring test, and the fact that one of the connectors isn't getting a good ground seem to point at a bad ECU.
....
here's the weird part. Im still only getting a p1309, Left side fault. these tests seem to show a fault in the right side...

Also if you remember, when I swapped OCV's I started seeing the Left side jumping angle 0-1 degrees and the right side remained at 0.... this was the oposite before I swapped them over


Now whats even weirder, I was logging a cruise log, while accelerating, My left side OCV went to 80 something degrees. I have never seen numbers this high, very weird.


What do you think about this

 

[bigBADbenny]

Might be a good idea to graph the log to get that maximum advance angle? Primarily so I can google it.
Eg is this something anyone else has experienced? Is this the cam pos fault we want to see?

Apparently vvt angle can be affected by oil pressure alone?

As an aside and in reference to engines with exhaust AVCS on non exhaust avcs ecu’s someone proposed an external avcs controller (edit this later)

Did each OCV see 9v or 4.5v? Eg was it wired in series or parallel?

I’d have to see the log for the circumstances, and was the ecu reset prior? Eg is that an attempted recalibration?

You’re brave doing the test piggy backed on the ecu, if you compare the results to the wiring diagrams, the idea the ecu tune calibrates and controls two seperate channels, you may gain insight.

Again with parts swapping OCV’s, if you can find a known good item to swap in, it could be informative. A local enthusiast tore down his failing OCV’s, the fault mode being galling wear on the actuator shaft. As far as I know, the solenoid is similar to a linear motor, which might explain that they work with reversed polarity.

It would be interesting to test the OCV’s using power to either pin and the OCV shell as ground, with a multimeter on volts and or resistance inline: is the shell isolated or not?

I guess you could extend your previous test by bridging one channel of both ocv’s then the other.

Hopefully you’re logging all the tests?
That’s going to be really interesting to see.
You can also share via google drive, or Dropbox, eg try drop the .csv in this folder: Eth207 my02 04 jdm swapped hawaii
Uploading to data zap is really easy after iirc registering, then anyone can graph on the site or DL the files.

 

[Eth207]

Might be a good idea to graph the log to get that maximum advance angle? Primarily so I can google it.
Eg is this something anyone else has experienced? Is this the cam pos fault we want to see?

Apparently vvt angle can be affected by oil pressure alone?

As an aside and in reference to engines with exhaust AVCS on non exhaust avcs ecu’s someone proposed an external avcs controller (edit this later)

Did each OCV see 9v or 4.5v? Eg was it wired in series or parallel?

I’d have to see the log for the circumstances, and was the ecu reset prior? Eg is that an attempted recalibration?

You’re brave doing the test piggy backed on the ecu, if you compare the results to the wiring diagrams, the idea the ecu tune calibrates and controls two seperate channels, you may gain insight.

Again with parts swapping OCV’s, if you can find a known good item to swap in, it could be informative. A local enthusiast tore down his failing OCV’s, the fault mode being galling wear on the actuator shaft. As far as I know, the solenoid is similar to a linear motor, which might explain that they work with reversed polarity.

It would be interesting to test the OCV’s using power to either pin and the OCV shell as ground, with a multimeter on volts and or resistance inline: is the shell isolated or not?

I guess you could extend your previous test by bridging one channel of both ocv’s then the other.

Hopefully you’re logging all the tests?
That’s going to be really interesting to see.
You can also share via google drive, or Dropbox, eg try drop the .csv in this folder: Eth207 my02 04 jdm swapped hawaii
Uploading to data zap is really easy after iirc registering, then anyone can graph on the site or DL the files.

Okay Ill try to log more tonight, I put my cruise log with the 80 degree left side VVT thing in that dropbox, im not sure if it worked or not, and I wasn't sure how to find "closed loop" so I just selected CL/OL fueling and closed loop fueling. I also couldn't find the cam or crank sensor stuff I hope its still valid. Also youll notice my gear parameter reads gear 6 when in neutral. I am pretty sure this is normal because I have seen other guys on the forums relate to that.


Each OCV wire was placed on he same 9V battery, I guess you'd call it parallel?

What are your thoughts on my tests? doesn't this kind of prove that the ECU is at fault? the OCV wouldn't actuate as good or with both polarities when plugged into the ECU

 

[bigBADbenny]

It’s hard for me to make that call, I’m not a mechatronics expert.

I’m not using dropbox pro, so perhaps you could try the link below to upload, or share logs from your own cloud service or “datazap dot me”?

When you get to it, share all relevant logs, perhaps we’re missing something basic we’ll catch in the data.

Dropbox

www.dropbox.com

 

[Eth207]

It’s hard for me to make that call, I’m not a mechatronics expert.

I’m not using dropbox pro, so perhaps you could try the link below to upload, or share logs from your own cloud service or “datazap dot me”?

When you get to it, share all relevant logs, perhaps we’re missing something basic we’ll catch in the data.

Dropbox

www.dropbox.com

I wasn't able to get the drop box to work so I uploaded them to data zap. I uploaded 2 sets of logs, cruise and idle. there are 3 cruise logs and 2 idle logs, the file name will describe what is happening in the log ex: OCV's connected vs disconnected.

Idle Logs

Cruise Logs

Hopefully those links will work

For whatever reason, the AVCS on the right side decided to be fully operational during my test drive, for the first time all 3 parameters were moving around the way they should (I think)

One notable point, when I was applying 9v to the OVC's during Idle, they individually actuated good, but when applying power to both, 9v would not do it. So I just went to the 12v battery and then they were able to both actuate at the same time, interesting.....

 

[bigBADbenny]

On the latter note, the OCV’s may have been getting 4v each from the 9v battery which is probably at the lower end of the actuation threshold?
Ok, in the logs we’re looking for out of normal range data, and data anomalies.
So I can see your long term fuel trim, af learning, is heading out of acceptable range negative.
But that’s a state of tune, usually a vacuum leak, issue, it’s not appearing to affect knock learning, at least off boost during cruise.
One anomaly I see is the rpm signal diving to zero after a quick peak, I see that as an anomaly: further investigation required, eg start by checking the integrity of the crank pos sensor plug: visual inspection, back probe the pins to the harness wires immediately after the plug and or at the ecu, use some dielectric grease in/on the pins, make sure the pins are clean on the sensor and the plug sockets aren’t over rounded, look for corrosion etc.

Then I’d repeat the cruise log with eg a 9v battery on each OCV, we want to check the cam sensors vs crank sensor, rpm & vvt angles whilst driving around, with the OCV’s pinned to max angle.
Generally vvt angle is high during cruise, so hopefully the car can get moving in that condition.
If not do the test warmed up in neutral using the vary rpm test method I already mentioned. We want to rule out issues with the crank & cam sensors before committing to say a new ecu.

You might address the crank sensor plug checks prior to these tests.

I’ll have access to my laptop next week so I can graph the logs with MLVHD, so no rush, the engine does not appear to be in danger, off boost, at this point.

Bear in mind I’m peeping those logs on my phone so results may vary, the rpm glitch is just after halfway through the cruise log, I’ll try upload a screenshot… It might be just a gear change but the cruise/idle log with actuated vvt is still something worth doing to see if the cam signals are steady.

 

[Eth207]

On the latter note, the OCV’s may have been getting 4v each from the 9v battery which is probably at the lower end of the actuation threshold?
Ok, in the logs we’re looking for out of normal range data, and data anomalies.
So I can see your long term fuel trim, af learning, is heading out of acceptable range negative.
But that’s a state of tune, usually a vacuum leak, issue, it’s not appearing to affect knock learning, at least off boost during cruise.
One anomaly I see is the rpm signal diving to zero after a quick peak, I see that as an anomaly: further investigation required, eg start by checking the integrity of the crank pos sensor plug: visual inspection, back probe the pins to the harness wires immediately after the plug and or at the ecu, use some dielectric grease in/on the pins, make sure the pins are clean on the sensor and the plug sockets aren’t over rounded, look for corrosion etc.

Then I’d repeat the cruise log with eg a 9v battery on each OCV, we want to check the cam sensors vs crank sensor, rpm & vvt angles whilst driving around, with the OCV’s pinned to max angle.
Generally vvt angle is high during cruise, so hopefully the car can get moving in that condition.
If not do the test warmed up in neutral using the vary rpm test method I already mentioned. We want to rule out issues with the crank & cam sensors before committing to say a new ecu.

You might address the crank sensor plug checks prior to these tests.

I’ll have access to my laptop next week so I can graph the logs with MLVHD, so no rush, the engine does not appear to be in danger, off boost, at this point.

Bear in mind I’m peeping those logs on my phone so results may vary, the rpm glitch is just after halfway through the cruise log, I’ll try upload a screenshot… It might be just a gear change but the cruise/idle log with actuated vvt is still something worth doing to see if the cam signals are steady.

Interesting. Maybe my crank/cam sensors are all bad, ill see about doing the test while driving hopefully it runs okay. It’s not dangerous for the engine right? It does stumble at idle pretty bad when they’re at max value

I forget what exactly it was, but I was looking at the OCV currents on the cruise log and it seemed to be dropping up and down at weird times.

Also if you’ll remember they use to stay at 9.41% for the majority of this testing. I noticed randomly while driving, both left and right would shoot down to 9.41% for a little bit, like it was wanting to go back to the way it was before, when none of it worked

it’s so strange how the right side just snapped out of it and decided to work

heres A picture of my OCV’s I just thought it was weird how one is less stained, but I guess the swap test proved no change from each side

 

[Eth207]

Interesting. Maybe my crank/cam sensors are all bad, ill see about doing the test while driving hopefully it runs okay. It’s not dangerous for the engine right? It does stumble at idle pretty bad when they’re at max value

I forget what exactly it was, but I was looking at the OCV currents on the cruise log and it seemed to be dropping up and down at weird times.

Also if you’ll remember they use to stay at 9.41% for the majority of this testing. I noticed randomly while driving, both left and right would shoot down to 9.41% for a little bit, like it was wanting to go back to the way it was before, when none of it worked

it’s so strange how the right side just snapped out of it and decided to work

View attachment 70414
heres A picture of my OCV’s I just thought it was weird how one is less stained, but I guess the swap test proved no change from each side

when I do my next test, do you want me to add/change any parameters? I kind of guessed which ones to do

 

[bigBADbenny]

Let’s say you get this sorted, eventually, there is a basic set of logging parameters used for tuning and keeping an eye on the state of tune. It’s a good idea to save that specific profile in romraider or btssm apps, as using that profile makes comparing multiple logs in eg mlvhd app easier.

For diagnosing a specific issue, save that profile under a different name (or vehicle in btssm) till you get to the issue, then revert as above.

The preferred list is in my logging faq, eg Eric’s set with a few added parameters, see the paragraph
Logging Parameters for H4DOTC:
From Torqued Performance

Make and share datalogs! • club.liberty.asn.au

Again that’s more for day to day logging after this issue is fixed.


Loooking at the cruise log, I didn’t see the rpm scale at the left side, I think what I saw as an rpm glitch is a normal gear change.

But try the cruise log again with remote power to the OCV’s, when I use mlvhd, maybe I’ll be able to see any cam sensor faults as vvt dropouts.

For this diagnostic, we can see ol/cl is active, you have coolant temp active, which leaves crank switch, cam switch/es, neutral switch, brake switch (in the RR switches tab) as some sanity checks to log or check live (although surely these would throw DTC’s: can’t hurt to check anyway.

 

[bigBADbenny]

Sure is interesting that with the dtc active, vvt inactive, lhs OCV current is active, rhs not. Does this represent the short referred to in the DTC diagnostics?

Gear position is a bit glitchy, but this is ecu estimated:Gear Position ESTIMATED (Gear Position) -> Current estimated gear position as determined by the ECU. This value is estimated based on RPM and vehicle speed.

So that could be a clue… I’m not sure to what extent datazap smooths data, but even a little will negate dropouts, looking again vs known good data next week

 

[bigBADbenny]

The fsm has two tests to find short circuits, did you get to these as per P1309?

The second test will be the one to do, since we’re not dealing with a burnt fuse.
The first test might have relevance as a voltage drop test.

I’d also look for continuity between pins or ground where it shouldn’t be, to be sure.

 

[bigBADbenny]

Interesting article on harness oil contamination:

Trouble Shooter | Subaru Oil Control Solenoid Valve

www.motor.com

“The PCM uses sensor data such as throttle position, coolant temperature, intake air and camshaft position to determine the optimal amount of cam phasing.”

Having a complete list would be good but the logic is buried in the rom?

So we have those parameters covered in the logs, I’d still recommend temporarily logging the switches I mentioned, plus the driving or idle test with AVCS activated directly with 9v batteries.

Don’t hesitate to make a log of 10-20 minutes of driving and if not safe to drive, idle with the throttle held at 1500-2500 rpm.

Rule out the oily cam sensors (activated vvt signal dropouts being the fault mode), view those switches at work and check the harness for short circuits and you’ll be able to decide on replacing the ecu, or be confident to not burn out the new ecu…

 

[bigBADbenny]

I see there’s also an open circuit test in P1309…
Here’s the fsm explainer how to…

 

[Eth207]

I see there’s also an open circuit test in P1309…
Here’s the fsm explainer how to… View attachment 70417

Alright I finally did a log while powering the solenoids. I wasn't able to get it to actuate with the 9v again (even independently, maybe a bad battery). So I just put both to my car battery.


You'll notice I had to rev it up every time I came to a stop, it idled horribly but I was able to take it around town without stalling.

Cruise log with external power

 

[Eth207]

I see there’s also an open circuit test in P1309…
Here’s the fsm explainer how to… View attachment 70417

Ill run these tests tonight, good Idea

 

[bigBADbenny]

I’ll fire up mlvhd and take a look, plus go through all your tests and post the findings on the local Liberty forum where there’s various specialists I can get advice from: tuning, service tech and mechatronic.

Based on their advice the FSM diagnostics tend to be sketchy, limited focus as regards contributing factors, even the DTC description, which is why, as we’ve seen, the scope of fixes is broad.

Regardless the fsm is useful for the wiring diagrams and pinouts, meaning you still use your own test methodology to find the fault.

I’ll also take a look at all the AVCS dtc codes in the fsm, the codes become more granular with successive generations, meaning P1307/9 might be early umbrella DTC’s that cover a wide range of faults that became granular later.

Don’t forget to reset the ecu each time you do anything to help the fault, potentially an active code will prevent AVCS calibration.
So ideally you’d reset the ecu with a warm engine, which should force recalibration on the first or second key on.

It’s a bit of a hassle needing a laptop for all this, you can potentially use btssm app on android with your tactrix & a female usbA OTG adapter.

So just to be sure, you've uploaded a total of six datalogs?

 

[bigBADbenny]

Ok so here's a glitch in Cruise Log- External power to Both OCV...
I always peep an overview of any datalog looking for anomalies, and I spotted sharp jumps in the custom parameter Engine Cycles which usually closely matches time.

In this very time compressed view, the glitch can me seen in the MAF graph as jumps to max, and I suppose that's where the 80' vvt angle you caught came from too.

If we zoom in we might see all affected parameters which might suggest which harness, ground or system is faulty.

 

[bigBADbenny]

Zooming in, we see almost every engine parameter is affected, briefly failing to high.

Don’t jump to any conclusions just yet, I’ve just started looking and there’s plenty more in the data. Plus the other logs I have not yet graphed

 

[bigBADbenny]

Could you please make another similar driving log, but with the external power to the OCV's removed: the parameter glitch to high isn't present in any other logs, so we want to see if the glitches are isolated to powered OCV's.
You could try that with the OCV's plugged into the harness or not plugged in (if the latter helps with drivability).
Stay away from high loads: I do see some knock on load, do add DAM/IAM to your parameter lists/profiles in RR.
Always include some idle, preferably warm in any driving log, where possible.

Since you've swapped an engine, in the case of a big glitch like this, check the main engine plug/s in the engine bay in front of the bulkhead for these considerations:
Is the weather tight o-ring in place around the perimeter of the plug?
Is the white spacer that sits in all the pins present?
Is there evidence of any oil (or water) inside the plug, or on the pins?
Are there any bent pins or rounded sockets?
Is the plug seating seating and closing properly with the lever fully going home?

Remember I mentioned that Gear Position looked a bit flakey, DRodman etc mentioned the neutral position switch being a known contributing fault?
Is your gearbox a 6mt?
So in Cruise Log- External power to Both OCV, in this section of the drive where you gently accelerated to 50mph, on and off the throttle whilst cruising, did you touch the clutch or gearshift?

 

[Eth207]

Zooming in, we see almost every engine parameter is affected, briefly failing to high.

Don’t jump to any conclusions just yet, I’ve just started looking and there’s plenty more in the data.

View attachment 70470

Very interesting.
Thank you again for all of your help I cant believe how generous you are with all of this!

So I am not positive what I am looking at, is this just like a more accurate view of the log as opposed to datazap? I also noticed the OCV duty cycles would randomly spike a couple times when I was looking through them.

By "max" do you mean the 48g/s on the left side of the screen? I see how the graph goes vertically out of view but why does the "max=" seem normal?


What would this mean as far as diagnosing? Ive never seen anything like this before