Rough Idle & Engine Shaking under load like AC

[bigBADbenny]

FBKC behavior, the lhs plot is filtered for transient data.

Screen Shot 2021-08-25 at 19.00.57.jpg (234.42 KiB) Viewed 3201 times

[bigBADbenny]

So the next step is to look at fuel learning and fuel correction, and use filters to see specifically under what conditions the greatest corrections occur.

[ilib11]

Way, bigBADbenny, that is hell of a lot investigation, thank you heaps, indeed!!!!

I'll try to add some context and confirm the next steps:

Comments:
1) could the power steering be an issue? - I do not think so, the pump has been re-sealed, fluid changed - no change to the engine behaviour, same problem before ad after. Apart from oil sweating fixed with re-seal, no other problems with power steering. For a test, I could take off its belt and see if anything changes, but that would take the alternator out of the equation too.

2)Are all the wires to the a/c pump connected? - yes, everything is connected. It is only a couple wires for the AC clutch. Engine ECU knows when the pump is on - I have two parameters logged, they behave appropriately :
Air Conditioning Compressor Signal (On/Off) <------ actual AC pump on/off
Air Conditioning Switch (On/Off) <------- a button in the AC control panel in the cabin

AC works fine - cools and blows etc.

3)Would be interesting to see TGV in this log, but cold to warm at idle... - I looked at it separately - from cold to idle practically no change, please see min/max values here http://forum.liberty.asn.au/download/file.php?id=20325.

4) The gas test, it looks like the engine responsed to something towards the end of the log - I did remember at the end of the test, had seen no affect from the gas sprayed out, in some desperate move, I sprayed some gas directly into air filter intake. I believe you have noticed exactly that moment in the logs.

5) Roughness, a single instance, cylinder 2. - Once I saw it in the logs, I tried to replicate the scenario but could not do that. At the moment all Roughness Counts are 0 for all cylinders



Next steps:
a) next step is to look at fuel learning and fuel correction - please let me know - do you need any fresh logs for this?

b) Log FBKC / raw knock count - will search the following in RomRaider and do the following logs, will try this weekend. Please let me know if there are any other specific parameters or driving conditions needed for this log:
raw knock count
IAM (Ignition Advance Multiplier)* - current IAM
Fine Learning Knock Correction* - current applied FLKC
Feedback Knock Correction - current FBKC
Fine Learning Table Offset - current applied FLKC cell. If this value remains the same while logged FLKC changes, then the ECU made an adjustment to the FLKC in the current cell (barring any sync issues normal to logging).

c) Check some loom or ground issue - will try to disconnect ground points and clean them and reconnect. That is a next hardware battle


P.S. I am having a second hand engine ECU (same partnumber, practically same donor car) coming in - will be more room to play

And again, bigBADbenny , cannot thank you enough !!!!!!!!

[bigBADbenny]

So on the left we're filtering transients and aceeleration, LTFT below -0.5.

You can see exactly where the car is pulling fuel... under vacuum, decelerating, afr's are too rich presumably because the maf measurements are off due to an inlet leak.

ilib11 driving graph scatter graph LTFT.png (78.87 KiB) Viewed 3200 times

[ilib11]

I am sorry, I am puzzled a bit re afr - rich means we have too much fuel, right? The extra air after MAF would make it lean, and the CPU would try to add fuel to compensate, right?

[bigBADbenny]

Way, bigBADbenny, that is hell of a lot investigation, thank you heaps, indeed!!!!

I'll try to add some context and confirm the next steps:

Comments:
1) could the power steering be an issue? - I do not think so, the pump has been re-sealed, fluid changed - no change to the engine behaviour, same problem before ad after. Apart from oil sweating fixed with re-seal, no other problems with power steering. For a test, I could take off its belt and see if anything changes, but that would take the alternator out of the equation too.

2)Are all the wires to the a/c pump connected? - yes, everything is connected. It is only a couple wires for the AC clutch. Engine ECU knows when the pump is on - I have two parameters logged, they behave appropriately :
Air Conditioning Compressor Signal (On/Off) <------ actual AC pump on/off
Air Conditioning Switch (On/Off) <------- a button in the AC control panel in the cabin

AC works fine - cools and blows etc.

3)Would be interesting to see TGV in this log, but cold to warm at idle... - I looked at it separately - from cold to idle practically no change, please see min/max values here http://forum.liberty.asn.au/download/file.php?id=20325.

4) The gas test, it looks like the engine responsed to something towards the end of the log - I did remember at the end of the test, had seen no affect from the gas sprayed out, in some desperate move, I sprayed some gas directly into air filter intake. I believe you have noticed exactly that moment in the logs.

5) Roughness, a single instance, cylinder 2. - Once I saw it in the logs, I tried to replicate the scenario but could not do that. At the moment all Roughness Counts are 0 for all cylinders



Next steps:
a) next step is to look at fuel learning and fuel correction - please let me know - do you need any fresh logs for this?

b) Log FBKC / raw knock count - will search the following in RomRaider and do the following logs, will try this weekend. Please let me know if there are any other specific parameters or driving conditions needed for this log:
raw knock count
IAM (Ignition Advance Multiplier)* - current IAM
Fine Learning Knock Correction* - current applied FLKC
Feedback Knock Correction - current FBKC
Fine Learning Table Offset - current applied FLKC cell. If this value remains the same while logged FLKC changes, then the ECU made an adjustment to the FLKC in the current cell (barring any sync issues normal to logging).

c) Check some loom or ground issue - will try to disconnect ground points and clean them and reconnect. That is a next hardware battle

P.S. I am having a second hand engine ECU (same partnumber, practically same donor car) coming in - will be more room to play

And again, bigBADbenny , cannot thank you enough !!!!!!!!

Bear in mind the ecu is usually coded to the security handhake set, ecu swap is the whole lot ot a trip to the dealership.

The knocks parameter I made for btssm: MY07 GT spec B II |A2UG002T|A|0xFF7C18|1|KNOCKS|Knocks|count|0|1|0|<0

Pretty sure its something I extrapolated from the romaraider site from the standard logger definitions...

Please just use the basic tuning parameters plus a few extras to suit, for my sake as the filters might work in more instances, eg afr, not lambda.

Logging Parameters:
From Torqued Performance:
A/F Correction #1(%)
A/F Learning #1(%)
A/F Sensor #1(AFR)
Engine Load (g/rev)
Engine Speed
Feedback Knock Correction
Fine Learning Knock Correction
IAM
Ignition Total Timing
Injector Duty Cycle
Intake Air Temperature
Manifold Absolute Pressure (PSI)
Manifold Relative Pressure(corrected)(PSI)
Mass Airflow (g/s)
Mass Airflow Sensor Voltage
Throttle Opening Angle(%)
Vehicle Speed (mph)
Optional (but mandatory for health check):
VVT angles (L&R)

Plus any parameters you feel the need to diganose.

Yes its kinda counterintuitive how the engine responds to leaks, but do post more logs and LV's.

Personally if my car is having issues I do them regularly enough to establish a pattern, which helps get to the issue.

[ilib11]

Thank you, will do more logs than post here too. I have not used btssm yet, will try to play with Romraider to get the same knocks parameter or learn how to use btssm.

On a side note I think I know how to preserve immobiliser settings and move it to another ECU without going to a dealership, would like to get it proven in practice first. I'll let you know how it goes.
Plus I have heard there was a common problem with 6cyl ECU - its ignition coil driver chip often got cold-soldered to the ECU pcb and that created all sort of issues. Common fix was just re-solder. I have heard many people suffered from spark plugs, ignition coils, etc replacement first before they looked to ECU as the last resort and it is where the problem was. But these stories were for 6 cyl, so I am not that excited. With the new ECU will have a chance to look closely into that area.

[bigBADbenny]

I’d line up an ecu unbricking service as a backup, locally try member Madman Sean aka Sick Chips.

In closed loop, the ecu is trying to maintain stoich, based on how the maf & af sensors are responding to the running conditions including other sensors.

Inlet and exhaust vacuum & pressure, air leaks essentially, will upset fuel learning and spark timing.

Fuel learning is said to be of concern at >+/-5%, so I’ve filtered for that…


viewtopic.php?f=18&t=36648&start=45#p451739

This graph intrigues me, you can see correlation between the first plot, and the ac plot, exactly where the propane test gas is introduced to the inlet.

I want cold start to warmup with no ecu reset, ac, tgv, afr & tuning parameters.

Once warm you can redo the propane and brake clean tests. One increases rpm iirc.

[bigBADbenny]

All the graphs posted above… you can get a rough idea just from looking at a Learning View.

MLVHD:
If you log open/closed loop, you can set filters accordingly.

[bigBADbenny]

This tip pinpoints CV axles, aftermarket or simply worn out.
This is for idle vibration but not necessarily under load like ac etc.

Try it anyway, the logs do show you have state of tune issues.

8ED9EBE8-8FEE-4027-86FF-AACFDDF5753A.jpeg (95.03 KiB) Viewed 3166 times

[ilib11]

Wow, this rabbit hole becomes deeper and deeper - will add a CV check to the list of tasks for the next service visit, thank you for sharing the info. ( However it looks strange for me how CV would affect idling but otherwise stationary vehicle. I am I right we are referring to Constant Velocity joints that connect transmission and wheels?)

Meanwhile I have done more logs & Learning View screenshots under different conditions.
They cover the following:
-warm up,
-Idle with added load,
-normal driving (2 pcs with standard and extended set of logged parameters)
-gentle braking.

The logs and their descriptions are here https://mega.nz/file/kfYFRApS#F6vwwdaAR7DOOxU5Km8YDayNytYR47TN78BVgGWr4UA


I also made a couple of videos here - it is engine on idle with aircon ON/OFF. At the time of the video the engine has no added RPMs in User Defined RPM Settings and it is vibrating. Once AC is ON you can hear the engine tone changes and you can notice an extra vibration.
https://mega.nz/file/YPAx2SyY#TatjOAkFKr0uTlcVao72h8ZC_AUQ-vTstdZZGjyjniY
https://mega.nz/file/wTJTVKaR#8OjITPB3dr6Te55uZLSZjQtLPfekUbPfRokYzmatt3w

Hope that helps.
With a huge appreciation for your attention and help,
ilib11

[bigBADbenny]

Awesome, I’ll try get to the logs asap.

Yes just do the test as described in the screen shot.

Just another heuristic counter intuitive tip from an actual expert.

They love rule of thumb, and I’m on board with that.

So I think this tip is for autos… and load like a/c is not relevant.

Let the vibration build up stopped, release brake and see/feel the vibration behaviour as you stop again.

[bigBADbenny]

Well I did look at your logs and the car needs to be health checked ASAP.

There’s an obvious fault whilst driving the car.
Eg the ignition advance multiplier, IAM, is all over the place but not returning to 0.00, and this was a short drive.
Whist it’s not appearing to knock, it is in rough knock correction mode.

I’d suspect a boost leak, in part because max boost is 12psi, and so on.
Did you try for boost?

Iam improves with load too.

My advice is to stay out of boost till you sort it out. I’ll post pics ASAP.

Health checks: load test battery, inlet pressure & leakdown, exhaust pressure test, fuel pressure test.

You did the propane idle test, this does not account for boost leaks, plus you probably didn’t get to the piping between the low mount turbo and the maf and intercooler. Do an inlet pressure test.

The first few are easy to do, ok?

If you need coaching, all the info is online and here. Just ask

Please edit your logs in excel, multiply lambda by 14.7 and call it afr in the header please
Or remind me how to

[ilib11]

Hi bigBADbenny,
Thank you for looking into the logs again! The results are scary! Just some clarifications on recent tests & conditions:

- What is expected max boost value? During driving, I did push for max open throttle. The car was in " I - Intelligent mode"' ( there are two more - S and S# available) and for this mode boost is not too high, 12 psi/0.8 bar seems to be ok as a max value. In other modes I remember I was able to get 1 bar or a bit more (14+ psi).

- Battery is 3 month since new. No changes in behaviour comparing to old battery.

- Fuel pressure test was done a month ago, values as per spec.

Need to find a place in Brisbane where inlet pressure & leakdown, exhaust pressure test can be done properly.

[subyroo]

Need to find a place in Brisbane where inlet pressure & leakdown, exhaust pressure test can be done properly.

https://www.subievolution.com.au/ or http://www.boxerservice.com.au/ in Brendale if you are near the Nth side.

Boxer service took care of my Forester for almost 13 years. Stifull used Subievolution recently, PM him for his opinion.