Can anyone make anything of this? - Odd idle at cold start

[jigga009]

This is a Subaru 2.5 application, with headwork, non-avcs GSC 282 cams, pistons, rods, etc. This is an issue that I have been poking around at on and off over the past year when time allowed, and so far, no obvious solution to me so far. With that said, I'm no expert, so perhaps something pops out at some of our resident experts here.

I have tried installing a new IACV, but that does not appear to have solved the issue.

Symptoms - The car cranks over and starts up fine, but doesn't sound happy idling in general...It is popping and banging out of the exhaust at idle. I have not driven the car on the road at all with it in this condition, so all I have is an idle log, as I don't really want to run the engine much if it is obviously unhappy about something. The car is not shaking by any means (even with my Group N engine mounts), but I can definitely hear that something is off.

 

History-(sorry for the essay, but I figure the more info the better?)

This all started shortly after I had the exhaust manifold off the car to install a few bungs for widebands. When reinstalling the exhaust manifold, I used a fresh OEM gasket with copper gasket sealer, since I noticed that it had been used in the past by my trusted shop. I drove the car this way for half a tank of fuel, and noticed on the newly installed 4 channel wideband that there was a difference in readings between both banks.  I could also smell an exhaust leak emanating from the front of the car. The car was not popping and banging at idle like it is now though. 

At the start of what was to be my 2nd shakedown trip on a new tank of gas with the new per-cylinder wideband setup, I noticed upon starting the car that the engine was popping and banging at idle cold start quite prominently out of the exhaust.

Back then, the guys at Link thought that it looked like an intake or exhaust leak. I used carb cleaner around areas of the intake manifold and injectors and found some leaks and fixed them.  Car still popped and banged and idle.

I then had the injectors cleaned and flowed but the car continued to run the same way upon re-installation and pressure testing .

I then pressure tested the exhaust system by forcing air up the system from the muffler. I quickly found a decent leak at a v-band connection  due to an installation error on my part when I reinstalled the exhaust manifold after getting the wideband bungs welded in.

Off came the exhaust manifold again, cleaned the copper gasket maker/sealer off the head and exhaust manifold, and then reinstalled a new OEM Subaru MLS exhaust manifold gasket and improved installation technique to prevent the v-band leak.  This time, I did not feel like using copper gasket sealer like had been used last time, or by my go-to shop previous to that. I assumed that it couldn't be doing much. I thought that a standard OEM MLS gasket would do the trick without any additions. 

Once re-assembled, the car sat for a few weeks as I decided to make a few more changes that I had been planning on doing, so needed to pull the exhaust manifold off again while making the following changes:

- switching from 12mm oil pump to a 10mm pump (as recommended by the guys at Killerbee due to some oddities I had seen on previous logs with oil pressure)

- Fixed some vacuum leaks that I found at the injectors

- install upgraded oil pickup tube and oil pan and baffle

- installed Fluidampr crank pulley

- installed Group N engine and transmission mounts

- Ran a dedicated -ve wire from the battery (in the boot) to the engine block, as well as upgrading the alternator power wire and ground links between the engine and chassis. Previously, only a positive line ran from the battery positive to the front, with negative going straight to the chassis in the boot. Set up this way, I was noticing a voltage drop at high RPM on previous logs, as well as relatively low cranking voltage. 

- Fixed an exhaust leak between the downpipe and turbo.

During final assembly though, I re-used the "new" exhaust gaskets I had installed and torqued down just a  few weeks earlier (but had not yet run the car on) again without copper gasket/sealant. I did run the air compressor again up the exhaust of the car to feel and listen for air rushing out anywhere leading up to the engine, but I did not notice any. 

I suspect re-using the new previously-torqued MLS gaskets without a coat of copper gasket sealer  *might* have been the wrong decision to take, but I'd like to get some more opinions based on my data logs in case there is something else that catches people's eye as far as a problem to look at.

Is there anything out of the ordinary that anyone can see on this datalog and map? It is a cold start log at idle, but I think i did blip the throttle slightly a couple of times. Any other checks that can be easily done?

Datalog and Map

The car has an almost full suite of sensors including Link Can Lambda in the downpipe (Lambda1), and exhaust gas pressure-compensated per-cylinder widebands as well:

Lambda 2 - cyl 1;

Lambda 3 - cyl 2;

Lambda 4 - cyl 3;

Lambda 5 - cyl4.

I did notice that exhaust gas pressure was almost non-existent for most of the log, which I would normally think might indicate some kind of exhaust leak (??), but I don't have enough experience with exhaust gas pressure to know what is normal at idle on an engine.

Techs at Subaru were confident that as long as the engine had not been run on the new MLS gaskets (which it had not been), it was perfectly okay to reuse them without the fear of leaks even if they had been torqued down on already. Apparently it is something they do all the time. Now I'm not so sure, but who knows...

Perhaps our resident experts here see something in the logs? Could fixing a few vacuum leaks at the injectors throw off the ability of the car to idle and require changes to the fuel table? Must there be an exhaust leak still? Or is there something else going on?.

Thanks in advance for your time. 

 

 

[remski2]

Hi.

Try chaning the "Spark Edge" to Falling. 

Next check your injector Min Effective Pulse Width... seems kind of high.

Also I see you have staged injection running.. just making sure that this is what you have.. running 8 injectors..

Not that of a common thing in a Subaru. 

You seem to be idling very lean.. 16-18AFR.. that might be a part of your problem as well. 

[Adamw]

As remski mentioned that spark edge is wrong, that has potentially fried some coils running like that. That would also make ignition timing wrong so check base timing again after changing the edge.

Other things that jump out at me is MAP at idle is 80-85kpa which is very poor vacuum, possibly cam timing wrong? 

 

[jigga009]

Remski2 -  The car is on a CDI ignition with M&W/MSD coils. Would that change your view about the spark edge being wrong? Am I still potentially looking at some fried coils given this info?

And yes, it is running 8 injectors.

Adamw - Thanks for chiming in. The vacuum is an interesting one for me also. The engine is on some new cams (switched from some Kelford custom mystery units to some GSC 282s, and am still getting to grips to what my new "normal" is with respect to vacuum at idle. I'm pretty sure that my timing belt hasn't jumped any teeth, but I can pull the covers off to have a look, and also verify that the ECU and engine are in fact on the same page timing-wise.

 

 

[jigga009]

Just checked some old notes I have regarding the CDI system, and I'm pretty sure I have them set correctly given that it's running standard (dumb) coils that have no built-in drivers.

I have the CDI boxes themselves set for Rising Edge, to account for them being used with non-factory dumb coils.

 

[Adamw]

If the CDI is definitely set to rising edge input then your spark edge is correct.  But it would pay to check as the M&W's come set to falling by default.  Selecting the wrong edge with CDI wont fry the coil - you will just have lots of timing drift, so unlikely your problem.  

CDI's can give a pretty shitty idle though, so that alone may be some of your problem.  

[jigga009]

Thanks for the input. I'm not really a fan of blindly throwing parts at a problem, but in order to rule out physical timing as a culprit here, I'm considering installing a new timing belt and tensioner, even though the one ones on there currently have less than 5K on them. Let me explain...

When installing my oil pump and reinstalling the existing timing belt and pulleys, I noticed that the cam pulleys on the offending left side cylinder bank (cyl 2 and 4) did not *quite* line up, while the marks on the cam pulleys on the cylinder 1 and 4 bank was 100% bang on. I did count the individual teeth on the timing belt to confirm proper belt positioning according to the Subaru manual, but I still can't help but wonder now if the gears on the offending cylinder bank are off just enough with the new cams to cause the banging and popping in the exhaust, the low vacuum you mention, and the occasional blips showing up in the map signal and my boost gauge as the car idles. There are concentric idler pulleys that are on the market to correct the issue with the timing mark alignment.

Here are pictures of the cam pulleys with belt installed -

https://www.dropbox.com/sh/xx8tktjz952qkph/AAC0b-G9Y6R2rbeFrs557Fyna?dl=0

Again, the side with the marks lining up seems to be the bank running AFRs I am accustomed to seeing on this map, while the other side of the engine seems to be running similarly also, but with a much leaner AFR than I would expect to see, with the popping, banging and sporadic blips showing up in the map sensor trace and my boost gauge needle at the same time.

Thoughts?

 

[jigga009]

Small update..

Had a chance this afternoon to check the physical timing on the engine, but found nothing out of the ordinary.  I was hoping to find some jumped teeth to explain the poor vacuum, but it's fine.

I also had a chance to talk to a Subaru shop that I get parts from (actually about ordering the timing belt pulleys and tensioner to replace everything there), and mentioned what was going on. They mentioned that the cam gear misalignment I was seeing was very typical with Subarus and likely not the problem.

 

They gave an explanation for AdamW's observation of poor vacuum at idle ----  Apparently, GSC's cams I have are AVCS cams, and running AVCS cams without the AVCS active (with WRX fixed cam gears like I do) will produce relatively poor vacuum at idle since you are essentially sitting in the "high RPM" position of the cams all the time. The real solution they say would be to either switch to dedicated non-AVCS cams, OR activate AVCS on my car, OR increase my idle speed from the 900-1000rpm that it was idling at during my logs. For comparison, they run 264 cams on their shop car, and have to idle at 1000rpm to prevent misfires, so were certain that that my idle speed of 900-1000rpm is just too low for the the cam I have, and that is why the car is misfiring at idle only and then behaving itself when I give it some throttle.

I just checked the log I posted, and noticed that while the engine was idling at 900-1000rpm the entire time, the target idle speed set in the ECU is actually 1488rpm, so it definitely is idling slower than it should be for the duration of the log, perhaps this explains the misfiring?

They think I must have run into the issue seemingly out of the blue because the engine must have had some intake vacuum leaks at when the new cams went in initially that my tuner remotely adjusted the tune for with the map I posted above, but some of the leaks may have sealed up as I started shaking down the car and introducing higher fuel pressure (pushing/seating the injectors down into the ports). Once this started happening, it threw the idle settings off again, making the car misfire at idle. Once I started diagnosing and pressure testing the exhaust and intake tract of the car (and found the vacuum leaks at the injector ports and some in the exhaust manifold), we are now seeing what idle/IACV settings which worked before on an engine with intake vacuum leaks is now doing on an engine with a leak-free engine.

They think I need to increase the idle speed from 900-1000rpm and that should bring an improvement in idle vacuum (although admittedly not as good as if I were running AVCS at idle), and should stop the idle misfiring as well.

I'm no tuner, but does this all make sense? 

I can't test this theory right now as I'll wake my subdivision up with the car, so it will have to wait until tomorrow. 

 

Question - 

In order to raise the idle speed to match the set target idle currently within the map, can you confirm that I would be adjusting or dragging the entire "Idle Base Position" table/graph curve up a few points to the point where the Idle target error is 0 at any given ECT?

I figure that if increasing the idle speed to where it should be doesn't solve it, and both banks don't start showing similar AFRs with increased idle speed, I'll focus more on the physical timing of the engine.

Thanks again.

 

[remski2]

Seems like the shop has no clue how the AVCS works.

AVCS should not run during idle.. no point in it.. it just robs your engine of oil.

I run my AVCS GSC2 cams at 1100 as this seems to be the smmothest. 

 

AVCS cams or not.. doesnt matter as you could run your AVCS cams and never turn on the AVCS solenoids.

 

Have you done a boost leak test ?

[jigga009]

Thanks for chiming in; I appreciate your expertise here.

I may have had their explanation down wrong. They did not say that AVCS is in use at idle, but something about the orientation of the WRX cam gears on the AVCS cams means that the combination effectively locks on into a situation you would have if you did not use AVCS at all? In addition, would lead to worse idling characteristics. It's something I have to do a bit of research on admittedly. I do know that often times people use AVCS to tune out the lumpy nature of cams, but again, it's all currently beyond my pay grade.

I may visit AVCS down the line, but from the research I have done on it, for all of the benefits it definitely gives, it does also seem to give some headaches when it doesn't work properly. Hence the reason I have stayed away from it and just put up with the loss in low end torque for now.

 

Yes, boost leak testing with my air compressor as well as smoke testing of both the intake and exhaust of the car.

Intake currently shows zero leaks anywhere. All of my injector ports ( some of which were found to be leaking when I first ran into this issue and started diagnosing) are now tight as a drum.

Smoke testing of the intake showed only minor thin trail of smoke emanating from the turbo's compressor housing/chra V-band joint, and the external waste gate dump tubes. I even tried pressurizing the smoke into the intake system to force any leaks out and nothing. It all showed up underneath the car at the dump tubes as well as a thin wisp at the V-band connecting the compressor housing to the turbo centre section.

Boost leak testing the exhaust system shows no major leaks. If I set the regulator of my 15 gallon compressor to 40 psi and send it up the exhaust pipe, I can get the v-bands on my twin scroll exhaust manifold to very slowly create a bubble or two when hosed down with soapy water. You couldn't feel air escaping with your hand next to them though.  I would imagine that they are so minor that they would seal up when the exhaust manifold gets hot. No air leaks at the exhaust manifold-head junction.  Smoke testing the exhaust system shows no leaks either, except from the exhaust waste gate dump tubes.

Apparently Garrett considers some very minor leakage normal with their turbos since they don't use o-rings between the CHRA and compressor housing on their larger frame turbos. TIAL also apparently considers leakage normal with their waste gates when cold as well. 

Is your idle RPM for when you are up to temperature already? What RPM are you targeting on a cold start with your cams?

Thanks again!

 

[remski2]

I start with 1400 cold and taper off to ~1100.

2nd reason to run ~1100 is to keep my oil pressure up when hot. 

Are you running ej25 series block ?

[jigga009]

7 hours ago, remski2 said:

I start with 1400 cold and taper off to ~1100.

2nd reason to run ~1100 is to keep my oil pressure up when hot. 

Are you running ej25 series block ?

Gotcha. Looking at my map, it looks like my tuner was originally targeting around 1400rpms when cold also, but with no leaks in the system now, its sitting at 900-1000rpm as soon as I cold start it. I've been a little occupied today reading up and watching youtube tutorials on the ins and outs of adjusting the idle before I attempt to touch anything on the ECU.

What size oil pump are you running? I'm guessing you are running the ACL/Orbit 12mm if you need to keep the idle speed up for oil pressure?

And yes, EJ257 block.

[remski2]

With EJ257 are you sure your Trigger Mode is setup correctly ?

It shows JDM v7-10

EJ's have a Reluctor for Trigger1 and Optical/Hall for Trigger2 

Btw... I see trigger errors in your log.. 

Why CDI over the factory coils that come with Ej257 ?

 

Oil pump from what I remember is 11mm. 

My oil pressure at idle when hot is ~34psi

[jigga009]

It is a EJ257 based block with JDM Version 7 heads utilizing WRX non-avcs cam gears and triggering pattern. The car is a GD WRX, so I decided for simplicity's sake when it was put together to stick with a WRX triggering pattern since options for the GD chassis were a little more limited when it was put together. Back then, AEM's V1 and Hydra EMS were the only standalone games in town, so I needed to maintain a triggering pattern that we knew would work with them. I eventually switched to a Vipec V88 when I got tired of bugs related to the the other two EMS solutions, and have been with Link since. 

As for the CDI ignition system, the car had been tuned to the past the mid 700's whp on a Mustang dyno in the past and ran into system voltage issues peculiar to the electronic demands of the car at elevated power levels that necessitated the switch. We did try new coils, different alternators, and even thought of installing an audio capacitor in line to buffer the electrical system, but eventually concluded that we needed to reduce the electrical demand on the electrical system to keep up with all the existing electronics in the car along with the increasing ignition demands as boost and power were increasing, as evidence of misfires were beginning to show up on the dyno traces as system voltage was starting to experience an inverse relationship to boost, whp and wtq. 

As for the triggering errors, I have to admit to not being knowledgeable enough as to what might be causing them. I can check some old logs going as far back as when I used to run a Vipec V88 to see if they were present on the same car, and confirm that the same triggering settings made if over the transfer a few years ago. Could the triggering errors be the result of the misfiring occurring at idle now? 

I'm all ears for any ideas or things to check if you have any.

Thanks again.

[jigga009]

2 hours ago, remski2 said:

With EJ257 are you sure your Trigger Mode is setup correctly ?

It shows JDM v7-10

EJ's have a Reluctor for Trigger1 and Optical/Hall for Trigger2 

Btw... I see trigger errors in your log.. 

Why CDI over the factory coils that come with Ej257 ?

Just checked my old V88 maps, and I'm pretty sure the trigger mode is set up correctly for my particular engine setup. It hasn't changed since the car originally had a V88 installed and tuned. Still not sure what to make of trigger errors though...

[jigga009]

2 hours ago, remski2 said:

With EJ257 are you sure your Trigger Mode is setup correctly ?

It shows JDM v7-10

EJ's have a Reluctor for Trigger1 and Optical/Hall for Trigger2 

Btw... I see trigger errors in your log.. 

Looking at Link's help file regarding Trigger 1 errors, it would seem that "Trigger wheel or trigger sensor mechanically moving due to engine vibration or cam-belt stretch" is one of them... Perhaps I will need to change the timing belt, etc. and install the concentric idler pulleys to bring the timing marks on the Left side head (where cam sensor is located) into 100% alignment? Would you be in agreement, or do you have a different read of things?

Any assistance you can render would be most appreciated  

 

Thanks.

[jigga009]

Was looking over the log file some more and the trigger errors Remski2 mentioned, and noticed something:

Why is it that the ECU is logging 3 trigger errors before the starter motor is engaged? It does seem to pick up another trigger error during cranking the engine over, but none while the engine is actually running.

No errors are logged while the engine is actually running, but it does log another 3 errors when the engine is turned off. Not sure what the deal is with the blip in RPM at the end either.

 

[dx4picco]

Trigger errors may induce unreliable rpm like you shown during shut off. 

Engine can be turning over without the ecu showing rpm. Ecu needs first to sync with the triggers and will then show rom. 

So the beginning is maybe starter motor engaged and during the sync process it gave 2 errors, synced showed rpm, and gave a 3rd error afterwards. Nothing to be worried about I would say

[Adamw]

13 hours ago, jigga009 said:

Why is it that the ECU is logging 3 trigger errors before the starter motor is engaged? It does seem to pick up another trigger error during cranking the engine over, but none while the engine is actually running.

That log shows the trigger errors already existed when the logger "reconnected", so those trigger errors happened at some time when you werent logging - probably the last time the engine stopped.

 It is not uncommon to get errors when the engine is stopped - this is usually due to the expected cam tooth never reaching the sensor before the engine stops.

But as you say, provided you arent getting trigger errors when its running you are fine.

[jigga009]

21 hours ago, Adamw said:

That log shows the trigger errors already existed when the logger "reconnected", so those trigger errors happened at some time when you werent logging - probably the last time the engine stopped.

 It is not uncommon to get errors when the engine is stopped - this is usually due to the expected cam tooth never reaching the sensor before the engine stops.

But as you say, provided you arent getting trigger errors when its running you are fine.

Thanks for chiming in; I'm very appreciative.

I've started pulling the front of the engine apart to inspect the timing belt, pulleys, etc. to rule in/out your suggestion of cam timing being off. Have a puller on order to pull the crank pulley off, which should arrive Monday. Also ordered a new Gates Racing timing belt to throw on. I figure the fresh installation marks on a new belt will reduce any potential errors associated with installation.

From my cursory inspection of the cam pulleys on both sides of the engine though, the timing marks aren't quite lining up, but I will have a better idea on Monday when I can get the crank pulley off to rotate the crank cog/gear (for crank sensor) into the right position to accurately check. 

Would one expect to see trigger errors during engine operation if cam timing was slightly off by a tooth or two?

[jigga009]

Good thing AdamW had me look at the cam timing.. 

I was inspecting things, and found this...I thought I was seeing things initially, but my finger nail quickly told me different. 

 Focus on what the the silver arrows are pointing at..

Gotta love Subaru and their "plastic" cam gears..

I know I've heard of a few horror stories of these things flying apart seemingly out of nowhere..Looks like it was only a matter of time...

[jigga009]

Update to this..

What I have been doing since last post:

I spent the past few days going over everything again.

- Went over the cam timing to ensure it was on point, and like Adam predicted it was in fact a tooth off on the left exhaust bank. Fixed that issue.

- Installed new intake and exhaust cam gear pulleys to replace the cracked one I found.

- Also installed Roger Clark Motorsport adjustable concentric idler pulleys to adjust cam timing on left and right banks to exactly the same, since the block and heads have been decked in the past. Now timing marks on both sides of the engine are 100% on point, with timing marks lining up everywhere.

- Switched out my MSD/M&W CDI coils for a set of known-good MSD Blaster SS CDI coils (in order to rule out coil issues). The Blaster SS coils also have a 2x longer spark duration than the previous MSD/M&W 8232 CDI coils, so I figure that it could only help at idle and with road manners.

- Replaced all the spark plug wires with fresh MSD spark plug wires, as Cylinder 2's spark plug wire had significantly higher resistance than the others (something like 2 or 3 kilo-ohms iirc compared to the 50-100ohms of the other 3 wires of different lengths), and would occasionally show "open" when moved around while testing with an DVOM. Seems perhaps these spark plug wires degrade over time? Previous wires were ~7-8 years old.

- Installed new NGK BKR7EIX Iridium plugs gapped to 0.027. The previous nickel BKR7E plugs had a gap of 0.025 that came out carbon fouled after all of my idle testing over the past few months, but the ignition test mode on the G4+ showed that the CDI system didn't seem to notice, firing the carbon-fouled plugs with authority...

- Used the Ignition "test" function on ECU to confirm that the 2 ignition channels for each cylinder bank were firing, so I know the CDI boxes for the Left and Right banks work.

- Boost leak AND pressure-smoke tested the intake and exhaust. No new leaks found.

 

Datalogs:

Cold start warmup extended Log - Dropbox Link

Current Map -  Map dropbox link

 

Lambda 1 = Downpipe wideband

Lambda 2-5 = Cylinders 1-4

EGBP and coolant pressure are also present if needed.

The idle vacuum was never much to write home about on the previous Kelford "278" cams either, but I do recall it being slightly better than what I'm seeing so far with the map set up as is on the 282 cams. With that said, my idle speed with the previous cams was higher than is currently.

 

Observations:

- Started the car up, and although I'd say it ran better and differently than before, so I'm guessing that the changes I made did something. If anything it sounds rather crisp... a bit too crisp...like "lean" crisp, which the AFR gauge confirms.

- I was able to take a longer idle log this time from cold start to almost 90 deg C. Log ended as engine died when I gave it a bit more throttle to see what happens. As you can see from the logs, the idle speed doesn't seem to change from cold start to warm. Not sure if it's the basic ISC settings that need tweaking or post start enrichment?

- IACV is definitely working - I disconnected it soon after startup, and the idle RPMS went down slightly. Once I plugged it back in, they went up slightly again slightly, so I know the ECU has command of it.

- I tried to blip the throttle a few times along the way, but its too lean lean, and the revs don't really want to go up without the engine stalling out.  I suspect it is genuinely running lean now given the new cams, no intake or exhaust vacuum leaks, new plugs, new plug wires, and concentric idler pulleys for cam timing and new ignition coils in place. I suspect it was in fact lean because I could actually tolerate sitting in the car with the garage door partly open during this test (not safe, I know, but I need methods to keep the noise the car makes down as much as possible while I get things figured out enough to drive it out of my subdivision). In addition, there is normally a sooty wet spot I'm used to seeing found on the ground behind the exhaust on my previous tuned cam setup after a cold start from enrichment/condensation. It wasn't there today.

- The disparity between cylinder AFRs definitely seems to lessen as the engine gets up to temperature (even though still lean overall). I suspect that closed loop lambda might be partially responsible for this, but it wasn't active all the time since the idle rpms sometimes fell under its activation threshold.

- Idle speed is still low compared to the target that the previous cams ran with. I'm guessing its something relating to IACV settings? Or is it fueling that would be adjusted?

Overall "seems" to run a little bit better as it gets up to temp, perhaps as it is less reliant on cold start enrichment at higher ECT's, but it is still lean compared to the old "278"cams...perhaps a side effect of the new 282 cams and adjustment to cam timing? With that said, downpipe CanLambda tells me that the engine as a whole is running lean. 

- Much-reduced penchant to pop and bang now. Just a few crackles every now and then, but no massive gunshots like before. Guessing correcting the cam timing and current lean AFR has something to do with that?

 

Additional History/backstory/context:

- Car was previously fully tuned running some mystery custom "278 degree" Kelford non-avcs cams with 2001/2 WRX non-avcs cam gears and drove around for a number of years successfully before eating a cam bucket.

- Kelford were unable to duplicate said cams as they were produced at a time when they themselves were not keeping tight records on what they were putting out, but admitted that the cam was definitely theirs due to certain features on the cams. 

- Upon removal from engine, damaged Kelford cams had "280" engraved on them. I always thought they were 278's, and purchased them new from as part of my engine package based on that assumption.

- Switched over to GSC S3 282 AVCS cams after showing owner of GSC the damaged cams that were in the engine, and requesting for something that would perhaps give similar performance/manners in an identical use case. His assessment after seeing the old cams was that Kelford supplied AVCS cams without the oil passages for AVCS, and that I can do similar (with the capability to later activate AVCS if I choose) with GSC cams by plugging the AVCS oil holes myself prior to install. He took into account my setup, and recommended S3 cams over the S2's (which I was initially eyeing up).

- Car was reassembled with GSC AVCS 282 S3 cams with original non-AVCS WRX cam gears as was before.

- Shop that did the work had some issues after with car not wanting to idle with the new cams due to AFRs being excessively rich at idle, but their assumption was that it was tune-related. They dialed down the fuel pressure as a temporary stopgap to prevent washing down the cylinders. After getting the car back, I installed a 4 channel wideband, and then started investing in equipment to properly boost test, smoke test and troubleshoot the car. Found boost leaks in injector ports, exhaust leaks, loose map sensor wiring that made the ECU think that it was at ~0.5 psi of boost while at idle (and provided fuel accordingly).

- Took time to troubleshoot the car, getting map sensor wiring connection repaired,  8 injectors cleaned and tested, boost leak tested, repairing boost leaks at injector ports, restoring fuel pressure to 55psi differential pressure at idle, new fuel filter, exhaust leaks solved, etc. Also, corrected cam timing, new ignition wires, spark plugs, synchronized cam timing between L and R cylinder banks.. Basically to clean up all loose ends from when the new cams went in. Now engine appears to run as I would think an engine would run if one switched to a more radical set of cams without making other changes.

 

Current Problem:

- Now when the car is fired up with the above corrections/changes, it understandably runs quite differently compared to with the previous cams or even the previous log I posted in this thread, and it runs about as well as I would expect if one switched to a more radical set of cams on a tuned map without making other changes. I'm unsure as to what should be getting adjusted and how much in order to get it to:

1) Go through the warm up process at RPM's that at least match the previously set target idle speed. 

2) Go through the warm up process at an appropriate AFR.

3) Settle at an appropriate idle AFR of around 1250 RPM when up to temp.

I suspect that once idle rpms are increased appropriately and fueled properly, one might see some more respectable idle vacuum figures.

As is now, with all of my intake and exhaust leaks fixed and cam timing corrected, the previous tuned map can start the car, but not much else on the new cams.

Hopefully the longer datalog I have this time gives a more comprehensive idea of how it's all running.

Any assistance would be very much appreciated.

 

 

 

[Adamw]

I would start by adding say 10-15% to the whole fuel table to get the lambda somewhere closer to target.  The idle valve is maxed out at 90% for the whole log so I suspect you will have to adjust the throttle stop so that the throttle blade is a little more open.  Ideally you want the idle valve somewhere around 20-50% when warm.  

[jigga009]

On 4/12/2021 at 5:35 AM, Adamw said:

I would start by adding say 10-15% to the whole fuel table to get the lambda somewhere closer to target.  The idle valve is maxed out at 90% for the whole log so I suspect you will have to adjust the throttle stop so that the throttle blade is a little more open.  Ideally you want the idle valve somewhere around 20-50% when warm.  

I added 15% more fuel to the entire map and took a log. It runs better, but it's still on the lean side once the cold start enrichments start to wear off and/or closed loop fueling kicks in (guessing). I would imagine that the map would need more fuel, but I'm hesitant to make any changes without consulting someone who knows this stuff.

 I have not had a chance to play too much with the idle stop yet. I only loosened the locknut for the adjustment screw and it *did* turn ever so slightly (maybe an 1/12th of a turn). I was more focused on the AFRs of the engine due to the fuel addition made to the map.

I am now also able to give the car a gentle rev without it cutting out due to being lean (which I was unable to do before), but I suspect that more fuel might be needed. Before adding the 15% to the fuel map, I could not do anything with the throttle while the engine was running.

I cut the engine off where I did as I noticed that the cylinders were leaning out into the 19's AFR one by one at ECT's above 70-80C, and wasn't sure what to do, and was getting a few lean pops out of the exhaust again. 

Map and Log

[Rozsko]

Just out of curiosity, did you manage to bring her back to "normal" life?