john banks

Looking at a troublesome STi5 this afternoon.

It has already puzzled two specialists and myself

I am not charging the guy for looking at his car and suggesting a few things he could get investigated, have his permission to post this, and we would both appreciate any input. Although he is interested in a remap at some point, I am not touching it with a bargepole at the moment.

WRX STi 5 Type R

Purchased by the owner with a Magnex backbox from a trader. Engine let go 6 days later at high speed. Engine not opened, but presumed to be a holed piston. Trader unhelpful.

Bottom end replaced with STi 4 block complete, attached to the STi 5 heads. VF24 turbo from donor engine used as status of VF28 on original engine unknown. All at owner's expense.

Subsequently a boost gauge was fitted and the car was overboosting - cleaning out the solenoid and pipework helped. In the last week the car was decatted with a Magnex twin dump downpipe and Blitz NurSpec.

Problems:

1. Boost creep from 5000 RPM upwards in high gears since the decat. Before the decat the car was holding about 16 PSI and was probably dropping off slowly at the top end as expected.

No difference with a Dawes as tried by one specialist, which kind of rules out solenoid or restrictor issues.
I released the circlip on the wastegate actuator arm and found 90 degrees free rotation of the wastegate flap.

2. Oil usage - goes from max to min every week

3. Goes really lean from 5500 RPM upwards in high gears as the boost creeps.

Trying to tie all this together, wondered if the oily solenoid and initial overboost could be caused by excessive blow by and recirculation of oil mist into the intake.

Never heard of a VF24 boost creeping before and with the free movement of the wastegate flap is puzzling.

Could the wastegate actuator be sticking and gummed up with oil and the problem is marginal in all gears but really bad in the higher gears? The ECU is seeing the correct boost which climbs to about 1.4 bar at 6000 RPM, from 5000-6000 RPM the wastegate duty cycle reduces to zero. So does the lambda voltage Didn't log the injector duty cycle because only looking at the logs retrospectively have noticed it leaning out at high revs (in addition boost climbs and knock correction goes to zero presumably because the ECU is no longer listening).

I have advised the owner to get the actuator swapped to see if that is the cause of the boost creep, if not refit the catted downpipe, but I am puzzled if it is not the actuator.

I recommended that the fuel pressure be checked, and if still lean to check the injectors. It is not beyond possibility that one is marginal causing the original engine to let go?

The MAF sensor, lambda sensor and knock sensor etc all seemed to be working fine and there were no fault codes.

The combination of boost creep, being lean and no knock correction at the top end seems a recipe for almost certain imminent failure, so he is driving it gently in the meantime.

Any suggestions welcomed. I'll post some graphs of the key bits of the Delta Dash logs soon.

David_Wallis

Time for a compression test.

What are the breather pipes like? are they covered in oil?

Oil seals in the turbo ok?

David

Thats a lot of oil to loose.. its going somewhere...

David

T-uk

total guess so check with mark aigin.

I am sure he said sti4 and 5 head gaskets are different.could this have anything to do with it,if sti 4's have been used?

john banks

Forgot to say the speed sensor input to the ECU seems to be correct in kph and then seems to softly clip at about 136 kph which I presume is the speed limiter removal box.

The logs are all the same top end section of 4th gear.

john banks

**REALLY** high airflow voltage


Note absolute boost and the revs - STI target is 15.7 PSI until about 5400 RPM then 14.6 PSI to the red line!


ECU is trying to keep the boost down with zero duty cycle - looks like the VF30 guys experience!


Lean as ****?



No knock correction... I bet it needs some.

[Edited by john banks - 1/16/2003 5:25:19 PM]

john banks

Thanks for the suggestions guys. Good point T-uk. I'm sure they are different. Maybe his CR is lower which is protecting him from the det - I can't understand why the engine is still together given the logs. It is a new lambda sensor BTW so I don't think it is just tiring at high revs as it recovers instantly when you back off. Wonder if the old one got fried by high EGTs?

AlanG

slow internet tonight then....

[Edited by AlanG - 1/16/2003 5:54:16 PM]

AlanG

If a lambda sensor fails or is failing, will the ECU not compensate for this until such times as it is renewed? i.e. fixed map

VF24 against VF28. Assuming a std ECU in the car, I would expect the boost to react differently, due to different components of the turbo's.

A

AlanG

Oil usage...
It's either being consumed by the engine or falling on the road.

No evidence of it being used out the exhaust?
Did you check the plugs for signs?
Check the intercooler pipe after the turbo?

john banks

http://www.mrtrally.com.au/performan...turbo-data.pdf

According to this the VF28 is basically the same as the 24, ie P18 same turbine, same compressor wheel and cover, same bearing. Even if not it's a heck of a boost creep.

The lambda sensor seemed to be behaving fine closed loop, and open loop.... until it went crashing down at high RPM. So I am inclined to believe it was actually lean.

Didn't have time to pull a plug out and see, exhaust looked fine. No oil on the owner's drive. Didn't have a full set of tools with me. Who can this guy go to in Scotland as he has tried the two main specialists already and been sold a Dawes and the other place wanted to drill his restrictor which is clearly not the problem.

[Edited by john banks - 1/16/2003 6:02:39 PM]

Andy.F

How much travel does the wastegate have with the actuator still attached ? (it's tight but you can move it with pliers)
If the actuators have been mixed/matched or adjusted it may be running out of travel. This will cause boost creep.

john banks

About 40 degrees Andy, same as mine (well T-uk's ), we compared them side by side. The actuator felt about the same tension, and was non-adjustable.

The oil loss is one issue, the boost control is puzzling.

[Edited by john banks - 1/16/2003 7:17:19 PM]

jameswrx

iTrader: (4)

As i've said a few times before, when I put my full hks on inc d/p I've been getting boost creep too. Exactly the same as you describe, checked wastegate movement, actuator changed, pipes changed etc, but it all points at the decat.

I pointed the blame for the boost creep on the HKS d/p which is a single pipe from the turbo (ie same bore as the rest of the exhaust) I thought this to be the prob as the d/p sits directly behind the exhaust compressor turbine and the plate the d/p bolts to the turbo with is in front of the wastegate effectively blocking a direct exit for the wastegate gases in comparrison to the turbine flowed gases which have a better less 'restrictive' route to flow the exhaust. This combined with the extra flow the hiper system provides is IMO enabling the turbine to gain speed with engine speed and gas flow thus producing the boost creep.

I know you have a twin dump, but I reckon this is gonna happen to the twin dump too because there is gonna be a different gas speed in the two parts of the d/p and ultimately the exhaust turbine is going to flow better as it has a less restrictive exit(pipe size).

I would say an open neck d/p is the only way your going to get a good mix of the same exhaust gases (via turbine vs via wastegate) and ofcourse a more restrictive exhaust is needed.

I've said this before and am sure there are loads of people that are getting this and don't even know about it.

I'm still investigating.

James

UkLegacyT

John,

my actuator went dodgy on my TD05, i set my apexi boost controller to 1bar, but at 5000rpm, the boost just shot to 1.5bar or more. swapped the actuator and it was fine, 1 bar throughout in any gear.

just a note too:
sti5 headgaskets are thinner than others, and also the pistons are different to others, iirc its the deck height.

ian

john banks

Thanks guys.

Andrew used an STi 4 headgasket, so the CR should be right.

James, I have used a VF28 based hybrid on the same type of downpipe and there was no issue with boost creep even with a 1.2 bar actuator but it had a ported wastegate

Anyone got a standard VF24 or VF28 with a Magnex twin dump downpipe and not had boost creep?

jameswrx

iTrader: (4)

any ideas about mine then John (94 wrx, td05, standard apart from HKS Hiper, HKS d/p, HKS induction)? car was standard until I got my hands on it and had this trouble

is it a common problem with the turbo (ie needs porting?), or do you think it could be exhaust? It would seem strange to think the wastegate couldn't cope with flowing such a low level, ie not allowing 0.80bar to be held.

My scenario is

WOT in 3rd at approx 4,000-4,500rpm car achieves full boost (0.85bar) then holds this til after 5,000rpm where it begins to climb upto around 1 bar at the rev limit, sometimes hits fuel cut just before rev limit (dependent on load or air temp)

Tried all obvious checks, boost gauge, actuator, wastegate(movement&blockage), different pipework, etc

P.S sorry to hijack your thread mate

James

john banks

Do other downpipes make a difference? Might be worth trying another. Same for Andrew's STi I guess and then he can tackle the oil usage problem. I was alarmed why it went so lean at high revs, but that seems a separate issue, a standard ECU should provide plenty of fuel beyond the boost this one went lean, so I guess it must be fuel pump/injectors/FPR etc.

[Edited by john banks - 1/16/2003 9:40:12 PM]

jameswrx

iTrader: (4)

if it's not lean anywhere else (ie as boost comes in@WOT) then fuel pump and or fpr seems more viable than injectors (if injectors blocked and or faulty you'd see a problem throught the run?) if it's not an electronic control problem

you say the boost creep appeared since decat (like me) therefore I think he should get some more backpressure or different d/p(if this is viable)

john banks

Think you are right on both counts. Try another downpipe and check the fuel pressure.

T-uk

could there be a lambda fault for reading with delta dash?.

the way your lugs are tuned in these days JB,I think you would hear the det.

[Edited by T-uk - 1/16/2003 10:46:30 PM]

john banks

Lambda sensor was fine and as expected until the upper midrange. It is a new Subaru lambda sensor on the car too.

With a Blitz Nur Spec it is quite difficult to hear det

The car wasn't going with the sort of conviction you expect at 1.3 bar at 6000 RPM either. It felt weak, but this is probably about or near to the most airflow you'd get out of a VF24. If the lambda sensor is to be believed it was the lean side of 14.7:1 which is seriously lean at that sort of boost level. Peak power is typically at 12.6:1, I run deep into the 11s. Andy F was running in the 12s at his leanest except for the odd test I gather. Leaner than that and you lose power.

[Edited by john banks - 1/16/2003 11:08:26 PM]

john banks

Just as a follow up - the fuel pressure was dropping at high RPM (substantially - about half of what it should have been) - a new pump fixed that. Taking the decat downpipe off has resulted in about 0.9 bar boost which is a bit low, but the car is happier now and adding timing (1.5 deg positive knock correction from Falkland's logs today). He is going to refit the VF28 and see what oil usage follows that.

Not surprising it blew up at sustained high RPM - I think the last engine let go because of a faulty fuel pump and this one was going the same way - it was not changed or checked.

AJbaseBloke

John,

That reminds me of something I saw at my dealer last year (not relevant in this case obviously): they had a Forester STi in that had let go at high speed - saw the very sad remains when I took my car in for a major service. The Forseter was run hard all its life, oil changed every 3500kms, no worries for 52,000kms, then went kaboom after "a bit of hesitation at expressway speeds" (according to the owner ).

Had all the mechs scratching heads as there were no codes, ECU checked out fine, all the sensors ran fine when bench tested...no blocked injectors either. Then they pulled the pump.

White stuff/gunk coated everything in the tank and had blocked the filter, leading to a pretty much instant lean out and the rest was history. Turned out to be contaminated fuel