Re-Bitten Hero

Thanks Rich

Hey Ian, what happened? S-reg Vectra? That's a bit of a comedown!

Looking back over posts, I suspect number 3 piston has gone pop... could be caused by a bad MAF sensor (though it was cleaned 6 months ago due to erratic control of boost, which cured that at least... hhmm...).

I'll get them to put a new MAF sensor in too, for definite now. Thanks for the warning on this John, Theo and others. Sometimes Scoobynet is worth its weight in gold!

But I miss my car

R

Ian Cook2

Rich, Its a long very boring story involving me having no money anymore ! LOL

johnfelstead

You had your MAF cleaned? Thats not something that is recomended on the MY99/00, it's too fragile and should be replaced if failing/contaminated.

EvilBevel

R, could it be that it was your boost control solenoid/pipework that was cleaned, not the MAF sensor ?

Hey IC ! Sorry to hear about your car too

Puff The Magic Wagon!

iTrader: (2)

Rich

We should compare stories when you come over

Bummer

Worthwhile getting a KnockLink too whilst you're at it. If it went in the same way as mine with a failed MAF causing det @ high revs, it *might* just have given enough warning to back off...

ian_sadler

Soz to hear bout the bang

When are you likley to find out what the cause was?

Re-Bitten Hero

John, Theo: no, I think it was the MAF sensor itself that was cleaned... maybe not such a wise idea then?

Ian: tell me all about it...

James: read about yours too How can I tell if it was the MAF sensor that failed? Or how can I get the garage to check it? Will there be an ECU fault code?


Got a fax from the dealer today - surprise surprise, new cylinder head needed Total cost including rebuild? A cool 4000 Euros (2800 quid)

Richard

EvilBevel

Richard, the short story is:

1) the MAF's on the MY99/00 are known to be very weak parts
2) the same MAF sensor is used on Golfs etc and same problem occurs
3) they sometimes fail instantly which is good, you get a CEL, engine cuts out, no damage
4) unfortunately they are known to degrade slowly, giving a leaner and leaner mixture until the engine gives up It still "gives a voltage" but the voltage is not correct (too low) for a given mass of air that goes past it. This can be due to oil vapour contamination or water that gets into the intake system. Sensors that were cleaned often fail very shortly afterwards.
5) the Subaru Select monitor is not the best tool to test a MAF, there are stories from people who had the MAF checked this way, it was deemed OK and they weren't
6) a better way to test the MAF is with DeltaDash operated by someone who knows what to look for. On WOT on a standard car, you should see about 4.2 to 4.3 volts
7) rough idle, especially after an ECU reset is a giveaway
8) a better way to avoid this in the future is like Puff says, install a knock detection device (like the Knocklink) and preferably also an AFR meter. Even on standard cars they are very useful as you can constantly monitor if everything is still working properly

There should be a "sticky" thread about this as people don't always read drivetrain, and it's such a shame engines blow because of this "Achilles Heel" of the Subaru, which costs a whopping 75 quid to replace (I replace mine now every year and consider it a consumable like oil or whipers)

My dealer estimates that 30% (at least) of the failed MY99/00 cars he sees/repairs have a failing MAF.

HTH

Theo

EvilBevel

Forgot: of course it's just "speculation" on my part, it could well be due to something entirely different. But I've seen too many stories of people who had a rebuild only to have the engine fail again on them after a few weeks

Re-Bitten Hero

Great stuff Theo, many thanks. Everything fits - I had noticed the temperatures were a bit high on my car, as I got gauges fitted about 3 months ago. But as they always read like this, I just thought the gauges were over-reading, on the grounds that "it's always been ok before so why are you panicking now?" The temps hit 255F, 124C when it went . Now I realise the gauges were right!! It'll be interesting to see what they read when the engine's sorted again.

One thing though - the water temperature gauge in the car never changed [img]images/smilies/mad.gif[/img]. What's the point in having this at all? Waste of space [img]images/smilies/mad.gif[/img].

Anyway a new MAF sensor will immediately go in. And I'm recommending to my MY99/00-owning friends that they do the uneven-idle check after an ECU reset, and maybe replace it too.

Replies much appreciated guys, thanks.

Richard

Scoobychick

iTrader: (1)

Good idea about the sticky thread Theo.

Richard, we'll deffo have to meet up, Cambridge meet sounds good, maybe we can persuade Dave TS to join us in his new toy? Perhaps we could even have a whipround for you to help towards the cost of your new cylinder head, although I don't think shirt buttons and leftover chocolate coins from Christmas are acceptable as hard currency in Germany are they?

Joking aside you know you have my full sympathy

Dave T-S

I'll be up for it providing i'm around

I'll bring the lil Golf along, and promise not to embarrass any Scoobs with it

Re-Bitten Hero

Ok, great stuff! Maybe we can do a Scooby meet on 30.12 or something? Otherwise maybe 03.01? Would have to be in the afternoon and I'd have to work out how to get there (no car in the UK), but maybe Sal could give me a lift back towards London . Would be great to catch up.

Maybe someone could do a Cambridge Meet thread? I'd like to keep this thread on topic, to collect wise words...

Thanks
Richard

Puff The Magic Wagon!

iTrader: (2)

Rich

I'll be around then - will be good to see you



Theo

Just speculating here, but it surely would be possible to run a voltage sensor across the MAF, reporting to an on-board screen, so that when you're giving it large, the voltage was displayed & if below a certain level, a warning light come on?

If you can detect it on a Delta Dash, then surely the same monitoring devices could be used to give a readout? Would it be something like boost/MAF voltage or rpm/MAF voltage or would MAF voltage on its own be enough?

Thoughts? Then if there's something that can be done cheaply by PCB or someone else, we can put one together. Components will surely only be a few pounds & if it can save a few thousand...

Scoobychick

iTrader: (1)

Richard yhm

Re-Bitten Hero

Sal: I got a notification to say you had replied to this thread, which told me I had mail

James: great idea... count me in on such a device... Also it'll be great to see what remarkable pyjamas you'll be sporting on a cold day

Cambridge meet thread here

Richard

[Edited by Re-Bitten Hero - 12/9/2003 4:37:29 PM]

Scoobychick

iTrader: (1)

Heeheehee

EvilBevel

James, yes that would certainly be possible, but it gets more complicated if you have modified cars with cone filters etc (they influence MAF reading, some very much so like the APS CAI)

Ultimately, the AFR meter already exists and gives the info about both the MAF (you can see very well if a MAF is failing with an AFR) and the rest of the inlet tract / mixture (leaks, blocked injector etc).

Boost isn't needed, the ECU does almost nothing with that info (apart from fuel cut safety), the ECU is relying almost purely on MAF / revs / throttle position sensor to calculate fuel injection (OK, simplification, but 90 % true, and still talking MY99/00 here)

But it can be done (but would probably only be useful at WOT, but then it matters most)

IF WOT and MAF voltage < 4.2v THEN light on/alarm. This voltage should be a variable and will have to be tailored to the car.

Anyone handy with PCB's ?

Also: this could already be present on the SECS ? (or should be easy to incorporate)

Just thinking out loud.

Puff The Magic Wagon!

iTrader: (2)

PCB (Simon) is SECS & thats what he does for a living, so I daresay he could knock up something to incorporate/not into SECS. I'll give him a call...

What is WOT btw? Having a non-thinking day

EvilBevel

Wide Open Throttle aka loud pedal mashed into floorpan

johnfelstead

Sorry Theo, but you are wrong on this one. You cant use an algorithm such as IF WOT and MAF voltage < 4.2v THEN light on/alarm. Think about it!

This is a trace of my car on WOT from 2000rpm to 7000rpm, the MAF voltage goes up in proportion to airflow (as you would hope!), at low rpm the amount of air flowing is obviously a lot lower than at 7000rpm, so its not posible to make a display just to look for failing MAF's unless you use all the important inputs such as rpm, air flow, throttle position etc, it would take some complex logic to be useful.

johnfelstead

It's also a bit of a bitch trying to use the lambda sensor fitted to the Impreza and a lambda link as an acurate guide for good AFR's on track. The longer you run, the more the lambda sensor heats up and therefor the lower the voltage it outputs for the same AFR.

A good example is at the nurburgring, at the start of the lap you may be reading 0.91V, 9 minutes later, even though the engine is running exactly the same AFR's, the sensor will be outputing 0.85V.

Doesnt sound a lot of diference, but thats the diference between running too rich and running too lean WOT. This is why the JECS ECU doesnt read lambda on full throttle and only uses it to control emisions closed loop at low rpm. To have an acurate on track AFR display you need an expensive wideband unfortunately. If your lambda link was to start displaying leaner than the top 2 greens then you would need to take a look, so it isnt totally useless on track, but you have to be aware of the behavour of these devices as the heat builds up and you will notice it drop a level as your runtime increases. On the road its a diferent story as you never really get the sensor "that" hot (unless you are a complete headcase).

EvilBevel

JF, agreed.

Was just thinking out loud as I said

But... if you look closely, you'll see that from the point of almost reaching max torque up until the redline, the MAF voltage should be within a certain band?

But I don't think I'm wrong in suggesting an AFR instead ? Safer and more to the point.

johnfelstead

look up. LOL

EvilBevel

Ah, the sneaky "in between" post

It's a case of diminishing returns John. Sure, the difference between 4.1 volts and 4.3 volts can be massive in AFR terms, but very simply put, if I go on track I don't worry if the AFR gives me 8%CO instead of 9 ... I *know* the sensor heats up and isn't very accurate.

But if it would go to 5% CO I *would* back off. I bet Puffs reading would have been even lower than that.

As a last resort... I did mention replacing MAF thingies every year, just in case. Will not catch 100%, but would you agree it would catch 90 % ?

johnfelstead

its got to help Theo. I beleive EcuTek find 2 out of 3 MY99/00's going for remap have a MAF that is degraded. Scary!!

Re-Bitten Hero

Very scary! Is it worth contacting Subaru about this? I mean, if that many are going, and the consequences are so catastrophic, shouldn't they at least order a check on the sensor at each car's next service?

Very interesting thread guys, thanks, more information to add to my Scooby information "database" (=sad techie brain ).

Now got a quote for the work and parts at 3200 Euros (2300 quid) so it's coming down a bit, plus I think I may be able to source a short engine from another guy... fingers crossed...

Cheers
Richard

johnfelstead

you cant check them at a service Richard, thats the problem.

you said the cylinder head was knackered, is that on top of a knackered bottom end? Short motors are just the bottom end less the heads incase you didnt know.

Re-Bitten Hero

Good point John. Trouble is, the fax containing all the details is in German .

It needs:

- part engine (cost 1000 quid)
- cambelt (80 quid)
- cambelt spanner (50 quid)
- cylinder head gasket, 2 of (30 quid each)

plus labour costs of 550 quid, all plus VAT of course.

R

johnfelstead

you need a lot more than that rich, i am a bit confused, are they replacing the entire bottom end with a short motor assembly?

Scan the documents and mail me if you like.