simonds1

My impreza has developed a problem which is driving me mad, especially because it's an intermittent problem so it doesn't happen all the time. I'll give you an example of what's happening.....

Lastnight on my way home, came off a roundabout in 3rd, foot to the floor and car behaved exactly as it should, took off like a legend. This surprised me cuz it'd been miss-behaving all the time lately, so next roundabout I got to I did the exact same thing - foot to the floor in 3rd, this time....absolutely nothing (no turbo spool up sound, no boost, seemed like no reaction to my throttle press at all) for a good 3 or 4 seconds, and then suddenly took off like normal. So next roundabout again, I did the same thing.....this time there was the same 3 or 4 second delay of nothing happening, but instead of taking off this time it started to stutter & jerk quite harshly, like as if the fuel was cutting out sorta thing. Can just never tell how it's gonna react like. Also when I'm just driving around normally it just generally feels sluggish. I seem to have to press the throttle much more to get any reaction out of it. Normally when I'm going along on the motorway, a little tiny touch of the throttle and it will just pick up beautifully and start accelerating, but now I seem to have to press it a lot more to get it to do anything.

I've had quite a few suggestions from people so far - MAF, air leak somewhere, wastegate sticking, spark plugs/leads and fuel pump also.

Your opinions on this will be massively appreciated. Thanks in advance for your help guys!

Maldge

Mine had the same symptoms (plus engine check light on). Changed the MAF and it's running fine now.

simonds1

That's interesting, the MAF seems to be the most popular suggestion among people. I'm not getting a check engine light at all though?

Maldge

Yeah, it's definatley a common fault with the Classics. I know that intermittant problems like you have described happen generally when there is a failing part. it might be worth trying the following which was advised to me on here by Splitpin:

Before going any further find the pair of black and pair of green plugs under the dash. These may be taped up around, or hanging loose next to the wiring loom running into the interior fuse box behind the coin tray. Sit next to the open door and crane your head up under the lower dash cowl and you should see them.

Once located connect the black plugs together, switch the ignition on and watch the flash sequence on the check engine light closely. Long flashes are tens and short flashes are ones, so two long flashes and three short = 23, three long and two short = 32 etc.

I think that code 23 relates to the MAF. Post back if you get a code and someone may be able to help.

simonds1

Awesome, thanks very much for that. I'll go and try it in a sec and then report back

simonds1

Okay I've just tried connecting the two black wires together now and there is no check engine light flashing at all. Y'know when you normally turn the ignition on , the oil, handbrake, battery & check engine lights all come on until you start it? Well that whole section was blank, no lights at all when I had the black wires connected. Is that normal?

I thought I'd put this pic up just to check I've plugged the right wires together!......



So of course, if I did do that right, and I haven't got any fault codes........what do you reckon I should look at next?

Splitpin

That's not remotely normal Simon. Certainly does look like you've connected the correct plugs. As a matter of interest, try connecting the green pair together and see what happens when you turn the ignition on. You should get the fuel pump, rad fans, and all the relays under the bonnet clicking on and off every few seconds. If you don't, let us know.

However, all of that may be moot as far as solving your problem is concerned. This might be the result of a dodgy air flow meter, but it also might well be something else. A degraded coil pack could also cause these sorts of symptoms for example, as could any of the range that has already been mentioned to you.

If you haven't actually seen the Check Engine light while the engine's been running, the black plugs won't help you diagnose as there will be no relevant trouble codes stored. In many cases some types of failure (e.g. coilpack or MAF) won't actually trigger the CEL at all.

Would help us help you if we knew what type and model year your car is, and also what its state of tune/modifications (if any) are.

winker

iTrader: (1)

Hi i have just had a problem with my classic might be the same i would drive low speed then try and pull away bad hesitation not all the time i agree with the other guys could be maf but with no light on it could be overfuelling i would give mine a good run and then it wasnt to bad but as soon as i was doing normal speed the hesitation would come back is the exhaust very sooty anyway i changed the lamba and the car now runs perfect.

simonds1

Right guys, just had a go at connecting the two green plugs together, and yes that works - lots of things clicking every few seconds, fans coming on & going off etc. I tried the two black connectors again straight after out of curiosity and I got it to do something this time. It gave me two long flashes, follwed by three short, so that's a 23, which is the MAF. BUT......here's the thing......when I was trying to see what might be wrong yesterday, I unplugged the MAF and tried starting the car. This put the CEL on, so the code that's stored is probably from that? Because like Splitpin said, the CEL hasn't come on while I've been driving so there would be no error stored.

Also, I didn't post info about my car because it's in my car info bit on here. Sorry I should've made that more clearer. I'll copy & paste the info here anyway to save you all going & clicking on it......


MY00 Subaru Impreza Turbo 2000

Engine/Drivetrain Modifications:
- Ninja sports cat downpipe
- Afterburner decat centre section
- Hayward & Scott custom made backbox (replica of the standard twin tailpipe one)
- Walbro Motorsport 255lph fuel pump
- NGK PFR7B Laser Platinum spark plugs
- Samco SM-TCS143 Blue Silicone Intercooler Hose Kit
- Sti panel filter
- Bob Rawle EkuTek3 Remap 290bhp / 329lbft

Was last serviced 3k miles ago and it was a big one so new cambelt, plugs, panel filter, pretty much everything! Just incase that makes any difference.

Cheers

Brun

The CEL will only show a "failed" sensor, not a failing sensor. As a sensor is failing, the ecu often compensates. If the maf is suspect then don't be nailing it everywhere as your engine could literally be at risk.
I would change the maf regardless as the MY99/00 maf in most cases, fail sooner or later

simonds1

I was wondering if the CEL would come on or not if it was a sensor that was on it's way out rather than completely dead, so thanks for clearing that up! One question though.......coulda faulty MAF actually cause fuel cut? Or would that part of the problem have to be down to something else?

Splitpin

Sometimes a failing airflow meter will trigger the CEL, more often than not, it won't until very seriously degraded. The actual criteria the 99/00 ECUs follow for activating the 23 error condition is first that the engine must be spinning at greater than 400rpm (i.e. running, in practice). Then if the sensor signal drops below 0.3 volts for more than a fixed period of time (if I remember right, 0.2 seconds), the error bit is set.

So this means that a sensor that persistently under-reads (for example as it might if contaminated or suffering FOD) will not trigger the CEL.

Similarly if a sensor follows the most common failure mode, and begins to drop out during conditions of high airflow, the voltage might not actually drop into the the 0.3volt/0.2sec trigger margin. Again, no CEL - at least until the sensor is considerably further degraded.

It is possible to alter the trigger criteria so the ECU will pick up a pending failure earlier, but whether EcuTeK or any of their mappers have twigged that yet is a question I couldn't say.

Not directly. There's only one thing that activates the actual boost fuel cut routine in the ECU and that's boost going outside the programmed maximum for whatever atmospheric pressure is prevailing. However, a sensor signal dropout can cause symptoms very similar to cut, and a dropout long enough to trigger the CEL (moot point here of course, but..) will put the ECU into limp mode.

Incidentally, the ECUs ability to compensate round a failing airflow meter is limited, and rather fundamentally compromised. It basically can't once you are into medium to high boost (i.e. open loop fuelling) mode.

It might be down to something else, but in the absence of any diagnostic kit you're probably going to have to run a process of elimination. Oh, incidentally...

Your surmise there is probably correct. If you've started the engine without the airflow meter connected, you will, as per the criteria above, have triggered the error.

Unfortunately the symptoms you are reporting are not consistent with the classic airflow sensor degerdation. It might be all, or part, of the explanation, but might be nothing to do with it at all, so the question at the moment is whether you start speculatively replacing stuff in the hope of fixing the problem or get some first-hand expert help.

If you don't have easy help on hand, the possible way to look at it is to ask how old your current airflow meter is. If it's more than about three years old (and almost definitely if it's one of the older style "green spot" or no-spot parts, it's probably worth a speculative replacement. However I would not be totally surprised if, in this circumstance, it didn't represent a total cure.

simonds1

Thanks for that detailed response Splitpin! It's good to be talking to people who know what they're on about, and hopefully it'll help me get to the bottom of this somehow

In regards to how old my MAF is..... I bought the car back in November 2006 and it was completely standard. Had no problems at all with the MAF, only issue I did have was failing MOT on emissions, which turned out to be a faulty lambda sensor. That was in Sept 2008. So I had that replaced and all was fine. I then had the car remapped by Bob Rawle in March 2009 and he said if my MAF wasn't one with a green dot on it, it needed to be. So at the time of remapping he fitted a new MAF with a green dot on it. So that green dot one has been on the car just over a year now. I don't know what the importance of that green dot is? Does a standard MAF somehow not work with a remapped ECU? I'm guessing now if I get a replacement, it will have to be a 'green dot' one? Another thing that might be of importance, when Bob was plugging his equipment into my engine in various places, he commented on how cheap & nasty my lamba sensor was and said it probably wouldn't last long because it was a cheap one not designed for a turbo car etc. That's why I was asking about the lambda sensor yesterday, because I'd remembered he'd said that. Car passed it's last MOT first time tho, so I'm guessing the lambda was still working then. That was Sept 2009. Could a faulty lambda sensor cause fuel cut?

My mate has got a fault code reader and has offered for me to go over and use it later tonight. Is there any point in me doing that or will it only show the 23 error code that I've already seen on the car? I don't know what fault code reader it is.

First person I talked to about this problem was Bob Rawle and he suggested I get the fuel pump checked out. I have got a local garage that I trust and he knows quite a lot about performance cars. He used to race-prepare cars and currently has a 550bhp Sierra Cosworth that he races. So I'm guessing he's got a good understanding of turbos etc, altho he's not really an impreza expert specifically. I will get it booked in to check the fuel pump, which incidentally, he fitted. And then I suppose we go from there.

Thanks again for all your help so far

Splitpin

That's not quite the correct reason, no. The "green dot" sensor is a standard one. So is the "green stripe" and so is the "no green anywhere" sort. Subaru have modified the design of the sensor since first introducing it, presumably to increase its reliability and robustness. The "no spot" was the first version that came with the cars in 99-00. Then there's one with a dot of green paint on the sensor head, and then there's the (I believe still current) one, which has a stripe/splurge of green paint on the mounting flange, along with manufacturer part numbers and a date code printed in white ink around the other faces of the flange. This version of the sensor was introduced somewhere around 2004-2005.

In practice any replacement bought from a Subaru dealer tomorrow should be the green stripe one (unless they've changed their marking procedures again recently). However this is a 100% compatible replacement for the dot version, if this is indeed what you have.

Can you check for us (if you don't know already) whether you have the green spot (on the head) or green stripe (+ white ink) on the mounting flange sort?

Almost certainly not. Once your car's on boost it runs in open loop fuelling mode, which basically means the lambda sensor has ceased to exist as far as the engine management is concerned.

If your mate's code reader is a generic OBD2/EOBD/etc sort there's no point him even plugging it in as it won't work on your car. 99-00MY Imprezas have the 16 pin DLC connector under the dash but are not actually OBD2 compliant - they use the proprietary Select Monitor serial communication protocol so trying to connect an OBD2 interface is a bit like someone who only speaks French trying to talk to someone who only speaks Japanese. They both have ears and mouths so each can hear the other, but neither has a clue what the other is on about (at least until the frog waves a white flag and surrenders).

You can diagnose your car very effectively if you have a laptop, a cheap ISO9141/vag.com compatible USB interface which you can get for less than £20, and some free software like ECUExplorer or EvoScan. DeltaDash is better again but isn't free.

This will allow you to examine the various sensor inputs into the ECU in real time, which, with a bit of advice on what to expect should soon clue you in on what's normal and what isn't. All this is assuming, of course, that EcuTeK haven't locked out or somehow altered the select monitor communications routines in their mapped ECUs as a "security measure".

The fuel pump is a possibility but again your symptoms don't fit neatly into this expected mode of failure. It's possible that you have a couple of different issues interacting with each other here though so it's worth following any potential leads. Actually diagnosing the thing will probably save you a lot more than twenty quid for an interface, provided you've already got the laptop.

simonds1

Okay I have just discovered that I think I was wrong by saying I've got the green dot type MAF, because I think this looks more like a stripe to me.......



I wasn't aware there were two types of green marking before you said. Also, it's got a number written on one edge that you can see in the pic is quite worn off. I can just about read it and it looks like it says "C36-700 R92". So out of the 'no green marking anywhere', 'green dot' and 'green stripe' MAF sensors, which is the one that's notorious for failing? Or is it all 3 types?

I'm not sure about my mate's fault code reader. All I do know is that he's got a 51 plate BMW M3 and will have bought it for that. I text him asking if it'd work on my car and he said yes, but he might not know about the imprezas not being OBD2 compliant if that makes any difference. I don't know if the M3 is or not. To be honest I don't know anything at all about this diagnostic lark! I just had a quick glance for the sort of lead I need to buy, and I found this........

http://cgi.ebay.co.uk/Diagnostic-Int...412138001r2155

But I'm guessing that isn't the correct type because it says it's "OBD2 to USB...etc..." and you said the imprezas aren't OBD2 compliant? If you could point me in the right direction as to what lead I need to buy that would be great And fortunately I already do have a laptop yes. It's running Windows 7 though, and I don't know will that cause problems for this sort of interface? Cuz that one I just found on ebay says at the bottom that it may not work correctly with windows vista. But then vista is one of man's worst creations ever!

Splitpin

Yep, that's a green stripe one. See where that "C36...." text is on the mounting flange? If you look on the lower side of that flange you should see another four figure number - use a mirror if necessary. That number is a manufacturing date code.

Notorious might be a bit too strong a word. The original no spot is the least reliable, the green spot's a bit better and the green stripe one is, IMO, pretty good.

Yes, that's the point. I'd put at least 50p on his not working on yours.

It looked good up until the point the listing says it's based on the ELM323 chip. These are no good for the Subaru SSM mode.

Have a look at the KKL interface here. This will work fine with your car and is a tenner cheaper than that eBay one. However I have no idea whether this interface (or ECUExplorer/Evoscan) are Win7 compatible so it might be worth an email to Gendan to ask about driver support, and a snoop around the appropriate forums to find out about compatibility.

simonds1

I spoke to my mate and he agreed that his fault code reader wouldn't work on my car, so I didn't bother going over in the end. I have ordered that KKL interface lead, and also downloaded the 'ECUexplorer' software ready, which it says works with all versions of windows. So I guess I'll just try and connect the car to my laptop using that lead when it arrives, and then go out for a spin and see what it comes up with.

Thanks again for all your help so far, and I'll report back when the lead arrives

simonds1

Just to update....

The cable arrived today and I installed ECUexplorer on my laptop etc, got the data logging working, went out for a spin......and do you think the bloody car would do anything wrong!?! Course it wouldn't! Typical innit.

Anyway, hopefully I will catch it faultering soon and then I'll report back on here.

Cheers guys

Splitpin

The watched pot never boils!

ECU Explorer isn't particularly clever/fast when it comes to logging but if you capture engine speed, throttle position, MAF voltage, O2 sensor voltage, ignition timing, injector pulse width and wastegate duty as it's acting up you'll probably have enough to work out what's what.

Kosy

iTrader: (2)

Interesting read. I have just bought a P1, can you tell me where the maf is located so I can check mine?

Thanks

simonds1

I've highlighted it here.......



Hope that helps

Kosy

iTrader: (2)

Thanks for that

simonds1

Right guys, finally here's an update! ......

I think I might be actually be on to something here (but probably not!). I've embedded two images below, which are sections of a log file I just created from ECUexplorer when I went out for a drive. By the way, this is with the lambda sensor disconnected - I unplugged it lastnight and also disconnected the battery for about 10 - 15 mins as advised by Bob Rawle, just to see if there was any difference in the car.. I'll come back to that in a minute though.

If you look at this section of the log, which is when the non-accelerating problem occured......



.....you'll see that at 13:25.2 I put my foot flat down. Throttle sensor voltage then went up to 4.28v and stayed there for about 5 seconds while I had my foot down. When my foot was flat down, the fuel injector pulse width was hovering around 7.17 for around 2 seconds and then it starts to rise very gradually, and then suddenly jumps to 15.87.......five and a half seconds after I originally put my foot down! My boost was also rising very gradually from 2.757 - 8.125psi (this took 5 seconds, bearing in mind my foot was flat to the floor).......then suddenly jumped up to 16.396psi, at the same time as my fuel injector pulse suddenly jumped up aswell. Literally as soon as the car felt like it was going to start accelerating properly, I had to take my foot off the throttle because there was somebody infront of me, so that's why the sudden increase in boost & fuel etc doesn't last at all. So this was 3rd gear, foot to the floor, and it took 5.5 seconds for my revs to go from 2,688rpm to 4,845rpm. That is abnormally slow for a 290bhp impreza don't you think!??

Now if you look at this section of the log, which is when the car seemed to accelerate very well........



......you'll see that at 07:18.8 I put my foot flat down. One thing I noticed here is that in the various places in all of my logs where the car didn't accelerate properly, the throttle sensor voltage went up to 4.28v when my foot was flat down........but in the places in the logs where the car did accelerate properly, it only went up to 3.76v with my foot flat down. Is there a reason for that do you think? So you can see as soon as I put my foot down at 07:18.8, boost goes up instantly, as does the fuel injector pulse. The car just accelerated instantly like it should. It holds 18.427psi of boost for a tiny bit then before I release the throttle. Every log I've done so far, this is the maximum psi reading I get - 18.427 everytime.

I'm reading this information as the 'fuel injector pulse width' being something to do with the amount of fuel being delivered? So the higher this number, the more fuel? If that's not the case then my theory is a load of rubbish! But if you look at the first log when the car suffered from the problem, the fuel injector pulse is only 7.17 when the throttle sensor is at 4.28v, and then in the second log when the car was okay, the fuel injector pulse is 17.41 when the throttle sensor is at 3.74v. So, less throttle but more than double the fuel??

Going back to the lambda sensor being disconnected - First thing I noticed on the first drive was that the car just wanted to rev high. It was idling at about 2,000rpm or higher and everytime I stopped at a junction etc it did the same. Actually driving the car doesn't seem any different to when the lambda was connected tho. Today it seems better from an idling point of view - doesn't seem to want to rev high anymore. It just seems completely normal now. But it hasn't affected the problem I'm having at all because it's still happening now exactly the same with it disconnected. I'm waiting to hear back from Bob to see what his thoughts are on that.

Thanks in advance for your help guys, I will be very interested to see if any of you can decide anything from this data, because if I'm being honest, it doesn't make any sense to me at all!

Cheers

Splitpin

Some interesting logs Simonds1, although one of the difficulties trying to diagnose them is that your car has a custom map on it, and thus none of us (apart from Bob) knows exactly what targets it's working to, and therefore it's impossible to know for sure, in all cases, what looks right and what doesn't. However...

The first observation to make based on these two logs is that the 4.28 volt TPS output is slap bang within the normal range for wide open throttle state. So's the 0.48 volt idle position.

Yes, that's very slow boost build - especially given we can see the wastegate duty is up around 81%, pretty much the same as it is in the second "fast" log. One question/clarification: Is it correct to believe you have the standard two port wastegate control inc 1.1mm restrictor, and your turbo is a TD04 with standard actuator, or...?

Also, the one practical limitation you're working in here is the refresh rate of the log, which seems to be over 0.4 of a second. In practice this means that some of the transient changes in input may be missed between samples.

As a matter of interest how much data are you actually logging? The less you get the faster the refresh rate so if you're problem solving, only monitor what you need to. If you're monitoring the binary switches (in the bottom half of the ECU Explorer screen), uncheck them, as you're better off having a faster refresh right now than knowing what those things are doing - most won't help you here anyway. Similarly you don't need the battery voltage, A/F correction (or learning) or necessarily ISCV position or coolant temp while troubleshooting this one so uncheck them too. You should end up with a refresh rate below 150ms which is far more useful.

Yes very much so. The other interesting/odd aspect of that is that the sudden increase in injector pulse width and MRP arrive after you've lifted off the throttle - the row you see them in, the TPS is down to 0.5v on its way to idle. That's unusual, to say the least. In addition there doesn't seem to be a huge jump in airflow sensor voltage of the sort necessary to satisfy that apparent doubling of manifold relative pressure, which is a bit perplexing, although, as above, this could have been missed "between" samples.

There's definitely a reason! Although the interesting aspect of this is that it's the 3.7-odd volt TPS reading that immediately sticks out like a sore thumb, If your foot was flat down. As above, 4.3 volts (or close to) is the correct TPS reading for WOT. If it's only reading 3.7, the ECU will think you're only on part (somewhere around 80%) throttle and as a result, depending on how Bob's mapped it, might be looking up different (i.e. incorrect) cells in a number of throttle-indexed maps, as well as changing execution path in a few areas.

This initially points to a suspect faulty throttle position sensor, although there may be other things in play.

Yes there's a reason for that. This is the maximum number the select monitor eight bit MRP parameter can send. You will find that your actual boost level will be higher than this. Explorer allows you to log a calculated MRP value derived from subtracting the atmospheric pressure from MAP. This calculated value doesn't suffer from the same upper limit and is just as accurate (the ECU internally does the same calculation to form the MRP data parameter in the first place). Select the calculated MRP and stop logging the "main" MRP parameter and see what happens to your boost readings next time.

Totally. The injector pulse width is literally the amount of time the injector for whichever cylinder is intaking is switched on. The longer it's spraying fuel, the more fuel you get.

However, as engine speed increases, the time window for fuel injection obviously gets smaller - which is how injector duty cycle is calculated. This (think you can display it via ECU Explorer's options) is the percentage of the intake cycle time the injector is switched on, and is a more useful indication of how hard the injectors are working.

You can convert pulse width to duty cycle by the following formula:

IDC%=((rpm/100) x PWms)/12. So, to use your numbers, in your first log, at 13:29.9 you have an RPM of 4462 and a pulse width of 10.75ms. This gives a duty of only 39%, which is far lower than you'd expect for your car when accelerating theoretically flat out.

The injector pulse width is calculated by the ECU on the basis of engine speed and the mass airflow sensor reading, the throttle position isn't used as a direct input into fuel calc, although of course, the wider you open the throttle, the more air the engine will suck in, the higher the MAF reading will be, and so the more fuel you'll get.

As above I think that 3.74 volt throttle sensor reading is an indicator of a (or possibly the) problem. In practice if you know you had your foot right down it must be a malread. The key differentiator to make is when you look at the injector duty cycle. 3704rpm and 17.41ms pulse equates to 53% duty cycle which is much higher than your first log and more what one would expect.

That "unusual" low throttle sensor reading in the second log sticks out like a sore thumb, albeit that in that second log you say the car accelerated properly.

It'd be informative to get Bob's perspective on these numbers as only he knows how it's actually mapped.

In the meantime I'd try a couple of diagnostic tests.

First of all, use ECUExplorer to test the throttle position sensor. Turn off all the logged parameters apart from TPS voltage. This will give you a very fast refresh rate. Then, with the engine off and the logger running, slowly press the throttle all the way down, hold it there for a couple of seconds and then slowly release it. Repeat this process a number of times, slightly increasing the speed at which you depress and release the pedal until you're at a full on stab/sidestep to release. Then analyse the log carefully (ideally graph it).

What you are looking for is a nice smooth rise in TPS voltage as the throttle is being applied, a consistent hold at WOT every time (always in the same 4.28-odd volt range), and a nice smooth consistent decline as you back off back to neutral (which might, with the engine off, settle at a fractionally higher number than the 0.48 you see with it running).

For example, if you looked at the log, you'd want to see (on throttle press) the numbers rising from the idle position like this:

1.50
1.54
1.58
1.62
1.64... and so on up to 4.28 volts. I'm sure you get the picture. A smooth upward (or downward) trend without any significant deviations.

What you don't want to see is something like

1.50
1.54
1.04
1.62
1.64...etc.

You also don't want to see something like

1.50
1.54
2.02
1.66
1.64...etc.

Any significant jaggies/sawteeth in your graph like that indicate drop-outs in the sensor signal. Equally if it starts reading 3.7-odd volts at WOT then you've likely got a worn sensor. Let us know what you find and post up the logs (or a graph made from them) if you're confused still further.

Also, as above, next time you log it, cut out any data parameters that you don't actually need so you can concentrate on what you do want to know. Switch to the calculated MRP and inj DC display parameters too.

Oh, I'd also reconnect the O2 sensor an do us an acceleration log with it on. However, before you do, you know you said that without it, the engine wants to rev, even at idle? Can you post us up a short log of the engine just idling like that, monitoring the same data parameters as you've got in the two above, with the sensor disconnected, and then another idle one with the O2 sensor plugged back in? Make sure the engine's fully up to temp before you do these two.

I'd also check carefully for signs of any air leaks (loose/missing pipes, bungs etc) in the inlet tract between the MAF sensor and the cylinder heads.

joz8968

iTrader: (13)

"And the No bull piece Prize goes tooooo... Splitpin!"

Seriously, you should charge for your services to S'Net/the general Subaru community...

simonds1

Splitpin, you are my hero! That is one of the most in-depth replies I have ever seen on a forum, and I am honoured to be on the receiving end of it!

I have sent the log files to Bob and am waiting to hear back from him. I've been in touch with him since the problem started so he knows the full story etc. Will be interesting to see what he says. I'll post an update on here once I've heard from him.

Well that's good to know. I still don't know why the 2nd log from my previous post only shows a throttle sensor voltage of 3.76v when I thought I had my foot flat down. I can't be 100% sure to be honest, so I think the actual reason is more than likely my foot was not fully down. According to the log it's only at that value for just under a second anyway. I think I had intended to put my foot flat down, but must've backed out at the very last second because it actually started to accelerate properly and somebody was infront of me.

To my knowledge everything to do with the turbo & wastegate is completely standard yes. I know 100% that it's a TD04 turbo, but the "two port wastegate control inc 1.1mm restrictor"......I'm not sure? More than likely yes because I don't see why any previous owners would've done anything with that. I certainly haven't.

Unfortunately on the log that the two extracts in my previous post were taken from, I was logging all of the analogue and digital data. The refresh rate was somewhere up at 409ms. Only now I've read your last post I understand all this. I didn't know what that 'ms' value was before, but now I understand. When I de-select all of the digital values it goes down to around 196ms, so I know that now for future logging.

I've just tried this now. The refresh rate with everything turned off was about 64ms. Here are two log files I made......

http://mbf.me/8aft - This is my first attempt, and any 'jaggyness' could well be down to my inability to smoothly apply and release the throttle. It's actually quite difficult to do smoothly!

http://mbf.me/98ec - This is my second attempt and should be a bit smoother

They are just mailbigfile links, so they will only be valid for 5 days or something. I didn't know another way of linking to a file sorry. So the throttle position sensor seems to be working 100% fine. Always went up to smack on 4.28v when my foot was flat down, stayed there constantly when I held my foot down, and seemed to respond very smoothly when I was pressing & releasing the throttle. I can't make a graph in excel because it says something about there being an invalid link to an external file that is inaccessable?! I don't really know how to use excel very well. I don't know if you can make graphs in ECUexplorer?

I can still do this if you want, but the high idling thing was only on the first drive after I disconnected the lambda sensor. It doesn't happen anymore, it seems perfectly normal. Actually, thinking back generally......every morning when I go to work the car seems to be in a different 'mood' as far as idling/revving is concerned. Some mornings it just wants to rev. It sometimes revs high to the point that when I change gear the revs actually go up between changes rather than down. This is only when I'm pottering along, not doing much over 1,700rpm sorta thing, and when the engine is cold. Never really thought anything of this before, but thought it might be worth mentioning now.

I'll have a look at this tomorrow and let you know if I find anything.


Here is a link to the complete log file that the two extracts were taken from in my previous post.......

http://mbf.me/uzEd

Just incase you wanted to look at the rest of it to see if anything jumps out.

I'll put the lambda sensor back on at some point tomorrow, and try and go out and log some data with that connected.

Thanks again for all your help, it's hugely appreciated! I feel we're getting close to a possible solution!

simonds1

Quick question I forgot to ask......

Remember when I said I disconnected the MAF and it obviously brought the check engine light on? I've currently got my lambda sensor disconnected, and have done for nearly 2 days now........but no check engine light? Does that seem odd?

simonds1

I've just had a look now, and all the pipes seem fine. From a visual point of view anyway, I suppose one of them could contain a microscopic split/hole that you just couldn't see with you're eye. One thing I did notice tho was this.......



This is towards the front of the car, just infront of the air box......and the pipe that I've highlighted seems as tho it's jammed somewhere in underneath the airbox. It's very tight sorta thing, there's no movement in it at all. I can't think when this could've happened? I have never had the air box off, and the only time I can think the airbox might've been taken off is at my last service, which was a major one. I don't know if this pipe is actually worthy of causing my problem or at least contributing to it a bit? I doubt it very much.

joz8968

iTrader: (13)

If I've got my orientation right, then that's a view of the DV from underneath, right?

If so, then that is the DV's vac hose that get a reference from the inlet manifold. That pipe should be totally free and with no kinks/collapses in it. If not, then it will affect the opening and closing charactersitics of the DV and could well cause the intermittent on/off boost issues you're experiencing.

I'd get a brand new length of OEM vac hose and replace it. Make sure also that the airbox lid is fully resealed correctly as well, etc...

joz8968

iTrader: (13)

I'm confused... Is yours a MY00 Impreza?

That's a pic of a early MY93-96 DV?!?!?!