StephenDone

I have just upgraded the induction on my MY99 WRX. I replaced the intercooler hoses with Samcos, the intake hose with a BPM, and the air filter with a GGR large K&N cone.

I took dyno runs before and after these specific mods.

My dyno results have me a bit stumped. Whilst I have picked up 10lbft from 4000 RPM to 7000 RPM, I have lost loads (up to 30lbft) between 1000 & 3200RPM.

Any ideas on why this might be ?

Here are some dyno graphs...

brickboy

Hard to say exactly but air flow mods can be an inexact science. It certainly could be intake air temp, especially on a dyno where the flow of cooling air is much less than on the road, or it could be resonator-related.

First thing to try would be the air temp, can you rig up a cool-air ram?

StephenDone

Thanks for taking the time to look at the graphs. Before I make any changes, hopefully I'll hear from someone else who's fitted a cone to an Impreza, has had the problem and has fixed it too.

Whilst I agree it isn't an exact science, anyone who has fitted a cone filter has effectively removed the resonator box, and so if that were the reason, everyone would have the same problem.

I think it's probably air temp too - that was my first thought. There are several ways I could get cold air into the cone, but they'll all cost money or take time to make, so again, I'll wait for some more votes before I invest the time/money.

Cheers

Steve

Davvers

You can get cool air into the cone quite easily. DIY stores sell aluminium corregated tubing for the exhaust on tumble dryers. This can be fitted in the inner wing and down to behind the orange intakes in the front dam - if you remove the resonator box, whixh you might as well do if you have a cone. Very cheap - and works - well probably!

AlexM

Stephen,

As you are running on the OEM ECU (i.e. using the MAF), there is a possibility that resonances within the intake pipe are causing the MAF to mis-read. I read that the maf can get upset if it is exposed to back flow or turbulence.

I seem to rembember reading somewhere that cone type filters that bolt straight onto the maf can also cause this kind of effect, resulting in incorrect airflow values.

Were youre runs at PE? If so it may be worth getting them to put the car on the select monitor to see what the MAF output is in the range where you've lost the torque.

heating of the intake charge is also a possibility as RR fans will never give as much airflow as when driving the car normally.

Cheers,

Alex

P.S. your dyno graphing servlet is very nifty!.



[This message has been edited by AlexM (edited 21 September 2001).]

StephenDone

Davvers: That's a cunning plan. B&Q on the way home then !

AlexM: Good point. So do you reckon it may be running rich or lean ? I have a GGR kit that has the STI spun ally inlet trumpet which is meant to sort that kind of thing out - the MAF's not bolted onto a flat plate like the 57i kit apparently has.

Have a look at this link, and say if you think I'd still be affected...

StephenDone

Ok,

My plan is to leave the MAF sensor where it is, but move the cone into the wing. I opened up the wing this evening and removed the resonator - there is plenty of space down there. A bit like the APS cold air induction kit, but bigger :-)

This will have two effects...hopefully:
#1 Draw in much cooler air through the orange intake below the indicator.
#2 Put around 12 to 18 inches of 90mm silicon hose between the cone filter and MAF. This should allow turbulence to settle more before hitting the MAF.

What do you think of this idea ?
It's a combination of the suggestions so far.

Cheers

Steve

john banks

How do you think it drives on the road Steve?

My K&N 57i gives obvious increased pickup between 1000 and 3000rpm. When the wife got me to take it off for noise reasons I really noticed the difference.

carlos_hiraoka

Steve,

Take a look at:

StephenDone

Hi John,

>How does it go on the road ?
>
Very well if you a flying, but rather sluggish if you plant it at 60MPH in 5th. Throttle response is excellent at high revs - better than ever.

Hi Carlos,

Yes, the BPM pipe mentioned in that thread is the one I have. Much easier to fit than the Samco one on a 99MY - I had that too before returning it. My site has details on how to fit it if you are interested.

Cheers

Steve

Sam Elassar

hi stephen
very interesting finding, were both runs done one after the other ?
i doubt that it has anything to do with the MAF meter, if it is going to mis read at the low end then it is most likely going to mis read even more at the top end ? you seem to have gained at the top end.

SO here is my hypothesis
you are running standard with power manifold and full decatted exhaust, subarus run very rich as we all know. so all your breathing mods might have got you running a little leaner in the mid range and hence you are loosing a little power. maybe

the other theory of drawing hot air is also suspect. if you are sucking hot air on the dyno at low revs you will also be sucking it at the top end ? so the difference should be across the range.

i am sure a lot of people will come on know and tell me that all i have said is utter rubbish as usual

StephenDone

Ok, remedial action is underway. Whether it works or not remains to be seen. But I enjoy this, so what the heck...

I have ordered some nice large (89mm) silicon hose, plus some bends, and have had some stainless steel joining pieces made by a friendly exhaust making company. I am going to drop my GGR cone into the wing, some distance from the MAF. This pipe is the same diameter as the MAF. I now have to find/make some maf adaptors, hence my other posting.

Hopefully I'll get cooler, more laminar air flow from this mod.

I'm not being as scientific as I'd like, since I can't find answers to some of the questions that I have.

What I do know:
---------------

# The MAF can get upset by turbulence, and therefore meter incorrecty. Moving the MAF away from the potentially disruptive influence of the cone _may_ help this - a nice long, smooth bit of pipe will be in between.

# Air intake temperatures are higher when using an 'under bonnet' cone filter. Moving the cone to the wing will reduce intake temperatures to some degree.

# Resonance in the new large & smooth BPM intake pipe may be causing the problems seen in the rev range before the turbo is boosting. Lengthening the intake pipe will lower the resonant frequency and possibly pull back the drop in power to a lower RPM. The original intake pipe would have resonated less and at a lower frequency, due to its nobbly shape and smaller cross section respectively.

What I don't know/understand:
-----------------------------

# Still haven't found anyone who can tell me what the resonator box does. Is it for noise reduction or something else ? I've still got the option of fitting a custom resonator to my pipework later, so I'm going to get on and try the above and see what happens, even if I do later find I need a resonator - the space in the wing is huge !

# I've been told that the car could be running lean with the mods, though I don't know why. I would expect the MAF sensor to fuel according to air flow - that's what it is meant to do ! My losses are low down, so there's no way I'm off the top of the fuel/air map...is there ? Unless the MAF's misreading due to turbulence/resonance.

# I've been told that the turbo will be spooling up slower, since the air it's chewing on is more dense. I agree with this, but since the air is more dense, the turbo won't have to spin as fast to make the same boost. I would have thought this would be better, not worse, since the wastegate can stay open more to produce the same boost - less exhaust restriction.

Any comments ?
I know I may be throwing good money after bad :-)

Cheers

Steve D.

StephenDone

Hi Sam,

Thanks for your reply.

> very interesting finding, were both runs
> done one after the other ?
>
There was a couple of months in between, though the ambient air temperature has dropped by about 2 degrees C. All else has remained constant. No ECU resets or anything else have happened.

> i am sure a lot of people will come on now
> and tell me that all i have said is utter
> rubbish as usual
>
Well I appreciate your ideas anyway, right or wrong. If ten people post, and one of them comes up with an idea that nobody has mentioned, then that's great as far as I'm concerned.

Cheers

Steve

TTZZ

Stephen,
it is very simpel, it is your low restriction airfilter causing this problem.
Change it to the standard filter and you will be amazed.

To keep the turbo on pressure at low revs, the underpressure between the airfilter and the compressor is used to keep the wastegate closed.

When you have a high restriction filter you will have more negatieve pressure, the down side is that on high revs it cost some air.
This means, with a standard filter you get more torq at low revs, and with a low restriction filter you get more torque at the higher revs. It is up to you what you want.

Sam Elassar

hi stephen
it makes more sense now, but the fact you have had the runs done on sperate days means that it will not very accurate, combine that with the fact that an air filter will give a 3-4bhp difference max.
your little dip in power could be due to maybe some dirty wastegate soleniod tubes, while you gains could be due to the car loosening up. unfortuantely the rr are not usually that accurate to measure small little differences even on the same day. i have had a couple of private RR sessions and not one run was the same as the other. you can just look at the general curve.

if you want my opinion there is a lot of other things to spend your money on that will give you better gains.

on a different note, did the turbo spool up quicker with the power manifold ?

Bob Rawle

The resonator is a heimholtz (spelling ?) resonator and is designed to tune the inlet tract to get best laminar air flow etc and also to damp out air "resonance" which would upset the maf.

By increasing the size of the inlet pipe you have effectively slowed the air, its not "more dense", this changes the "system" that the turbo is seeing effectively forcing it to work harder ie not spool as quickly hence the result you see.

Since slowing the air changes the reading that the maf produces the ecu is now sensing "less air" and is therefore adding "less fuel" as a result, in fact the air volume is the same or even a bit greater, thats why the engine is now leaner in that area. The maf works on the principle of heating a wire or film plate to achieve a set temperature, the more air (faster) flow the greater the cooling effect and the more energy it needs to heat it up, thats how the variable output is generated, so slow the air down by increasing the diameter of the pipe and the cooling effect is less for the same volume, less cooling equals less output equals less fuel. The maf also has air temperature compensation built into it.

Basically you changed the "calibration" by doing this, you need to compensate for the changed callibration to get back the loss by adding more fuel in the right places.

The wastegate is opened and closed by the ecu to maintain boost targets. Its part of a closed loop system.

The car's run rich at wot, part throttle response is different and more susceptable to mods When you open the throttle the ecu "traverses the map" to get to the final point, as it does this it is going lean.

Classic case really.

hope that clarify's

StephenDone

Thanks for all the replies everyone...

Hi Jan
======
That's a very good suggestion. Unfortunately, since I've changed my intake hose, I cannot revert to the standard airbox easily. I will try picking up some air from nearer the turbo, and see if this has a higher vacuum than further back, closer to the air filter - I expect it will have, since the hose is much narrower near the turbo. This is easy to try, so I'll give it a go. btw, What boost did PE set on your car - your torque curve is superb.

Hi Sam
======
I am comparing power at the wheels, and I have found this to be quite repeatable. I am happy to rely on these figures. As for the power manifold, I got more torque at lower revs, so didn't get the 'wall' of power liek I used to - much more progressive now.

Hi Bob
======
Thanks for the information. At my next dyno run, I'll get the mixture checked for leanness/richness. I'll also think about adding back a custom resonator box later on. Unfortunately, I can't change my ECU, since my insurance won't allow it - strange corporate insurance policy ! Hopefully PE will be able to invisibly upgrade ECUs on 99/00MYs soon.

I'll keep you all posted on how it goes...

Cheers

Steve

[This message has been edited by StephenDone (edited 25 September 2001).]

john banks

How about running your DSO off the lambda sensor and going for a drive to see? 0-0.7V from lean to rich.

[This message has been edited by john banks (edited 25 September 2001).]

madou

helmholtz

sorry to interrupt guv'nor :-)

<BLOCKQUOTE><font size="1" face="Verdana, Arial">quote:<HR>Originally posted by Bob Rawle:
<B>The resonator is a heimholtz (spelling ?) resonator and is designed to tune the inlet tract to get best laminar air flow etc and also to damp out air "resonance" which would upset the maf.

By increasing the size of the inlet pipe you have effectively slowed the air, its not "more dense", this changes the "system" that the turbo is seeing effectively forcing it to work harder ie not spool as quickly hence the result you see.

Since slowing the air changes the reading that the maf produces the ecu is now sensing "less air" and is therefore adding "less fuel" as a result, in fact the air volume is the same or even a bit greater, thats why the engine is now leaner in that area. The maf works on the principle of heating a wire or film plate to achieve a set temperature, the more air (faster) flow the greater the cooling effect and the more energy it needs to heat it up, thats how the variable output is generated, so slow the air down by increasing the diameter of the pipe and the cooling effect is less for the same volume, less cooling equals less output equals less fuel. The maf also has air temperature compensation built into it.

Basically you changed the "calibration" by doing this, you need to compensate for the changed callibration to get back the loss by adding more fuel in the right places.

The wastegate is opened and closed by the ecu to maintain boost targets. Its part of a closed loop system.

The car's run rich at wot, part throttle response is different and more susceptable to mods When you open the throttle the ecu "traverses the map" to get to the final point, as it does this it is going lean.

Classic case really.

hope that clarify's [/quote]

StephenDone

Hi John,

I don't want to cut into my loom to read the voltages. I am on the look out for some ECU connectors, so that I can place a small extension lead in beteen the ECU and loom so that I can chop things there.

Anyone know where to get 'em ?

Steve

Gazzz

Bob,

Tell me if I'm wrong,
using a different pipe (i.e increased diameter) won't affect air speed and qty passing by the Maf, as its diameter and surface remain the same.

Maf will still read the correct mass air admitted (and not volume, as it is compensating air temp. on its own).

I would suggest Steve to have a look on what TTZZ described as the underpressure between the airfilter and the compressor, used to avoid wastegate openning at low revs.

You can solve this, even on a stock car, by using a calibrated spring+ball assy plugging the line up to steady boost pressure (carefully bench set your assy...)

This efficient and low cost mod. will let you scrap the restrictor originaly fitted in the line connecting compressor outlet to wastegate + solenoids lines.

Cheers fellows,

Gazzz !

Bob Rawle

You are assuming that the mass air flow meter is the "gating" csa in the air intake system ... not correct.

Interesting comments about lower pressure next to the turbo inlet holding the wastegate closed.

When I last looked the wastegate is in the exhaust turbine housing, not the compressor housing, thats connected to the exhaust not the inlet, so the only thing acting on the wastegate is exhaust gas "pressure" which try's to blow it open, thats why the actuator has a spring in it, to keep it closed till the ecu decides its time to switch boost pressure onto the actuator to open the wastegate.

Still I'm sure the answer is out there as they say.

Oh forgot to mention that I have actually logged the comparitive maf output's under the conditions described, my car may be unique of course.

john banks

These sorts of threads do not irritate mammary glands but provide VERY useful info. Keep it coming. Also saw Pat's comment in number 3 bigend thread stating that RR often show Scoobies rich but this doesn't necessarily translate into road given temperature differences - best read the thread rather than I mangle it - may or may not be relevant. But at the end of the day you don't drive it on the RR so you need to know it is not lean on the road- or am I being simple?

CharliePsycho

An interesting dynamic of fluid flow in pipes is that a turbulent flow actually passes more fluid than a laminar flow.

I even did a picture


The top is laminar flow, and although indeed in the centre of the pipe there is a higher flow, it drops off quite sharply. The turbulent flow (bottom pic) has a lower flow, but this is uniform across the pipe and give a higher overall flow.

Who said 'A' Level chemistry is worthless???

Anyway, how does this effect the MAF... FIIK, a lot depends on what is meant to be there, obviosly in an ideal world you want good dense turbulent flow all the way, however; is that what the turbo wants? is that what the MAF wants? where is the MAF, centre or side? (note it does not matter on turbulent flow)

Tips that may help the conclusion, the original pipework is far from smooth, especially the corrugated bit to the Turbo. This has a less restrictive effect than you think on a turbulent flow, so it may be designed to expect that (alternatively it may just be cheap)

If you are using the stock ECU I think it would be fairly safe to assume a turbulent flow along the length of the tubing (unless somebody can show me some special baffles in the stock airbox to smooth flow) to the turbo. So, the K&N nice uniform cone and trumpet *may* be too laminar flow?

Ah I feel experiments with hairdryers and airboxes coming on!!!

...©

StephenDone

Hello Mr Psycho,

The MAF sensor is roughly central. It's got all sorts of nobbles on it, but the main business bit is in the middle.

You are agreeing with me (I think). The MAF, sensor when connected to the air box will get a high reading, relative to the total amount of air going down the pipe - ECU is mapped to account for this. When you get rid of (some of) the turbulence, the flow rate down the pipe remains pretty similar, but the reading taken from the centre of the pipe is lower, hence the lower fuelling, lean running, and consequent drop in power.

Steve

john banks

Unless you were running rich before like most Scoobs then you might be more stoichiometric - ie 6-8% CO? What voltage would this be from the O2 sensor? (range 0-0.7 lean to rich).

[This message has been edited by john banks (edited 27 September 2001).]

CharliePsycho

Well, unfortunately I am probably not agreeing

I think I am saying that reducing the turbulence has 2 very bad effects
1) it reduces the total airflow
2) it makes it 'look' like increased airflow to the MAF

The question is, does the K&N hve more or less turbulence for a given airflow than the stock airbox?? Gut feeling is less turbulence <B> somebody who knows K&N filter really well tell me if I am wrong </B>

So laminar flow would make the mixture richer... The question is what about the turbulence in the feed pipe to the Turbo? from the pics on your very excellent website I would say that the new hose is wider as well as smoother; so you have a more laminar flow (maybe) but a wider cross-section ~ probably you have less pressure loss and more flow... *Probably*.

Looking at the dyno graphs though I suspect that the effect you are seeing on the RR is possibly from under-bonnet heating; low revs - low airflow ~ nearly all from under bonnet... High revs ~ high airflow being drawn through the engine bay, but cooler. The effect probably wouldn't be so evident on the road.

I think you have the right plan - APS type wing induction for cooler air (OK OK this *may* not make much difference on the road) this will certainly make you tests easier to judge on the rolling road. THEN start experimenting with the induction hoses (like swerl baffles after the MAF maybe??)

It has to be said, it is problems like these that make it interesting

...©

StephenDone

Hi Charlie,

I have to disagree. After replacing the air filter, with the same actual flow rate, the MAF sensor would see decreased air flow, according to my assumptions anyway...

The filter I have is a GGR K&N, not a 57i. It _should_ reduce turbulence. I.e. Turbulence should be less than std air box.

It you set up a test where you drew air through the MAF housing at a constant flow rate, a housing with more turbulence would see more flow at its centre than at the edges. The housing with less turbulence at the entrance would see a more even rate of flow across its cross section. Agreed ?

The MAF sensor inside the housing is at the centre. Therefore, with turbulence, it would see give a higher flow reading than without turbulence, although the flow rate is the same. Agreed ?

If you disagree, which of my asumptions above is incorrect ?

Like you say, I am going to do the cool air induction anyway. I'm certainly not convinced it will fix this, but it's a nice project, and I might learn something, or get the opportunity to start some more threads to annoy the t1t5 off everyone :-)

Cheers

Steve

[This message has been edited by StephenDone (edited 27 September 2001).]

StephenDone

Hi Bob,

What do you mean by CSA, in the context of the MAF ?

I mailed TTZZ offline to talk more about his suggestion. Concerning the MAF, what TTZZ meant was the vacuum against the wastegate actuator diaphragm. I think the wastegate actuator is a two way valve, either connecting the diaphragm to the intake or exit of the turbo compressor.

I spoke to Mervin at Power Engineering about this. He said it couldn't be the 'lack of vacuum', since the spring in the actuator has a strength equivalent to about 10-12PSI of boost. I.e. you need 10PSI of boost before you even have the pressure to force the wastegate open. Mervin said the amount vacuum present was insignificant, relative to the spring strength.

Like Gazzz, I still don't understand how the MAF can misread. The MAF has a heated strip, and a temperature dependant resistor of some kind. Is it a Wheatstone bridge arrangement Bob ? The cooling effect is proportional to the number of molecules of air that pass over the thremistor. I agree that the air may be moving slower, but that it will be more dense. Slower air will cool less. But denser air has a better cooling effect, since a unit volume contains more molecules. I.e. it amounts to the same thing as far as I can see. If this weren't the case, correct fuelling would be significantly affected by altitude, barometric pressure and air temperature - these all affect air density.

I could imagine another reason why the MAF may read differently when no longer connected to the air box: The MAF bolts straight on to a flat panel of the air box. Sucking air round the square edges of the air box hole causes turbulence at the edges. Because of this, the flow of air through the MAF may not be constant across its cross section. I.e. the portion of cross section where the sensor is may see more or less air than other parts of the tube, consequently over or under reading. I would guess that the centre of the MAF housing (where the sensor is) would see most of the flow, since the turbulence slows down the air at the edges. I.e. the MAF would over-read the air flow, and the ECU would be mapped to account for this.

For the car to fuel correctly with the air box, the ECU must be mapped to account for the poor metering.

I have stuck on a spun aluminium intake trumpet, which reduces turbulence and encourages better air flow. The exponential shape of the trumpet encourages smooth flow across the whole cross section. In other words, I may have improved the metering accuracy compared to the air box, though the ECU is now trying to compensate for an error that no longer exists. I.e the MAF sensor sees a certain air flow, and the ECU is mapped to assume this sensor is over-reading. BUT IT'S NOT ! And now the mixture is too lean. Just what Bob said a week ago. This is yet another possible reason - I'm not saying it's right, but I'd be interested to hear what you all think.

Like Bob says, the truth is out there. I'm just sitting on the fence at the moment :-)

Cheers

Steve

[This message has been edited by StephenDone (edited 27 September 2001).]

Gazzz

Steve, Charlie,

I've been reading all above comments but don't get it wrong -using hairdryers!-
MAF has not to be considered as in a wind-tunnel : nothing blows through, and this makes things a little different.

-first, flow is turbulent in any scooby pipe from 2000rm to top revs, cause of the centrifugal compressor -pitch/angle of rotating blades-, even without considering lines connected to the pipe,

-second, it sucks air through the pipe, MAF and filter. (Static air pressure from the filter to the blades is lower than ambient from about 2500rpm to top revs. : a Burbling -or Vortex- sucked flow as to be considered)

Back to what I wrote about the brass restrictor, I guess there is a difference between my 98' and latests, according to the -highly respectable- PE info:
I tested the car replacing the restricted line (from compressor outlet to the T-fitting) by an "open" one: wastegate openned from around 5psi with huge lag... I don't expect to find any spring in there, even broken!

NB:Interesting chat with you guys

Gazzz !