Nice idea I saw the other day:

The MAF is only good for measuring enough air for c. 300bhp and presents some restriction to the air flow.

So... create a bypass tube of similar dimensions which runs in parralel to the MAF between the air filter and the turbo.

Advantages are some reduction in pressure drop, MAF that can now measure up to 600bhp-ish flow* and it costs pennies.

Anyone tried this?

Mark


* Obviously a remap needed needed here to take advantage of more accurate air measurement for cars > 300bhp - i.e. avoid using estimation

Am I missing something here? The turbo inlet and filter are the same size or (very similar) to the maf so how will this help?

What about the extra air not being metered in the other pipe?

You either need 2 MAF and combine the signal some how or assume that the air flow down each path is the same and therefore double the signal. Niether of which is a good solution.

Either way I can't see the point.

So your engine will be sucking in a huge amount of unmetered air thats bypassed the MAF ?

I assume if you have two inlets then the volume of air being drawn will remain the same however it will be over a much greater area therefore the flow over the MAF sensor will be slower. This would effectively cause the ecu to read less air and lean the mixture which could be compensated for with bigger injectors and technically extend the MAF range. It's a similar principle to a larger MAF housing but I'm not sure it will be very easy to moderate these changes within the ecu.

Exactly Flea:

The air is not unmetered, the MAF reading will still be proportional to the amount of air being drawn into the engine, but the reading will be about 1/2 the value for a given flow.

The advantage over a large MAF is that you don't need to actually buy one.

It does mean that the fuelling maps will have to be completely re-written, but the advantage is that estimating the air flow from throttle angle or RPM (or whatever) which the ECU uses over c. 300bhp can be avoided.

Except it won't be half will it. It will be an unknown proportion of total flow so will not be more accurate. Assuming all else is equal(length, bends, diameters, etc) there will always be more flow down the non metered side as it does have a bloodly great sensor in the middle of it blocking the flow. So what you will be metering will most likely be less than half the air flow. I don't see how this is more accurate.

Also i dont know the in's and outs of the ecu but what will it make of a maf reading thats circa half what it should be? will it assume the maf is duff and disregard it in which case its pointless. If the ecu could use the signal it still wont get the correct a/f mixture as it does know what % of the total is being measured. So again pointless. If your serious about this route you really need 2 mafs and somehow modulate the signals together to get a true airflow rather than an assumed one.

So you save on a second maf but spend extra on pipework and the hastle of trying to route it. And I guessing the remap involved with be more comprehensive and thus more expensive.

You are still estimating at the air fuel mixture but just using a different parameter to do so.

Already testing this idea.

I have a 2 split Y pipes of equal dimensions. Both MAFS are connected within the Y piece. So it goes like this:
Air Filter>Y piece>Splits in Half>two AFM>Joins back>Ypiece>Turbo>Engine.

Only one AFM is connect to the ECU the other is just there as a dummy to keep the dimensions correct. This way the connected AFM will measure exactly half of the true air coming into the induction system thus potentially doubling the ability of the AFM.

Problems:
Space. It isn't easy to fit all this under the bonnet and keep the AFMs a safe distance away from the turbo. Well the Bravo has an issue not so sure about the coupe.

The major problem is re-scaling the ECU to account for this and to be honest I don't think it will work as there are not enough point to ensure a smooth transition. This throws the map down the other end of the scale meaning that when it comes to idling you have a problem. The AFM will be good for above a certain point but below it you now have a new problem as you are putting more points up the higher end of the scale meaning the lower end points will have to be removed. Its a good idea but it just switches the problem around so instead of 'maxing out' you are ‘minimuming out’

It’s the same principle for injectors you put too bigger injectors on and you can’t pull out the fuel as you are re-scaling the map. The Bosch ECU has its limits and there are just not enough points to go too far.
I can see that the AFM can cause problems on our cars but its best to stick with it or cough up and go standalone with a MAP sensor.

Hyperlink: The point is that it doesn't actually matter what the proportion is or indeed if it remains constant, thats for the map to interpret.

This is a massive oversimplification, but all the tuner needs to do is to work out the MAF signal x is equivalent to y kilos of air flowing into the engine per second which means give me z duty cycle for injectors... adjust the map accordingly. He would start off with the assumption that the signal is 1/2 of the previous and tune it using the A/F ratio.

There is no difficulty with this as it stands. A large MAF would be doing essentially the same but costing more.

2 MAF signals from two sensors is also no benefit - its just a voltage signal 0-5v.

Hi TurboJ:

Kewl - using a dummy MAF sounds like a better idea, the restriction to flow should behave identically for both pathways no matter how much air the engine is drawing.

As for the problems - I can see that the granularity of the sensor reading could be problematic, especially at low loads - the A/D conversion is only one byte IIRC - but that would give you 256 points which I thought would be enuf?

Are you saying that you've had a remap and have got a really bad idle or somesuch? I would have hoped the lambda adjustment would be able to handle this - but then maybe I'm just placing too much faith in the old Bosch ECU...

As for space... why do you guys want to tune Bravos instead of just getting a Coop?????

Actually - perhaps a compromise: if the dummy sensor was, say, 2x as restrictive as the original MAF, then the total measurement could be valid for up to c. 450 bhp without stretching the points quite so much... but I presume you're testing this as well.

The only way I can see it working is to build the bypass pipe, then perform proper tests to determine exactly what percentage of air goes each way, then intercept the voltage reading before it gets to the ECU and boost it accordingly with some electronics wizardry.

The splitting of the air is NOT the problem this is actually very simple a Y piece is the perfect solution. It’s just symmetry so what passes through one pipe will be the same through the other.

The problem is getting the ECU to understand what you are doing, unfortunately our Bosch ECU's don't like being pushed too far. I'm not saying that it’s not possible but realistically it will take far too much time on the dyno to get right when it would be far cheaper to map a standalone from scratch. I never pushed this idea any further as it would take too much of my time to perfect so instead I scrapped it and am sticking with the standard MAF. I may come back to this idea sometime in the future when I can book untold time on the dyno and have a better understanding of the Bosch system.

There is a more elegant solution, take the MAF innards out of the existing pipe and mount them in a larger diameter one. Going from a 4" to a 5" pipe gives you a 50% increase in flow

In that case try intercepting the voltage that the MAF is feeding to the ECU, and double it. dont even need to touch the ECU or do any mapping if it works.

Then again maybe doubling it wont work, maybe a more advanced formula would be necessary, I dont know

The point is that the stock MAF will reach it's maximum output at a particular airflow. By halving the airflow and then doubling the output you are right back where you started i.e. although the MAF won't be running out of range, by doubling the voltage it will appear to have reached it's limit to the ECU.

The idea is that you remap the ECU to work with the reduced signal from the MAF, but you then have a lot more margin before the MAF runs out of capacity. The problem is that the airflow between each map point in the ECU is doubled, giving potentialy poor calibration.

there is a good reason why you shouldn't do this, but I can't remember what it is I think it is something to do with the measurement of density gets thrown off which equal flow would not solve. Not that it is possible to get equal or consistently proportional flow because they will not have the same air over them coming from outside the car. In thid gear and fourth gear the map is the same but the flow over the pipes will not still be equal it will all depend on what is between the MAF and the outside world.

There is no reason why you cannot create a circuit that gives an aggregated voltage proportional to the 0-5v range for the air passing through 2 MAFs. It's not exactly rocket science.

I think a bigger tube is a much more simple and elegant solution and there are ones out there that will fit the standard Bosch MAF element on the coupe. Moreover it is the CORRECT solution because it is what Bosch do to overcome this. An 80mm housing is used on an Audi TT 225 for instance, and there is one that is used on a Porsche 964 3.6 turbo (can't recall the diameter).
It is standard practice on lesser VAGss to change the MAF housing and mappers who do a live map style mapping can cope with it.

Your problem is that PT may not be able to and you've got no-one else to turn to. If they can't cope with a larger MAF housing then how the hell are they going to deal with totally unmetered and potentially inconsistent flows of air. So the whole discussion may be academic...

By "the stock MAF will reach it's maximum output" do you mean it cant physically produce more than 5V, or that the ECU cant accept a signal higher than 5V even though the MAF can produce it?

That is not what is happening. The voltage is NOT being doubled it's being halved. You are halving the airflow so the voltage will also be halved. Which means you are doubling the ability of the MAF but the problem is the ECU.

@ JohnS, That is exactly the point I am trying to make. PT are the only people whole map this system well and at the end of the day they are a business. This experiment is more research and development so to get PT to do it for you (if they even agree) is going to cost a bomb because they are gonna want the hourly rate no matter what. If the Bosch mapping software was made free to the public then further development could be under taken by enthusiasts but for now it standard or standalone.

Just been thinking and do we know for sure that the flow rate and electrical signal is linear? If it is not a linear relationship then half the flow won't provide half the signal but either more or less than half. What impact would this have on the mapping?

Also what "resolution" does the maf/ecu operate at? By effectively halving the working scale of the maf what will the impact be on the maf signal and the ecu? Could this lead to a slightly stepped profile of air flow vs. voltage?

If the Bosch software was available I can guarantee there would be at least 5x as many blown up engines on here a year.! A bit of a double-edged sword really. I have some equipment which I've got to work and do basic mapping changes live but I don't have access to a normal 20VT so it isn't going anywhere fast.

I think you could possibly try to get a Porsche tuner to map the car as some revisions of the 993 have the same basic ECU (Motronic 2.10) although a differrent sub-version. I believe they also change the MAF housing. The 964 RSR or 993RsR apparently uses our ECU but runs without a MAF That's more doable without boost to worry about


To answer the question about range the MAF actually goes out of range (#FF in the ECU) somewhere about 4.8v if I remember correctly. For two MAFs you could produce a circuit that if you get 2.6v in one MAF and 2.4 in the other it reports say 2.5v to the ECU. It is such a basic circuit. But still the ECU is the issue not the signal

Cheers - some nice info there John.

So to sum up:

a) Same sensor in a larger housing is a better solution
b) PT may not handle it which is more likely 'cos...
c) ... no-one's done it so we don't know if...
d) ... the ECU can fuel better higher up the flow without losing too much calibration lower down* so...
e) ... we won't know if its worth it without a guinea pig

All in all glad I asked



* I still don't think this will happen tho

John did you encounter any maf problems when tuning past 300bhp.

Is it possible to get into the ecu and alter the voltage of the maf and if you can how easy would it be.

John is on Motec. He has MAP not MAF.

Ah well that's that question answered then.

Recalibrating the MAF voltage isn't a trivial process but if the map location was known it could be achieved. It is not altogether necessary to do this, simply matching injectors to a MAF housing would effectively achieve the same process however you then have to recalibrate all the ignition timing as the standard values will be on an incorrect scale.

Tell me more or should i phone you?

Having a few problems at the moment