Air Filter Box Modification

[Pingu]

Looking at the intake system of the M Roadster, I thought that the air filter box could do with some improvement...

I built a flow bench to measure the flow rate through the air filter box...

but this rig would only measure comparative differences.

So, I built another rig. The data from this rig can be used to calculate actual flow rates...

I tested the standard air filter box and the volumetric flow rate was 444 CFM @ 28" H2O...

The flow rate through the MAF and trumpet was 608 CFM @ 28" H2O...

This was the target. I could never get better than this without supercharging.


There is an obvious restriction at the front of the air filter box. Air has to squeeze through the hatched area...

The flow rate through the standard half-box with MAF was 535 CFM @ 28" H2O


I cut away the side and top of the box and increased the area that the air can flow through (the image shows the trumpet twisted to 60°)...

I used shaped steel and fibreglass to reshape the air filter box...

The flow rate through the modified standard half-box with MAF was 542 CFM @ 28" H2O


The next modification was to twist the trumpet tube...

O° (Standard)

The flow rate through the modified standard half-box with MAF was 542 CFM @ 28" H2O

15°

The flow rate through the modified standard half-box with MAF was 532 CFM @ 28" H2O

30°

The flow rate through the modified standard half-box with MAF was 552 CFM @ 28" H2O

45°

The flow rate through the modified standard half-box with MAF was 534 CFM @ 28" H2O

60°

The flow rate through the modified standard half-box with MAF was 559 CFM @ 28" H2O

I suspect there were some errors in some of the measurements, but most of the actual measurements for a twisted trumpet were better than a standard trumpet.


The final modifcation was to the intake tube. The standard intake tube directs air through a tube of diminishing cross-sectional area (4185 sq mm reduces to 3287 sq mm) to only one half of the air filter...

I removed the intake tube...

and cut away a large section of it...

The air is now directed towards a much larger area of the air filter...

The standard air flow is shown in red and the new air flow is shown in green...

The air flow no longer doubles back on itself.

The flow rate through the fully modified air filter box was...

Trumpet 0° = 481 CFM @ 28" H2O
Trumpet 15° = 482 CFM @ 28" H2O
Trumpet 30° = 484 CFM @ 28" H2O
Trumpet 45° = 484 CFM @ 28" H2O
Trumpet 60° = 485 CFM @ 28" H2O

an improvement of about 9%

The modified air filter box fitted in the car...

 

[andyglym]

Interesting stuff

 

[Antm72]

Nice work

 

[Grumps]

As always John I understand every bit of that!

 

[Ianmc]

Does it make any difference on the road?

 

[Jack Ratt]

Why not just fit the K&N type filter with a straight flue ?

 

[Pingu]

Does it make any difference on the road?

Hard to tell. It was wet today, so I couldn't give it the beans.
Also, the difference between 300bhp and 310bhp is not really noticeable.
I'll use Testo to find out if there is a difference.

 

[Pingu]

Why not just fit the K&N type filter with a straight flue ?

No reason, except I would rather build performance than buy it.
Ideally, I would build a completely new air box that could feed filtered cool air into the trumpetted MAF.
The BMW paper filter is very good - MUCH, MUCH better than an oiled Simota. I've not tested K+N, but would be happy to test it if someone has one.
A dry Simota filter is better than a standard paper filter, but they need to be oiled to work. The flow rate is almost halved when oiled.

 

[Pingu]

Why not just fit the K&N type filter with a straight flue ?

re the straight flue. My tests have shown that ram trumpets make a massive difference to flow rates.
I suspect that the vena contracta is much larger when "rammed" air is pushed through an orifice.

For example, a simple 90mm square-edged hole flowed 677cfm @ 28", but adding a Plasticine ram ring, the flow rate was 732cfm.
For a 29mm hole, the figures were 105cfm and 130cfm

Flow rate is not just about bigger is best. If you run a tube directly to the MAF, you would need to consider tuning the length of the tube, and if the tube length is good for one rpm range, it would be poor for others.

The only way to know what's good is to test it.

 

[Jack Ratt]

re the straight flue. My tests have shown that ram trumpets make a massive difference to flow rates.
I suspect that the vena contracta is much larger when "rammed" air is pushed through an orifice.

For example, a simple 90mm square-edged hole flowed 677cfm @ 28", but adding a Plasticine ram ring, the flow rate was 732cfm.
For a 29mm hole, the figures were 105cfm and 130cfm

Flow rate is not just about bigger is best. If you run a tube directly to the MAF, you would need to consider tuning the length of the tube, and if the tube length is good for one rpm range, it would be poor for others.

The only way to know what's good is to test it.

Square holes, square ducts, square bends, square anything, creates turbulence in any type of flow, gas or liquid, so by eliminating it you have to increase or rather maintain optimum flow rate.

Round section, straight ducting is a must if you want to get all (or as much as you can) of the inducted air into your engine. You know the flow will be reduced as it navigates the inlet system as it goes into your engine, this is because of the very short duct lengths and the continual change of flow direction through the inlet manifold. You still lose a lot of flow even if you have a supercharger fitted, but more gets into the engine because of the extremely high volume you start with.

Now, if you could just turn your zed engine through 90 degrees, with the inlet facing the front of the car. Engineer individual throttle bodies for each cylinder that are fed through straight inlet ducts then I think you'd have a world beater.

Many years ago, over a few beers, I discussed this idea with a mate and we mulled over turning an inline four cylinder motorcycle engine through 180 degrees so that the inlet was pointing forwards and the exhaust venting out the back. Totally opposite to how they are actually built. These engines have a carb per cylinder and each carb is on a straight duct. We figured that this would give a proper ram air effect. Just need to engineer a fuel monitoring system to prevent lean running.

Another bright idea that we found in the bottom of a beer glass was to build our own blower, by bodging a centrifugal motor from a Dyson and fitting it into the induction system.

There's nothing wrong with experimentation, so carry on regardless as we say down the pub

 

[Pingu]

Square holes, square ducts, square bends, square anything, creates turbulence in any type of flow, gas or liquid, so by eliminating it you have to increase or rather maintain optimum flow rate.

I didn't mean square shaped hole. I meant square edged hole. The hole is round. But I agree with what you say. I'm not sure of the physics, but I suspect that the vena contracta of any shape is the same distance from the edge. As a circle is the furthest edge-to-centre of any shape, it is the best.

Really sad to admit, but a couple of years ago I bought some buckets and made some aluminium shapes all of the same cross-sectional area. Water flow through each of the shapes was very similar until a vortex was created. I still have the buckets and the blocks - I may have to try the same experiment with air.

Engineer individual throttle bodies for each cylinder that are fed through straight inlet ducts then I think you'd have a world beater.

Already half way there. Just need to drive sideways more often

Many years ago, over a few beers, I discussed this idea with a mate and we mulled over turning an inline four cylinder motorcycle engine through 180 degrees so that the inlet was pointing forwards and the exhaust venting out the back. Totally opposite to how they are actually built. These engines have a carb per cylinder and each carb is on a straight duct. We figured that this would give a proper ram air effect. Just need to engineer a fuel monitoring system to prevent lean running.

I think that every engineer has this idea. We thought about scooping air straight into the back of a set of DCOE Webers in a Mini, but (like you say) how would you control the fuelling.

Did you also come up with fan blades on aircraft wheels?

Another bright idea that we found in the bottom of a beer glass was to build our own blower, by bodging a centrifugal motor from a Dyson and fitting it into the induction system.

Nothing wrong with a 12v supercharger. You could use a double-pole switch to activate the nitrous oxide at the same time. You may need some extra strong cylinder head bolts

There's nothing wrong with experimentation, so carry on regardless as we say down the pub

 

[Jack Ratt]

The vena contracta effect will always be there, at every bend or hole (jet) and it's effect can be nullified quite easily, it's just that the tight confines of a car engine bay just does not give us the room to do this.
In an ideal world you would remove all the bends from the induction system. This is near impossible to do so the next thing is to give the flow the best chance to recover velocity after the vena contracta.
Duct lengths need to be around 4-6 diameters long for the flow to straighten out and regain volume and hence velocity after the vena contracta. Again, you can see this is not practical to do.

 

[Nodzed]

I've not tested K+N, but would be happy to test it if someone has one.

I have one, PM me your address and I'll post it to you, would be interested as I use them on my cars (could be going back to paper filters )

 

[NZ00Z3]

Have you used TestO to discover how much air the engine needs? It would be good to compare with your new designs capability.

 

[Pingu]

The vena contracta effect will always be there, at every bend or hole (jet) and it's effect can be nullified quite easily, it's just that the tight confines of a car engine bay just does not give us the room to do this.
In an ideal world you would remove all the bends from the induction system. This is near impossible to do so the next thing is to give the flow the best chance to recover velocity after the vena contracta.
Duct lengths need to be around 4-6 diameters long for the flow to straighten out and regain volume and hence velocity after the vena contracta. Again, you can see this is not practical to do.

I've just tested the shapes using the rig. They each flow similar rates. The differences would be due to tolerances in the rig and the shape dimensions.

I ran two tests at different air-speeds and the results were consistent for all the shapes.

I struggle to believe it, but it would seem that the shape is irrelevant. This must be why shape is not to be found in any of the maths.

 

[Pingu]

Have you used TestO to discover how much air the engine needs? It would be good to compare with your new designs capability.

Not yet. The ground has been too wet and I have a binding brake. Testo is next.

Maths could tell me how much the engine needs...

8000rpm, 3.2 litres, 100% volumetric efficiency = 452cfm, but I don't know what the "suction" pressure is. My values need to be converted to those of the engine.

 

[2DunnsZ3]

WOW!!!!!! I was just wondering what cold air intake was of choice! And now I feel like I got a 4 year degree on air and fuel!!

 

[the Nefyn cat]

Another bright idea that we found in the bottom of a beer glass was to build our own blower, by bodging a centrifugal motor from a Dyson and fitting it into the induction system.

Bit late, I know, but

View: https://www.youtube.com/watch?v=cbGWgvJN1_8
is one way of putting it.
And the reversed head bike engine idea was tried, can't remember by who, and failed. Seems the air-speed in the average inlet manifold is somewhat higher than most motorbikes can reach anyway, there's no gain to be made that way.

 

[Pingu]

Bit late, I know, but

View: https://www.youtube.com/watch?v=cbGWgvJN1_8
is one way of putting it.
And the reversed head bike engine idea was tried, can't remember by who, and failed. Seems the air-speed in the average inlet manifold is somewhat higher than most motorbikes can reach anyway, there's no gain to be made that way.

re the video...I found out all about the difference between moving air and pressurising air when I was making my flow bench. I started with a big 1000cfm fan that was used for ventilating a woodwork shop. It was rubbish. A vacuum cleaner motor did a much better job at compressing the air. This was due to the high speed of the turbine AND the small gap between the turbine blade tips and the housing.

 

[Pingu]

I tried to do a test this afternoon, but when I floored it at 1800rpm in 2nd all I got was wheelspin.

I used Testo and three manometers, but I could only read two of them. The sun was shining on the screen for the throttle trumpet manometer. Here are the test points...

1. Testo data from OBD port
2. #4 throttle trumpet
3. Plenum elbow
4. Air filter box (before the filter)

Here's the results graph...

The first throttle application just resulted in wheelspin (even though the road was only moist and the car was already doing 15mph).

The maximum depression in the plenum is 23 cm H2O. This is really good information as it shows the approximate depression that any air filter modifications should be tested at.

I need to do a proper (no wheelspin) test to confirm the maximum depression and to compare the A/F rate against the throttle position.