August West

Thought I'd post a composite of two dyno runs that I did, one with the stock airbox lid and K&N filter, labeled as the test run, and one with the Fabspeed Cup Airbox lid and K&N filter, labeled as the base run.

As you can see there's no appreciable difference in horsepower/torque numbers. The stock lid dyno run actually put out .7 max horsepower more, but that could be attributed to atmospheric conditions (the stock airbox lid run was done on 5/24/13, the cup airbox lid run was done back in 2/18/13).

Not sure how else this can be argued...the cup airbox is really only giving you an increase in induction noise. Don't get me wrong, the induction noise is awesome but to say there's a 8-10 horsepower improvement is just false.

These runs were done on a Mustang dyno...for those of you wondering why the numbers seem low.

NineMeister

A more meaningful comparison would have been to test both lids on the same day in the same weather conditions, with the same tires at the same pressure, the same gearbox oil temperature and with the same fuel and with the engine at the same cylinder head temperature and breathing exactly the same intake temperature. In other words there are just too many variable changes between your February and May dyno runs to make an accurate or valid comparison I'm afraid.

NineMeister

Sorry to sound like the teacher giving you a physics lesson, unfortunately nailing down variables to measure real differences is imperative if you are going to get meaningful results. This mighthelp to cheer you up though, especially if you read between the lines:

Here is the result of 9m dyno intake system change which I carried out on a 8000 mile 964RS that we had upgraded with a 9m Motec conversion. At the time I was fascinated with the detail of why and how the 9m Motec kit managed to gain 25hp above a perfectly mapped stock 964RS engine, this test and a few more like it gave me the answers. First of all notice that the run test numbers go from my expected best to worse. I use this approach so that any gains from reduced gearbox friction (due to oil temp raise) improve the results and hence slightly "normalise" the results. For the test I started all dyno pulls at the same cylinder head temp and air intake temp, and of course all within a few minutes of each other on the same day. To speed up the test I had a spare air box base and bypass tube pre-assembled ready to fit quickly between runs.

Results
Run 19 (green) = air flow meter bypass (round to rectangular tube) with standard air box, no air filter or lid
Run 21 (red) = standard air flow meter with standard air box but no filter or lid
Run 24 (blue) = standard air flow meter, standard air box and standard (new) paper air filter element

As you can see, there is more to be gained from removing the paper air filter than comes from removing the "restrictive" barn door air flow meter. So, what does that tell you about the effectiveness of a Mass Flow conversion which makes 300hp when the stock air flow meter (with optimal fuelling) can support over 315hp?

All interesting stuff proving that it is always good to ask the right questions but that you should only ever believe the answers that you get which come from verified data, not guesswork.

August West

Fabspeed Cup Airbox Lid Run in February: 55degF, 48% humidity, Rear tires were Bridgestone Potenza S-02s with no tread left.

Stock Airbox Lid Run in May: 70degF, 35% humidity, Rear tires were new Bridgestone RE970 AS Pole Positions with full tread.

I would think that with all else being equal, the February run would have generated more power because of lower air temperature and decreased rolling radius of the S-02s with no tread...smaller radiused tires produce more torque, right? (I'm sure there will be differences due to sidewall deflection, tire pressure, etc.)

An additional variable would be the difference between our California Winter Fuel Blend and our California Summer Fuel Blend... correct me if I'm wrong but I think the winter fuel blend generates more power than the summer blend, which has ethanol mixed in. Again, this would have favored the February runs over the May runs, but the dyno results don't reflect that.

The May run with the stock airbox was done on a hotter day, with new rear tires, and with the summer blend fuel, and yet it still pulled marginally better numbers (statistically equivalent, in my view).

Yeah, it would have been better to do it all in one day to eliminate many of the variables but it wasn't one of my objectives to do a lid comparison...it just so happened that I'd changed my lid from one dyno session to the other and thought I'd post the results!

August West

Glad to have someone of your caliber, Colin, in having discussions like this! Physics was one of my favorite subjects, so teach away!

NineMeister

I understand and respect your motives, clearly you have a handle on the variables so now understand where I am coming from.

When testing a 911 engine the most critical variables are cylinder head temperature and air intake temperature, a 10 degree swing in either can cause a greater variance than you observed between your tests. Hotter days, higher air temperatures and lower air densities result in reduced cylinder head cooling efficiency, the knock on effect being higher intake manifold temperatures and reduced combustion chamber mixture density. I could go on (and on), suffice to say that some of the items you mention (and many that you don't) will cause significant error to the point where the comparison between the results is fairly meaningless. One other example of potential variable error are the tire rolling losses. I once did a back to back test between two tires and found a 25hp variance at the wheels, the difference nothing to do with rolling diameter, pressure or temperature.

The trouble is that dyno testing is relatively easy, whereas the empirical analysis of the physics is anything but unless you nail down all unnecessary variables, but if nothing else comes from this discussion, my hope is that it might get more folk questioning information that is presented by vendors as verification of some notional performance gain when it comes from dyno tests that are not scientifically accurate.

f3nr15

At the risk of revealing my ignorance...

I don't think the results are meaningless.

I think that for street driven cars, which are driven in various states of tune, with different fuels, variable tyre pressure, and all weather, one could reasonably assume from the testing that a Fabspeed Cup Airbox is unlikely to result in a SIGNIFICANT change in performance.

For track based performance where you are trying to control as many variables as possible and extract every ounce of extra performance, the testing and results are less valid.

Just IMHO...

Robbie Ravioli

All very interesting and exactly what I expected to see.

The effect from the air filter is actually very easy to measure. The only effect an air filter / air box could have, is chnaging the resistance of the air passing it. On an atmospheric engine, this resistance has a direct effect on the power produced, as the manifold pressure under full load conditions (throttle wide open), would be reduced, according to the following formula:

Atmospheric pressure - pressure drop air filter - pressure drop clean air duct - pressure drop throttle - pressure drop manifold = intake port pressure.

As we all know -> more manifold pressure (intake port pressure) = more POWER.

Hence the only thing you do to prove if the difference in power came from the air filter / box is to measure the pressure before and after the air box (pressure drop across the filter) -> in the air duct before the air box and directly after the air box in the clean air duct.

At low air flows, the difference is not measureable, at full power (max airflow) I would expect to see a pressure drop of 40-60 mBar. If we go worst case -> 60mBar pressure drop on a stock filter, and assume 0 pressure drop for whatever the tuning boys come up with -> 60 mBar more manifold pressure. At nominal conditions : 60/1013 = 6% -> 250bhp * 6% = 15 bhp maximum potential.

This is obviously a not completly realistic calculation, but according to me, 25 bhp would be unrealistical. To achieve 25 bhp, the base engine needs to produce 416bhp (416 / 3.6 = 115bhp/L), which means the engine requires updates to a level the 997 GT3 4.0RS has (492 bhp / 4.0 = 123 bhp/L) -> a little more than an air filter.

A very good air filter has a pressure drop of about 35 mBar at full power, anything less and you might as well remove it completly, as it wont filter anything anyway. However, by changing the air filter / air duct, you also modify the resonance frequencies of the air passing the air duct.

As manifacterer actualy tune the volumes and the length of the air ducts to maximise the benefit of the resonances (Porsche especially, look at the 993 varioram engine) for torque and power, you may very well loose some power when changing the air duct. As these resonances are actually very difficult to measure on an engine, you need an engine on a engine dyno at stabil conditions, using in cylinder pressure transducers to optimise, expensive and difficult.

Further more, the engine management system cannot measure these resonances with the air flow sensor (not the volume flap in a 964, neither a hot film type i.e. Bosch HFM5), there is a map inside the EMS that predicts these resonances. When changing the air ducting, the map has to be modified as well, otherwise you'd have more air, but not the corresponding fuel amount and correct ignition -> no benefit.

Still sounds cool though -> perhaps use an sound pipe like Audi does in their V10 S6 -> connect the air box via a rubber hose to the interior -> great sound, no loss of power....

I will leave my stock filter in, and change it every 25k mls, but than, I am not such a good driver that I would make any difference if I had 250 or 255 bhp.

:-)

Cheers,

alexjc4

Great test

As for the physics; I can see two scenarios:

1. All the uncontrolled variables happen to have conspired to give an erroneous but coincidentally identical (within the measurement error) result pretty much across the rev range.

or

2. The results are in fact the same.

Occam says 2

rodders

... so it does make sense to use the K&K plus drilled box, just to listen to the induction - but dont spend any money on fancy kit.
It does add to the fun just to hear that lovely noise

c didy

This is valuable concrete data with attention to matching parameters. Thank you, NineMeister. The controversy surrounding restrictions on air intake has been examined many times and here is a nicely controlled evaluation.

Since there is a small but notable gain from removing the paper element filter your results beg the question: What filter/box combination minimizes intake restriction?

Those looking to maximize the return on investment should probably look for the answer to this over removing the barn door.

964Heaven

Mmmm

boxsey911

I agree! With an extra 5 bhp Kimi Raikkonen might get one of our cars round a track a couple of tenths quicker but it would make no difference to most of us.

August West

As much as I love Occam's Razor, I suspect the answer is somewhere in between 1 and 2. There definitely are a lot of variables at play and it could very well be a coincidence that the curves are so identical.

I'm glad to hear Colin state that 315 horsepower is possible with the stock AFM. I've been bit by the horsepower bug and would love to get 300 horseys at the wheels without spending beaucoup bucks on a Motec upgrade et al.

I've been eyeballing those Red Devil fuel injectors from fiveosports: http://www.fiveomotorsport.com/bosch-porsche-911

and will probably have to look at upgrading my headers. Of course a weight loss program is probably the most cost effective route at this point. Too bad my sport seats are so comfortable and that air-conditioning can be so critical at times.

Must. Have. More. Power.

NineMeister

Let's be specific. That's 315hp at the flywheel, through the stock air flow meter, not using it. The stock Motronic ceilings at 305hp and that's it, there is no cheap solution available that actually works, ok?