Originally Posted by Sticky
As I think you have indicated there are many variables. There are ways to cheat the dyno, we all know that. Whether you lower tire pressure, reset the ecu, move the fan, don't give a cool down, these all affect it.
The reason I give credit to someone like RPI is because of the way they dyno and test. The same reason I believe Active Autowerke and Evosport when they post numbers. They simply have too much at stake to inflate numbers to where they can't be reproduced. These companies have spent decades in some cases building their reputation.
I believe items like the RPI scoops make far greater gains in the real world than they do on the dyno.
As for your comment regarding the GPS and incline, the newer boxes give you the % of incline or of the decline. You will know exactly where you stand. I believe the GPS is the easiest and quickest way to see the acceleration difference. No, I won't have volumetric efficiency and exact air flow numbers, but I don't have those now, and what I am primarily concerned with is the acceleration numbers (differs for each person and their goal with the car.)
Like I said, combine all sources ideally. I'm sure even dataloggers have variables that influence them.
I would like to see where you believe you increased efficiency, some things could potentially carry over to the E92. It would at least be valuable to see.
I'm not even attempting to suggest that dyno operators are deliberately manipulating the results obtained on dynos. What I am saying is that static dynos will produce artefacts (inaccurate results) with genuine ram air-intakes unless used with appropriate variable speed fans which replicate the same air speed passing over the car as the wheel speed of the car. This is not a factor when data logging on the road.
I find it bizarre that most dyno operators give their results as "at the flywheel" figures when they actually record "at the wheels". Why not give the figures they actually accurately record? Likewise, whatever percentage they add to the "at the wheels" figure to give their "at the flywheel" values cannot be a fixed percentage to add to all cars of the same model, and even to the same vehicle with genuine engine performance improvements. For example: a OEM MC produces 325bhp "at the flywheel" and 275bhp "at the wheels" on a dyno: this represents an 18.18% increase on the "at the wheels" value. After modding my MC's air/exhaust system my car now consistently produces 321bhp at the wheels, as recorded by my datalogger, so if I add 18.18% to 321bhp, I get an "at the flywheel" value of 379.35bhp, whereas if I assume that the losses are still the same, then I should only add 50bhp to 321bhp to give 371bhp. As you can see, there's a difference of 8.35bhp (2.25%) between the two figures. Which is correct?
Again, the figures produced on dynos are generally variable upon the weather conditions at the time of the dyno run, yet many dyno operators don't attempt to standardise their results to account for altitude, barometric pressure, temperature, and humidity, and this means that dyno results, even for the same car tested on different days on the same dyno are not comparable. Whereas the results produced by my datalogger are standardised for altitude, barometric pressure, temperature, and humidity and give the "at the wheel" figure, so that someone testing a car in Mexico can produce the same standardised result that I can produce at an elevation of 100feet in cold and wet UK, so that we can accurately compare results.
Therefore if you acknowledge that a few percent here, and a few percent there, in all the potential errors that can creep into static dyno testing, as I see, then whilst you are attempting to be objective and scientific in using static dynos there are still too many possible flaws to make them as reliable as "on the road" dataloggers which use the cars ECU for their data. For me, it's a no brainer which is the most accurate method.
As for the mods to my MC that can be carried over: it seems to me that I've been unknowingly working in parallel to BMW and some of my mods have also been incorporated into later BMW's M car air intake designs. For example, the scoop in my MC's bumper is called a "brake duct" and is a straight-through funnel with a snorkel into it from the air-filter box. I discovered by taking temperature readings of my brakes, that this duct doesn't actually have any effect on brake cooling, so I decided to make a small blanking plate to almost completely block the back of the duct to make it into a genuine scoop for the air intake. I found that this increased the ram pressure inside the air-intake so that I got the same ram pressure inside my air-filter box at approximately 10mph lower than when OEM without the blanking. The first time I looked inside a Z4 MC's brake duct in late 2006, I was expecting to see a straight-through brake duct like on my MC, but in fact, there was a partial blank at the back of the Z4 MC's brake duct like mine - it's even an additional part fixed to the duct like mine, rather than a one-piece scoop.
I started modding my air-intake system when I saw that the inside of the air-filter box contained a trumpet which makes the air take an "S" shaped route inside the box before it passes through the filter (reverse-flow silencing). I cut this off so that the air can pass directly through the filter by the shortest route.
I've found the most successful mod of all though has been the gutting of the OEM silencers to remove the "reverse-flow" and make them "free-flow". This mod has allowed all my additional mods to the air-intake (pre-throttle) work better. My engine now feels so "light" and willing to pull throughout the entire rev range.
I've added an additional air intake duct into the system, so that I've now got three air intakes and not just the OEM two ducts. If you think that each of the mods might only add about 2% performance gain, since I presently have a total of seven mods to the system, you can see how I've obtained a 15% increase in power. If you think that a static dyno might have an error of around ±2%, then it would be possible to miss every single one of those mods, whereas the VE can spot them all.
Hope this helps.