So today I Dynoed my Z3 at my engineering school! The dyno that was used was MAHA 3000, the teacher said that it gives fairly conservative power figures and that the indicated power from the engine is much more accurate than the indicated wheel horsepower because of the way it measures it. This dynometer measures all the losses occuring between the engine and the wheels independently after the dyno pull (you put the clutch in and keep it in gear) you can see it as the green line in the picture. I also asked about the DMF thing and the teacher said it shouldn't effect drivetrain losses, it simply changes the place where torque will occur.
Sorry that all the text is in finnish, there is a translation at the end of the pictures 
I also have an Excel file already made if you want to plot a comparison from it Lambertius!
If I forgot to mention something, ask away!
The mods that I have on my M44 Z3 (156 000 kms driven):
Fogged airbox mod
Fulda Steel 4-1 style collector 2"
2,5" Custom made mid-section pipe
2,5" Powersprint high flow catalysator
2,5" Simons catback exhaust
Exhaust manifold
Dynoresults
Translations:
Normaaliteho, Pnorm = The power from the engine that has been corrected with ETY-norms (so not the real one!)
Moottorin teho, Pmoott = The Engine power
Pyorateho Ppyora = The wheel horsepower (not to be blindly trusted according to the teacher)
Havioteho Phavio = The measured horsepower lost from the drivetrain, wheels etc.
Suurin teho = biggest power @ 6190/172 km/h
Vaantomomentti = The Torque (this is also corrected to the ETY-norms, so out calculating the norms the real torque is 189 Nm)
Suurin saavutettu RPM = The highest achieved RPM
There are a lot of interesting things to cover with your results, but before we get into that:
- The DMF will affect drivetrain losses, not because it is a DMF just because it is massive. I only raised it to illustrate that the low wkW from our cars is likely a result of the drivetrain designed for much torquier 6cyl engines
- According to your dyno results, the OEM drivetrain losses are ~25%. I'm going to work with that value in the future (until someone can demonstrate otherwise) simply because it is close to what we would expect.
- Different brand dyno, different country. I'm going to be making illustrative assumptions. If anyone takes the following as being tacitly true, don't.
Below are the wheel results from Puksuttaja's excel file:
Its hard to know how accurate any dyno readout is compared to another - but let's just for the sake of interest assume that this is accurate within a reasonable error range (<+/-2.5%), it makes the results more interesting.
We know from my own testing that the muffler being changed resulted in a 6% improvement in peak power, so it would stand to reason that a whole exhaust system could result in more. Since Puksuttaja hasn't fiddled with his intake, it is possible that he wouldn't see the low-end losses that I did initially (till I reduced the pulse chamber volume).
So assuming that his engine was making 98kW~103kW (assumed losses over time of 5% and OEM rating), and the dyno he used is accurate - then the peak power improvement for Puksuttaja would be 9%~14%. Now this is based on assumptions, but they aren't unreasonable when considered alongside my own recordings.
Below is a comparison against the results of my own car. Unfortunately I can't really correct the results from the two dynos because we don't have an OEM recording to try and draw a comparison with. What we can do is look at the behaviour of the curves, which we can learn from.
Ignore the numbers - they mean nothing in this case.
Puksuttaja's curve is actually very similar to my OEM curve in shape - with the notable exception of the large torque spike in the middle. This spike is
exactly what I would expect to see from a 4-1 header. You should be able to see the point with the ideal torque response, and that is exactly what you see. Then like my OEM run (and unlike my ITB run) the power and torque begin to drop then level (presumably from the DISA). The restricted volume of my pulse chamber combined with the long runners in my ITB setup create a very flat torque curve - a friend who has been minding my car actually commented how linear the engine response was when he had to drive it the other day, that you could easily run it to the red line without realising it because it never feels choked.
Unfortunately there are a lot of assumptions involved with all this, but it doesn't seem to be unreasonable to me to assume that Puksuttaja's exhaust setup could be worth 9~14% improvements over stock on its own, simply because I know the muffler on its own is worth 6%. This of course leads to the potential for extraction + induction improvements in the range of 25~30% in peak power. The only thing I would be cautious of is with enlarging the exhaust diameter and running a large volume pulse chamber that you may be optimising the entire system for a rev range in may not ever reach. In Puksuttaja's case he still has the low volume of the OEM intake, which may compensate for lower end-losses from the enlarged exhaust diameter.
Regardless, I would really want to try a set of headers and a 2.5" exhaust with my induction kit - plus the lightened flywheel I'll be ordering soon!
Should make a significant impact on drive train losses!
