Quote:
Originally Posted by wstar
You know exactly what the testing methods are, basically none. They're randomly inaccurate. Go to dyno shop, take a few pulls. Get more mods, go back to same dyno shop, take a few pulls. Repeat. We know they'll have random inaccuracy, but that's a completely different thing than fabricating the results or trying to ignore bad numbers and/or post ones that are known to be wrongly high.
The fact that the randomly-inaccurate dyno results of multiple members in different states are showing results in the ballpark of each other validates the general idea here, which is that it's not crazy to believe 30+ rwhp from a full exhaust on this car.
Your testing methods, while obviously more self-consistent, aren't going to be perfect either. They may be more accurate than random dyno shots, but there are always uncontrolled variables.
|
I agree with you 100%, through careful conditioning of the vehicle you can cut down on the variabilty quite a bit. Some of the controlable things I have found are variance due to oil and coolent temp of powertrain, +- 3 hp. Heat bloom transients in the engine compartment +-6 hp(amazing isnt it?), electrical load due to fans +- 1 hp.
In my experiance there is also usualy a random variance of +-2-5 hp (depending on the car) for factors that are not easy to control or explane. This makes the testing quite imperfect. I also feel that SAE correction is imperfect, espcialy when heat is combined with extreme humidity.
In my opinion dynojets have better repeatabilty than load dynos (I suspect that the elertical resistance changes in the coils with heat on eddy current and oil properties like viscosity with temp change on hydralic loaded dynos. Dynojets run "shallower" into the high load cells on the maps and don't accurately represent what loads an engine really sees, one reason why they are not optimal for tuning.
I feel that dynojets are the most repeatable but the least accurate for this reason if you know what I mean.
So I think its important to try to be as accurate as posible but as you said, no results are going to be perfect. By being careful you can eliminate more than half of the variabilty which is significant.