I have built a few evos in my time where the tyres and strapping techniques can vary the reading on a dyno by 110wkw.

As rob said, the condition of the knurling on dyno rollers come into play, tyres used (harder compounds will give a better reading than softer tyres, semi slicks and slicks are prone to overheating and sliming up on the dyno.)

Hubs dynos are accurate for calculating drivetrain loss, if you can run an engine on an engine dyno, and take tyres out of the equation.

Run a jockey wheel and do speed dependant boost control rob, then feed the rear wheels speeds into the ecu and do traction control vs the jockey wheel with a varying slip allowance. Front wheels can be unreliable, and gps is too slow to react.

I know exactly what you're talking about with the multiple wheel speed sensors and traction control, because that was one of the highlights of my university thesis, which helped me get my previous job at Adaptronic. 

People have also suggested boost by gear, which I can already activate in my ECU at any time (I already have gear detection working), but I enjoy the challenge of trying to handlle those things myself with the right foot instead of letting the ECU do it.

On the topic of tyres getting sloppy on the dyno...yep, my Federal 595 RSRs turned into a sloppy mess on the dyno. 

grosso
What size injectors do you run? Thinking of getting bigger ones, could keep engine temps down.

I have Injector Dynamics ID2000, which are much bigger than necessary. Installing bigger injectors won't change your engine temps unless you richen the fuel mixtures, and even then it won't make much difference.

robk
Remember that the 620 crank hp that I mentioned was assuming a little bit more than 378rwkw if there was 85% ethanol in the tank instead of 68% ethanol. The car was tuned on 85% ethanol, but that was before MRC replaced their dyno rollers to help eliminate wheel spin. With the new dyno rollers I've only had the car on their dyno with 68% ethanol at a dyno day and that's when it made 378rwkw. I'm sure that it will make about 385-400rwkw if I put it back on the dyno again with 85% ethanol and no extra tuning. The flex-fuel tune is already set up to automatically (and safely) increase the boost and ignition advance with any increase in ethanol content. 
 
I should be able to get over 400rwkw on E85 with a few more psi of boost and/or more ignition advance, which would be awesome, but I have to decide whether I want to put that extra stress on the engine and transmission. I've heard of E153 transmission gear teeth shearing off with about 450+rwkw, so I want to stay well away from that sort of craziness. If I had just a little bit more power then that would help me to get a 10 second pass at the drags, which is why the idea crossed my mind in the first place, but I should be able to do it without the extra power. I was toying with the idea of using my launch control button to also allow some more power to be applied only while the button is pressed. 

Indeed, fair enough if that's the figure you feel comfortable with. Well, to get to just above the Veyron on that list, you only need 646bhp. (Veyrons are indeed heavy!) So I think you're safe there, it's probably only 1-2psi or a couple of degrees of timing more and the gearbox is still in safe sub 400kw territory ;)

Im personally not a fan of boost by gear, it can be a bit of a killjoy. If you have got your boost running closed loop, and have a decent gauge of base duty vs boost pressure, you can always run throttle dependant boost. Set up your aim table in throttle vs rpm, set the values for you aim boost, and in you base duty/normal position table, set up your rpm/boost axis and plug in your base duties, make sure you dont have too aggressive pid settings though, it can end up making it more of a handful.

That said, if you have ethanol based boost aim, to do throttle dependant, you need to make your ethanol content a boost aim trim.

I just find as you have boost aim tied in with throttle, you can keep the car under control a lot better, think of it as a medieval torque request table.

Its a bit of work, but easier than doing steering angle based boost control on a fwd car haha. (Awesome when it works, but a pain to optimise aim values)

I have a Tial 44mm wastegate and it seems to be the perfect size for my application. 

grosso

robk
I know exactly what you're talking about with the multiple wheel speed sensors and traction control, because that was one of the highlights of my university thesis, which helped me get my previous job at Adaptronic :-D

People have also suggested boost by gear, which I can already activate in my ECU at any time (I already have gear detection working), but I enjoy the challenge of trying to handlle those things myself with the right foot instead of letting the ECU do it.

On the topic of tyres getting sloppy on the dyno...yep, my Federal 595 RSRs turned into a sloppy mess on the dyno lol.

grosso
What size injectors do you run? Thinking of getting bigger ones, could keep engine temps down.

I have Injector Dynamics ID2000, which are much bigger than necessary. Installing bigger injectors won't change your engine temps unless you richen the fuel mixtures, and even then it won't make much difference.

It was leaning out at the top off the rpm range due to waste gate being to small. Did you say you use 50mm gate? If I put a bigger one on I could run lower boost, so I could do 7 laps without over heating the cooling system on a hot day.

If you are leaning out, then your tune is wrong

That's true though my point is that the tune should take into account things like this - i.e. it should back off timing when the injector duty cycle reaches 100% to save the engine.
A tune isn't just making sure the mixtures are right and that you get maximum power. It is equally important for the tune to accommodate failure scenarios - high intake temperatures, high oil temperatures, 100% injector duty cycle, high coolant temperature etc.
Even though I have a basic Microtech, we have as much of this programmed in an possible. As soon as things get outside of operating range, timing is reduced and fuel is added. The engine will run terribly in this condition though it will not blow up!
That was my point... :-)

Correct, though best to dial back the tune instead of KABOOM!
Rodney's tuner shouldn't have given it back to him with injectors running at 100% duty cycle..... Not cool