Desk top Dyno real world results

Hey Everyone:

OK over the years everyone has been saying how the desk top dyno and other computer programs weren’t accurate enough to believe in ..

Seeing how over the years I’ve only ever had my seat of the pants and my Desktop to say how much my engines where making I now have real world numbers.

Well I just got my motor home for The Race Shop and have my dyno sheets in hand. I then plugged it into my Desktop dyno to see just how close this program is to real life.

And I always said the Desktop would have a 10% window. So my 950 H.p. guessament motors could be as low as 855 or as much as 1045.

Deck top says my latest build should be 958 h.p. @ 7500 Rpm and 730 t.q. @ 6000 Rpm
OK Real world.
The Race Shop DTS 4000 Dyno results 954 h.p. @ 7400 Rpm and 725.7 t.q. @ 6200 Rpm.

So I would have to say these programs are pretty dam good.

That's a great comparison Dan. Nice to see that one can get pretty close with a DD. One could plug the numbers into a DD and if the real world dyno is off, it's time to start checking where it's short.

The programs can be very accurate but you do have to put in the proper information and in detail.

Was it just the basic Desktop Dyno 5? Thats what I have, but there are so many different variations that aren't included in it.

i need this!!!!!

I say give or take 50hp on the bad side.

Northwest outlaw wrote:I say give or take 50hp on the bad side.

even if it is off that much, that's still good considering that you never have to touch a part or get out of the chair

i play with it all the time. So far this is the 528 twin im working on.

Everybody's entitled to their oppinion and the desk top dyno's do seem to tell the truth about whether you have made an improvement or a detriment; as to final figures, I don't have any faith in them at all as those have been so wrong, so often.
I'll believe the real thing which ACTUALLY measures the power coming from the crankshaft.

Northwest outlaw wrote:i play with it all the time. So far this is the 528 twin im working on.

Jeeeeeze, your gonna need all that tire......look at the torque at 2000rpm!!!

Bill

Northwest outlaw wrote:i play with it all the time. So far this is the 528 twin im working on.

Something's wrong with your simulation there? Look at the rpm where the horsepower and torque cross.

Yup. The HP and TQ curves should ALWAYS cross at 5252 rpm....since the formula for calculating HP is: Torque x RPM divided by 5252 = HP

Right now I have a 408 Cleveland build on the dyno. It's a pump gas engine for a Pantera that uses a dual plane intake and a set of CHI 4V heads. The customer has been using that program Dynomation to predict performance, work on camshaft selection etc, just for fun. What we've found is that even when using accurate data and a lot of that data the program has to make certain assumptions about the engine that may or may not be correct. For instance the engine makes 500 horsepower at 5850rpm in it's present state and around 480lbs/ft torque. The program predicted something like 512 horsepower at a very similar rpm which I say is pretty darn accurate. But it predicted torque numbers that were around 515lbs/ft which is somewhat optimistic. What's interesting is that the engine seems to out perform the simulation when it comes to making torque that's farther away from the peak in other words it's better than predicted 500rpm below and above what was projected by a small amount.

What I think happens is that the program overestimates the effectiveness of the tuning of the intake manifold. While this tuning is happening at about the rpm suggested by the program it's just not quite as effective as what's indicated by the software.

Another area that probably is overly optimistic is that it tends to assume that the valvetrain's behavior is more optimum than in the real engine or so I think. This engine has a hydraulic roller cam and this does cause some valvetrain geometry issues in the Cleveland. Nothing serious just because the cam is quite mild but things like this, control issues, lifter bleed down etc and their effect on the performance of the camshaft/valvetrain would be difficult or impossible to model. So the program assumes that the valvetrain performance is somewhat better than it really is in the running engine.

Dynomation is a very comprehensive program and the customer has used a LOT of data from other testing when essentially building his model of the real world engine in the computer. I'd say that without question this could possibly make the program give results that are closer to our real world build and if the program allowed less data input and thus relied on more assumptions about the actual engine it could possibly be quite a bit less accurate.