I've ripped apart dyno diagrams, decided they all have negatives, and am building one myself! Ultimately, the process is to practice Watts test for Horse drawn power, and never forget that you are replicating a tractor pulling sesion, not a sprint race.
The basic platform is this
http://www.steamengine.com.au/misc/horsepower/
Back to the real world
See at once
http://www.allstates.com/dynomometers.html
Do an image search in google, and use both
Engine Dynometer
and Engine dynamometer.
Go through all 16 pages of images, and make notes.
A few of my thoughts. I'm not a practioner, unlike
A7M, or some of the guys here like
The Frenchtown Flyer,
Mark P, etc, so take it all with a grain of salt.
Cheapest good stuff in Australia was Dave Bennets Go-Power Stone Bennett dynometer, a simple device made in the early 60's. Most other systems are marketing ploys, in my opinion. Taylors are one of the better heavy duty units of varying types. They seam to have a back-up. Super Flows look like a good peak power device, and it has lots of scope for a dyno cell and is the too for repeatibile peak power.
You need to discover what system suits you best.
You need to know what you are aiming for before you buy a thing. However, a dynometer is a sure fire way to make money if it can unlock inforamtion on transitiona drivability, and the spin-off of that should be the abilty to get the numbers on peak hp.
People are looking for drivability and a good BSFC curve, area under the curve torque, not peak power. Market boffins are looking for peak power. It is foolish just to get a dynometer with no flywheel becasue it only focuses in on peak load conditions, not transitional charecteristics. Companies are focusing so much on driveabilty these days, and some of the staber controls are very well aranged to control all the inputs
The set-ups I like are the US Taylors Dynometers, which are used by Ford Australia for outback testing to simulate grades and speeds up to 150 mph while sitting at 70 mph.
Superflo dynometers or dyno cells are just fine, but they all must allow basic loads in rpm each second to be controlled, and to allow situations like detonation on over-run to be expolered. This is the most critical component of the ability to deliver economy. I've seen engineers rabbit on about how propane is fine under load with heaps of advance, but then behave like a lean fired blowtorch when on over run. Dynometers with counter wieghts which allow coefficent of drag x frontal Area, road to tire friction, trans and diff loses, torque multiplication, and brake specific fuel economy to be run are better candiates.
In addition, the power absorbtion can only be controlled with heat control and /or correct volatage, physical inertia or some force by air, water and resistance measurement. Some dynos require over 100 US gallons of water per minute just to cool 500 flywheel horsepower. Some have heat exchangers with water recirculation, others have eddy current generators, and take a huge amount of cooling.
Some roller configurations are poorly designed, but have lots of flywheel weight. There are pluses and miniuses.
One other thing. Most attention is being leveled at pulsed fuel delivery (the method of making an EFI engine behave like an independant runner multi Weber set up) and ignition. An engine must be mapped to its transmission combination. An example is when you put an automatic transmission behind an engine mapped for a manual. The car invariably knocks its head off. People bad-mouth inertia dynometers, but it is certainly the best way to optimise and ignition system.
Larry Perkins, famed Aussie race engineer and touring car driver, said it best. I don't care if it makes 462 or 460 hp. If it makes 460 hp at 6500 rpm, then its better than 462 hp at 6900 rpm.
