The term "dyno tested" is used a lot these days, and most people tend to accept dyno results as absolute truth. Unfortunately, dyno results are no better than the testing equipment and method used to get the results. In other words, unscrupulous operators can make the results come out to be almost anything they want.

The real purpose of a dynamometer is to measure changes in performance and to tell the user whether modifications helped or hindered performance. In other words, a dyno is like a bathroom scale. It tells you whether you're gaining or loosing. Almost any dyno can tell you whether you're gaining or loosing, provided you make your comparisons using the same dyno, under identical conditions, or corrected to the same standards. You also need to use the same dyno operator, but that's another story. The problem comes when a dynamometer is used to determine a peak power level. Just like bathroom scales, different dynos are likely to yield different results. So, which results are accurate? Do you want a fudged result to make you shine or do you want to know the true horsepower your car puts down on the road. Those that wants a dyno figure for bragging rights beware, some one with a dyno sheet indicating 20 - 25 % less power might hand you your breakfast.

Both acceleration and sustained-load chassis dynos measure power delivered by vehicle's drive wheels. In theory, this makes more "real world" sense than simply measuring the net engine power output on an engine dyno. After all it is not the crank that's on the road but rather the wheels.

An acceleration chassis dyno, such as the Dyno Jet really ought to be named a "calculating dyno" since it attempts to measure power output by calculating power based on the amount of time required to accelerate the dyno rollers from one speed to another. This is possible when the weight of the rollers (actually the moment of inertia) is calculated into the equation, and mathematically it sounds good. The problems come with actual application and a number of inconsistent variables that cannot be built into the computation.

By its very design, a power test on an acceleration dyno is a very short test, lasting only a matter of seconds, and during the test, the vehicle's engine and drive train are in a constant state of transition. By its nature of transitioning from one speed to another, gear changes are usually required. This creates torque spikes when the shifts occur. At anything other than a direct 1:1 ratio in the transmission, the engine torque (power) is being multiplied, and an acceleration dyno has no way of ascertaining the transmission gear ratios of the vehicle being tested. If the vehicle has a manual transmission, there's the problem of gear changes that momentarily remove the drive force from the rollers, or worse yet, initiate some momentary tire-to-roller slippage. Automatic transmission vehicles also have the problem of the torque converter clutch being unlocked during acceleration. The amount of slippage in the torque converter, being a function of stall speed and load, is another inconsistent variable.

While the weight of the rollers and their resistance to acceleration must be known to calculate the power required to change the speed of rollers on an acceleration dyno, just as important is the mass of the vehicle's drive train, especially the drive wheels and tires. Obviously it takes more power to accelerate the wheels and tires on a dually than it does to accelerate single wheel/tire configurations, but an acceleration dyno has no way of accurately computing the moment of inertia of the vehicle's drive train. Consequently, the peak power number generated may be inaccurate.

Equally important regarding realistic acceleration chassis dyno results is to have the roller inertia weight closely equal the weight of the vehicle. If the effective weight of the rollers is 4000 pounds and the vehicle weighs 8000 pounds, obviously the vehicle's engine can accelerate the rollers far faster than it can accelerate the vehicle on the road or vice verse. Calculations can still be made when there are such large variances, but the more unequal the roller moment of inertia and vehicle weights, the more removed from a "real world" simulation the test becomes. If the moment of inertia of the rollers is substantially less than the weight of the vehicle, it is questionable whether the rollers can fully load the engine the way the vehicle would load it. On a turbocharged vehicle, full boost may never be achieved on an acceleration dyno, and that of course, results in lower power output. On an acceleration dyno, you simply can't duplicate the load a vehicle encounter on a continuous uphill grade etc.

Exactly where peak power occurs is another problem with an acceleration dyno. Because everything is in transition during a test, the RPM where peak power, or even changes in power, occurs can only be approximated. Still, in apples-to-apples comparative testing, an acceleration dyno can reliably indicate whether changes to the vehicle result in performance gains. Similarly, if multiple vehicles are tested on the same acceleration dyno under similar conditions, that dyno can usually indicate the relative power differences between the vehicles. The reason we say "usually" is that recognizing and quantifying tire slippage on an acceleration dyno is difficult to impossible. The real problem with an acceleration dyno is putting any faith in the calculated peak power number. These problems escalate in itself when tuning a vehicle for everyday driving conditions. Anything other than wide open throttle and the tuning results are very questionable. In many cases what appears to be a safe tune on an acceleration chassis dyno is every thing but that. It is virtually imposable to do a cell by cell tune on an acceleration dyno and in most cases the full midrange and under the curve potential of the engine can not be fully optimized.

A sustained-load chassis dyno does not calculate the test vehicle's power output. Instead, it measures power output directly by imparting an electrical load or water absorption load on the rollers and measuring torque. It can sustain this load indefinitely to allow conditions to stabilize on the test vehicle. It can take readings at any desired engine speed or roller speed to exactly determine a power curve and the peak power output RPM. The test vehicle can be locked in direct, 1:1 drive with the torque converter clutch (on automatics) locked to eliminate any torque multiplication or slippage. Similarly, engine RPM, wheel speed, and roller RPM can all be monitored simultaneously to immediately identify any tire slippage on the rollers. A sustained-load chassis dyno is simply more accurate.

On a sustained-load chassis dyno, the weight of the rollers has no significance since the load is usually measured at a steady speed with the load imposed on the rollers. This also means the weight of the test vehicle is insignificant. It could be a Geo Metro, Supra, Viper, Corvette, Ferrari, one ton pickup or a dually it makes no difference.

Perhaps the best way to explain the differences between acceleration and sustained-load chassis dyno operation is to envision the way they load a test vehicle compared to actual road loads. An acceleration dyno is like a drag strip, deriving its power rating from how fast the dyno rollers can be accelerated. A sustained-load dyno is like an unending uphill grade, measuring the power necessary to climb that grade at any given speed.

So why doesn't everyone use sustained-load chassis dynos for performance testing? There are many reasons, but there are two reasons that are important when it comes to testing tuning and aftermarket power products. First, a sustained-load chassis dyno is mechanically more complex and expensive because of the load generators and load-measuring units that must be connected to the rollers. These things, taken together, mean that the sustain-load dyno, and sometimes the permanent installation facilities required for it, is substantially more expensive than what's involved for an acceleration dyno. In fact, the total costs of a sustained-load chassis dyno may be more than five times as much as for an acceleration dyno. Second, many of the power products produced by various manufacturers for the performance market generate their best results during short tests. Let's put that another way; these manufacturers of quick and dirty power modules and programmers really aren't interested in revealing the problems their products cause under a sustained load.

In all fairness, the fore-going reasons why many manufacturers and tuning shops choose acceleration chassis dynos may reflect the old "chicken and egg" syndrome. If a company is unable or unwilling to invest in a sustained-load chassis dyno, then they also will lack the facilities to more fully develop their products for safe operation under sustained loads. Who knows which came first, an inadequately developed product or the inability to develop a safe, reliable product? Either way, let the buyer beware.

In searching for a quality dyno we at EngineLogics set the following criteria:

Accuracy.
Repeatability.
To tune al aspects below and above the curve we need the ability to hold at steady state under load with no wheel slip.
Worldwide Standardized factory settings (No fudging).
Cost
We chose the twin-retarder Dyno Dynamics - 4WD/AWD Lowboy 450 Chassis Dynamometer with its up to 2700 HP and 4500 ft lb torque capability.

(Thanks in part to Banks)