Separating the truth from wishful thinking, one pull at a time
Over the past few decades, dynamometers, which are devices that measure horsepower and torque, have gone from equipment only used by car manufacturers and high-end engine builders to an indispensable tool for performance enthusiasts at every level. Whether you’re getting a tune to dial in a new turbo, looking for bench-racing bragging rights with big numbers, or just enjoy watching videos of other people’s cars and trucks fail spectacularly, dynos have become a prominent part of the automotive world.
Unfortunately, dynos(and the information they provide) are widely misunderstood. Our goal today is to brush away the cobwebs of ignorance and replace them with the cobwebs of knowledge, by explaining how they work, what they measure, and why certain things everyone ‘knows’ are either half-truths at best, or just plain wrong at worst. Strap down, because this is going to get bumpy…
Defining Terms And Busting Myths
It’s very likely you have heard the cliche, “horsepower is how fast you hit the wall, torque is how far you move the wall.” This is a terrible analogy, for many reasons, and if we catch you saying it after you’ve read this article, we won’t be mad at you – just very disappointed. While they are mathematically linked to one another, they are measurements of two different things. Torque is force; specifically, of rotational force or more simply twist. Horsepower is the rate at which work is done. Note that in order for there to be horsepower to measure, something has to move, while torque has no such requirement, and in fact horsepower can be thought of in highly simplified terms as how quickly torque is applied over a distance.
Way back in 1776, while America was getting its revolution on, Scottish engineer James Watt had invented a new kind of steam engine that was a lot more powerful and efficient than the ones that had been around since the 17-teens. Watt wanted to sell his new engine to mine owners in order to replace draft animals for tasks like running pumps and hoists. For marketing purposes, he rated his engines in terms of ‘horsepower’ which he defined as the ability to lift 550 pounds one foot in one second after some experimentation with the actual work done by draft horses. As it turns out, horses can actually deliver nearly 15 horsepower for short periods, but for sustained activity over an entire day, it’s actually pretty close to what can be expected.
His work with steam engines (and his marketing savvy) led to the eventual adoption of Watt’s name as the metric unit of work – one horsepower is equivalent to roughly 745.7 watts, because nothing can ever be easy in converting back and forth between the units we’re familiar with here in the US and the metric system. Regardless of whether it’s being recorded in horsepower or watts, the actual thing being measured is how quickly work is being accomplished.
Torque, on the other hand, doesn’t require anything to actually move. Clamp a bolt in a bench vise, put a wrench with a 12 inch long handle on it, and hang a one pound weight from the end of that handle – congratulations! You’ve just applied one pound-foot of torque to that bolt. Hang a second identical weight, and you are up to two pound-feet. Keep the same weight, but double the length of the wrench? Also two pound-feet. Note that nothing actually has to move in order for there to be torque; this is where a lot of people get things mixed up in their heads, and why the cliche we mentioned is so useless in understanding torque versus horsepower.
To keep this a reasonable length, and avoid losing you with a bunch of math, we’re going to skip over all the boring stuff, but know this: The formula for calculating horsepower is as follows…
Horsepower = (Torque x RPM)/5252
You’ll notice that we have another value in there for RPM, which tells us how quickly the torque is being applied over time. The 5252 constant is derived from Watt’s original calculations and experiments, rounded off a bit, but it’s the universal standard way to calculate the output of a mechanical device. Because of that constant, every dyno chart you will ever see that plots horsepower and torque on the same scale will have those two lines cross at exactly 5252 RPM. It also means that once you know horsepower and RPM, you can calculate torque, and vice-versa; any two of the three values will get you the third one with just a little algebra, and this is one of the keys to how dynamometers work.
Spin The Wheel
There are two main types of dynamometers – engine dynos, which measure horsepower and torque at the crankshaft, and chassis dynos, which read output at the wheels. For most of hot rodding history, engine dynos were far more common because they are simpler to build and install than a chassis dyno, but the ability of a chassis dyno to test a vehicle without first removing the engine from it has caused their popularity to explode over the past few years.
One old-school term you may have heard is “brake horsepower” and this phrase comes directly from the way an engine dyno works. The engine is set up on a stand, ready to run, and the crank is connected to the dynamometer load cell and brake. The brake does what it sounds like it does – it resists the turning motion of the crankshaft and absorbs the power being fed into it (usually by pumping water that circulates through a very large reservoir outside the dyno room), while the load cell measures how much twisting force is being exerted on the brake. When you know that force (the torque being produced) and the engine RPM, you have two parts of the equation and you can solve for horsepower.
As you can see, engine dynos of this type directly measure torque, and require an RPM input in order to calculate horsepower. Chassis dynos are another story – while some include a load cell, the most common type is an ‘inertial’ dyno. These actually directly read horsepower, and run the math in the other direction to calculate torque based on engine RPM.
In an inertial chassis dyno, the vehicle to be tested is strapped down with the drive wheels in contact with a drum or a set of rollers. These rollers have a precisely-measured diameter and mass, and the computer that controls the dyno uses that information to read actual horsepower being delivered by measuring how quickly the vehicle accelerates the rotating mass of the rollers. The faster it speeds up during a pull, the more work is being done per unit of time, and the more horsepower is being produced. Add in an RPM signal from the engine, or calculate RPM by knowing the tire diameter, final gearing in the differential, and transmission gear ratio, and the dyno can calculate torque. Sometimes you’ll see a chassis dyno sheet that just shows horsepower versus MPH – this is a bare-bones way of testing that doesn’t require any calculation of gearing or engine RPM input, but it also doesn’t show any results for torque.
Here’s Adam LZ using a chassis dyno to get hp ratings for his S15:
While an engine dyno with a sufficiently robust load cell and brake can hold an engine at wide open throttle at a steady RPM, a pure inertia chassis dyno has to accelerate through the RPM range being tested in order to work. This is fine for most tasks, but for a tuner who wants to concentrate on specific engine speed ranges, most chassis dyno manufacturers offer upgraded models with load cells in addition to inertial operation. By attaching an eddy brake (a kind of simple and durable electromagnetic device that converts energy into heat) and load cell to the drum or rollers, the computer can receive input of both torque from the brake and horsepower from the inertial acceleration of the drum or rollers. This also can reduce the necessary size and weight of the drum or rollers, making it convenient for mobile dynamometers or shops with limited space.
It’s also worth mentioning that there are also chassis dynos available that use separate load absorber units that bolt directly to the wheel hubs, and operate in the same way as an engine dyno but with an electromagnetic instead of hydraulic brake system. These are very popular for all-wheel-drive cars and for situations when the user doesn’t want to permanently devote floor space to a lift or in-ground installation.
What’s The Difference?
You would think that precisely-calibrated machines designed for the single task of measuring the output of a vehicle’s powerplant would settle arguments, not cause them. But if you’ve been around car guys for more than a day, you know that there are endless disagreements about the validity of engine dyno testing versus chassis dynos, one dyno brand versus another, and even different operators of the same kind of dyno.
Back when color TV was new and we were putting boot prints on the moon, American auto manufacturers were involved in a war of one-upmanship when it came to the claimed power output of their hottest engines. To get the most impressive numbers possible, engines would be dyno-tested without any of the power-sapping components they’d need to actually be practical in a car; things like water pumps, alternators, fuel pumps, or exhaust systems and mufflers. Referred to as “gross horsepower” this was really more like “horsepower at the brochure” instead of at the crank, because it was wildly unrealistic.
Since these power and torque claims were basically useless to consumers, by 1972 the standard had switched to SAE (Society of Automotive Engineers) net horsepower, which defined testing conditions and required engines to be configured with the same components they’d have when they hit showrooms. This is still the horsepower figure used in US advertising, and it’s important to remember that it still doesn’t take into account any frictional losses between the flywheel and the tires – things like transmissions, differentials, driveshaft bearings, and other necessary but less-than-100-percent-efficient pieces of hardware.
As a result, it’s typical to guesstimate that the readings at the tires will be 10 to 15 percent lower than the crankshaft output when the transmission is in its most efficient 1:1 gear. Some will argue that because of the variable drag of the driveline, the only fair way to compare engines is at the crank, while others rightly point out that nobody drives or races an engine dyno. In any case, you will often see articles testing new cars on a chassis dyno, then comparing the results plus that estimated loss to see how closely they agree with the advertised SAE net rating.
It’s also very common to see big differences between the same car on different chassis dynos. Some brands have a reputation as being ‘stingy’ or ‘generous’ compared to others, and there are tricks a dyno operator can employ to bump up numbers, should that be desired. It’s been said that the man with two clocks never knows what time it is, and that same concept applies to dyno results. Unless they’re on the same machine, with the same operator and the same correction factors for temperature, humidity, air density, and load cell calibration and/or inertia factor, it’s impossible to make a direct comparison. Where dyno testing really shines is when tuning or checking the improvement (or lack thereof) from modifications, when it’s possible to get meaningful, consistent results.
Hopefully, we’ve cleared up some misconceptions and helped readers to understand how dynos can be useful tools as opposed to just another source of arguments with no possible resolution. Just because James Watt used fanciful horsepower figures to sell engines to people doesn’t mean we should still be doing the same thing centuries later…
The AWD Dyno Dilemma
An increasing number of high-performance vehicles are being equipped with all-wheel-drive in order to maximize traction and get as much power to the pavement as possible. Unfortunately, this can lead to issues when trying to test certain makes and models on a chassis dyno. The first and most obvious one is that the majority of dynos have only a single drum or set of rollers, and trying to test a car that powers both ends with a system that is only set up to let one end turn is gonna make you have a bad day. Unless the vehicle is among the rare few with AWD that will let you safely detach a driveshaft and only spin one pair of wheels, you’ll need to have more than one roller.
The second issue is that cars have different wheelbases, so one of the two rollers is going to need to be adjustable to accommodate variations between axle centerlines. Pretty much every chassis dyno that has AWD capability has this handled, but it brings up the third issue – whether the front and back rollers are physically connected and always turn at the same speed.
Example of an AWD dyno setup with Rob Dahm’s 4-rotor RX7 FD:
The simplest form of AWD dyno is just two inertial rollers with one on an adjustable track to set the correct wheelbase. But some vehicles, especially those with sophisticated electronically-controlled center differentials that can change the power split between axles on the fly, will absolutely lose their minds if they detect a big enough difference in wheel speed front to rear. This is often addressed by activatingf a specific “dyno mode” in the ECU, either baked in from the factory for diagnostic purposes or as part of an aftermarket tune, but there are a handful of cars that can actually damage their transaxles on a dyno with unconnected rollers.
The most complicated and expensive solution to the problem, but also the one that eliminates all of the major problems with AWD dyno testing, is a system that mechanically links the front and rear rollers via a driveshaft or chain, regardless of the wheelbase setting. This type of dyno is also useful for the rare 2WD vehicle with an ECU that throws a fit when the undriven wheels are stationary but doesn’t have a dyno mode to disable that failsafe.