Forty years ago when I first started working with engines, BHP figures were something you made up. Basically, since only car manufacturers had engine dyno's, the rest of us had to “guesstimate” our power outputs. I f another car was faster than ours then they obviously had more bhp. Call me cynical but not a lot has changed over the years. People still make up their power figures (guesstimates often based on other people’s claims) although to be fair some just pluck them straight from the air. When rolling roads came along that was all set to change. Now we could measure the power of the engine by taking a bhp reading from the driving wheels. The trouble was, back then none of us knew what we were actually measuring and today many still don’t.
MAHA were probably the first volume rolling road supplier, selling their machines through Sun in the U.K. The RAM 11 was a good solid piece of kit but not without its problems. Operators found that by running in a different gear they got a different result in terms of power at the wheels. Also if you were short of power you could get a fair gain by inflating the tyre pressures! Keep reading to find out why.
Emerald's Rolling Road
Emerald's rolling road is a 4-wheel-drive Sun Ram 12. Our rolling road is used in two ways, steady state mode for mapping and also in acceleration mode which allows the computer to produce repeatable power curves. The rolling road measures a maximum of 330bhp at the wheels up to a speed of 160mph. Depending on transmission/tyre losses this equates to a maximum flywheel figure in the region of 360 – 390bhp. I f necessary these figures can be doubled but any engine making over 500bhp will have traction problems. The software logs the power output at 1mph intervals and can measure the power losses through the transmission and tyres using a coast-down test. The results can be plotted and extracted for further analysis. Comparison graphs can be produced between several runs and these assist the operator when making adjustments to the engine. As well as the traditional analogue dials, the rolling road has a digital read-out which is used heavily when mapping. To date we have mapped many systems, and not only our own. Unfortunately at present we are unable to take on mapping work for systems other than Emerald as our diary is nearly always booked several weeks ahead.
How Rolling Roads Work
Before we consider how a rolling road measures power output, it helps to understand what we are measuring. We often talk about bhp (brake horse power) as if it was something that existed, rather than the reality of it being a convenient number - which we have calculated. Bhp is a rate of doing work and in order to do work you need to put in some effort and then you need to see a result for that effort. The effort is, in our case, the torque generated by the engine. The result is the distance moved by the flywheel (expressed in rpm). The simple formula for calculating bhp is based on 33,000ft lbs of work being done in one minute. This amount of work is regarded as one horsepower. The formula is: HP = Torque x rpm over 5252. From this you can see that when the rpm is at 5252 the HP and the torque are the same. If the power and torque lines do not cross here on the graph then someone is telling you porkies! We use the term "Brake" horsepower because the engine torque is measured on a device called an Engine Brake, or dynamometer as it is more correctly known. Dyno's measure torque at a given rpm and then we calculate the bhp from there. With an engine bolted to a dynamometer we take the torque reading directly from the flywheel and without any form of gearbox.
To coin a computer phrase: "wysiwyg" (what-you-see-is-what-you-get) in an ideal world you can measure the power at the flywheel and get the same result every time – but this seldom actually happens because we do not live in an ideal world. The problem is that we are dealing with an engine – which is a dynamic device. If you run the engine up to the test rpm quickly and take a reading, (known as a flash reading) it will be higher than if you let the rpm stabilise and then read the load. What happens is that the longer you run the engine, the more heat soaks into the combustion chambers, the spark plugs and the inlet manifold. The power drops off a little as a result. The traditional method of power testing automatically stabilised the engine before taking a reading because you had to set the load, and hence rpm, manually.
In the old days...
It worked like this. As the operator you open the throttle and the load holds the engine rpm back. You then back off the load and as the rpm climbs the engine suddenly comes on cam - and the revs soar. Quickly you wind the load back on and get the engine down to the target rpm where you want to take a reading. You take a reading off a spring-loaded scale and then measure the engine rpm. Now you close the throttle back to idle, write down the data and think about the next reading that you want to take. Computers changed all that. Now you can have the load controlled by a computer and this can put the load on, and off, so fast that you can hold almost any rpm regardless of what the engine is naturally trying to do. You can also tell the computer to let the engine accelerate at a given rate and you can record a whole power curve in a few seconds. The end result is something of a flash reading but for comparison purposes there’s nothing wrong with that.
BHP at your wheels
A rolling road takes its power reading directly from the driven wheels of the car. This means involving gearboxes, drive shafts, differentials and tyres. A lot of people talk about bhp @ the wheels as being the only meaningful number to quote: "It’s what you race with" they will tell you. In a way that is correct, but then the gearbox gets in the way of the true picture. We measure the torque at the wheels but the rpm is measured at the road wheel roller. Put the car in a lower gear and the torque at the wheels increases – but the rpm of the roller is reduced. In theory the resulting bhp should be exactly the same – but it never is. The lower the gear that you run the car in, the higher the bhp at the wheels. This is because we have rolling losses (some call them transmission losses) that increase with increased roller (and hence road wheel) rpm. The biggest single rolling loss is the tyre.
Remember that you have not one, but two contact patches on a rolling road. The tyre is compressed in two places and the faster it spins the more often it is compressed. The tyre construction, the diameter and the tyre pressure all have a direct influence on the rolling losses. As an experiment we measured the power at the wheels of a Golf GTi. Then we put another 10 psi into the tyres and checked the power at the wheels again. The power went up by 4bhp! Can you imagine what happens to the rolling losses when the tyre is compressed by several bodies sitting on the back of the car trying to find enough grip to prevent wheel-spin? As long as the bodies stay on the back during the run-down, which measures the rolling loss, you get the right result in the power graph. If the bodies all jump off when the car is knocked out of gear and allowed to run-down, you lose the tyre compression and the losses are less – distorting the resulting graph plot. In order to make any sense of rolling road power figures you must measure the rolling losses and add them to the power at the wheels. When you do that you can run in any gear and get the same result on the power graph – almost.
The way forward
More modern rolling roads tend to have one very big single roller and you park the car on top of that roller. The main reason for this system is so that you get less tyre distortion with a single large diameter roller. With an older system using twin rollers, you get more rolling losses because you have two contact points, not one. Car manufacturers like Mercedes who make powerful engines and very heavy cars will only allow their dealers to have single roller dyno's to limit the tyre loading – it’s a safety issue.
Several factors prevent you from getting exactly the same result in every gear. First off a lower gear means more torque at the wheels and hence a little more tyre slippage than when you run in a higher gear. The run also takes less time, so the engine accelerates faster and gives you more of a "flash" reading. Our Sun Ram 12 rolling road allows us to alter the acceleration rate so that we can adjust it for different power outputs. The software in our system uses the road speed, measured by the rear roller, to obtain engine rpm in order to scale the power curve. We take an rpm reading at 60mph and the software works out the revs at any given road speed from there. What this doesn’t take into account is tyre growth. As the revs increase the centrifugal force makes the tyre grow – which alters the gearing slightly, putting the rpm out by a tiny amount. When you take all these "fudge factors" into account, it’s a wonder the rolling road is as accurate as it is. But it can be accurate, and more importantly, repeatable.
With careful setting up of the acceleration rate to match the engine power, and accurate setting of the engine rpm, (dashboard tachometers are often out), you can get a meaningful number from a rolling road. I know that when trying to improve the engine in the 'Red Shed' our rolling road is depressingly accurate enough to give the same power curve time after time – despite my best efforts to increase the power output! I always call our final figures "simulated" flywheel figures but they are close enough to engine dynamometers judging by the comparisons we have available. Ken Snailham at Q.E.D recorded 218bhp on his dyno and the same engine showed 220bhp on our rollers. We've had similarly close results to the Lotus Service Centre dyno and J.E. Engineering’s dyno. We also see close to factory-quoted power outputs on most standard cars that we have run in the past.
A big problem with rolling roads is tyre grip. As power outputs reach the levels where you question the sanity of the engine builder it gets harder and harder to put the power down. On a rolling road you usually have more grip than you do on the road or track but I leave problems like that up to the driver – the throttle tends to work both ways after all. The answer would appear to be to eliminate the tyre from the picture. A hub dyno bolts directly to the drive shafts so there is no slip at all. This, on the face of it, would appear to be the answer to all our problems. But you still come back to the question of what you are measuring? By removing the tyres from the picture you get much lower rolling losses but not enough inertia in the system to measure any sort of run down. For setting up it’s probably the perfect answer but not much use for comparing one engine to another as you can’t get back to what the engine is making at the flywheel.
Mapping with a Rolling Road or Dynamometer
For mapping, the rolling road’s acceleration mode is less useful, apart from for part and full throttle runs. You have to be able to switch to fixed (constant) rpm running in order to map and engine. This means operating in closed loop. In closed loop mode the dyno’s absorption unit holds the roller rpm regardless of load. Without a closed-loop rolling road the operator is required to dial in the load in order to hold the rpm. Having done so, the first change to the map that increases engine power also moves the engine to a different speed and load site on the map. The operator has to re-adjust the load to get back to the target cell. I’m not saying that it can’t be done, but it really needs two people and a lot of time to get a half-decent job done. Letting the electronics do the donkey-work is a lot faster and a lot more accurate. You use the acceleration run for full throttle comparison after the main mapping has taken place. After all, you can’t do full throttle runs until the engine is close to being correctly mapped in the first place.
Finally - the arguments for and against dynos and rolling roads... I have worked with both and if you want to develop an engine, the dynamometer is the place to do it. It’s more accurate and a lot more convenient to work just with the engine and nothing else. But for final mapping I believe the rolling road is the place to do it. The engine, including all ancillaries, inlet and exhaust systems runs exactly (wind and ground-effect apart) as it is going to on the track. For more on Dynamometers see our separate dedicated section here - Engine Dyno