# is it possible to install some device which will measure in real time the torque my diesel engine is providing at the moment?

It is easy to make a meter which will show the engine RPM, many cars have one installed. But to know the power in kW that the engine is providing right now, I also need to know the torque, because:

$power = RPM * torque/9548.8$

Besides RPM all in metric units (kW, Nm). BTW, does anyone know how the 9548.8 number is derived?

But back to my main question: is it possible to install some measuring device which will measure and show in real time the torque my diesel engine is providing at the moment? I want to know how many kW is used to move the car in real-time. I want to do something like this with Arduino.

• Let's take a look at the constant. Power in this sense is moment of force times angular velocity. A problem here are RPM. Let's turn them into angular velocity. So we have number of circles that is made each minute. A single circle is $2\pi$ and each minute has 60 seconds, so formula would be $\omega=\frac{2\pi RPM}{60}$. When we calculate that, we get it to be approximately $0.1047195512*RPM=C*RPM$. Now lets turn multiplication into division. $C^{-1}\approx9.54929658548$. And since we're using kilowatts, we divide everything by 1000, so in the end we get around 9549.29658548 – AndrejaKo Apr 5 '11 at 19:08
• So I guess that the number comes form someone approximating it at one point. – AndrejaKo Apr 5 '11 at 19:10
• 9548 is 1000 x 60 / (2 * pi). You divide out the 2pi/60 to change RPM to radians/sec, and dividing out the 1000 gives you the answer in kW rather than just plan watts. – JustJeff Apr 6 '11 at 1:00
• @AndrejaKo - That should really be an answer! Unfortunately, while the mod tools allow me to convert answers to comments, I can't go the other way around. – Kevin Vermeer Jun 7 '11 at 22:10
• @Kevin Vermeer♦ You're right. I decided to post it as an answer. This way, the TeX got fixed too. – AndrejaKo Jun 7 '11 at 23:03

The power the engine outputs at any given time will vary with the load on the engine.

The load on the engine when in normal operation depends on the drive-train efficiency, gear ratio, wheels slippage, the weight of the car, inclines in the road surface, etc.

Real Dynos put a known constant load on the wheels which allows them to calculate horsepower at the wheels.

Doing this just for the engine would require putting a constant load on the engine, which would mean removing the engine from the car and applying a constant load to the flywheel or having a good estimate for the drive-train efficiency, but that tends to vary over various parameters.

The best your going to get for a car on the road is an estimation of wheel horsepower. You can make a decent guess based on the weight of the car and its acceleration. You could try to make some sort of correction for inclines/declines based on another accelerometer axis.

If you wanted to get a guestimate of the flywheel horsepower you can include a factor for drive-train efficiency and adjust for the current transmission gearing. Alternatively you can estimate based on drive-train efficiency and engine RPM.

If you wanted to contrast that with the actual energy released in the engine, you can calculate that from fuel flow, amount of air taken in, and measuring residual oxygen in the exhaust using a wideband O2 sensor. You'd find that a LOT of energy goes straight out the tailpipe.

• Thanks for lots of info. However guess/estimation is not useful for me, because what I want it for is after installing the torque sensor measuring power to also install a fuel flow sensor, and based on that measure the liters/kWh of generated energy for different powers, loads and RPMs to be able to find out at which combination my engine is most efficient in converting fuel into kWh. – miernik Apr 5 '11 at 19:56
• @miernik Unfortunately there is no reasonable way to do what you want, there is a reason that dynamometers are expensive, calibrated devices. The only other option is to instrument your vehicle to the level that F1 cars are, but the sensors and modifications alone for such an endeavor would cost many thousands of dollars. The best your going to get is an estimate of wheel horsepower. – Mark Apr 5 '11 at 20:27
• @Mark: another thing came to my mind: is the torque a given engine produces at a given RPM always the same? Or does it varies while I drive? If it would be always the same, then if I get a graph/table of power or torque vs. engine RPM for my engine, it is enough to measure the RPM and I can get the actual torque and power from that graph/table? Or would that not give correct results? – miernik Apr 6 '11 at 18:28
• @miernik Assuming everything else remains constant, the torque will be the same at a given RPM and throttle position. But that assumes constant air density (altitude, temperature, humidity) and assumes that engine performance remains constant (the air filter never gets dirty, etc). You'd need to dyno the car at multiple throttle positions and interpolate between them as power vs throttle isn't strictly linear. – Mark Apr 7 '11 at 13:29
• @miernik Its not really any simpler or harder, its still just reading a throttle position. A gas engine controls throttle by controlling the air sent into the engine. The engine computer measures how much air is going in and provides the proper amount of fuel for the target AFR. The Diesel just controls the amount of fuel sent in and injection timing since 'running lean' isn't an issue. The graphs you see for torque and horsepower are normally at WOT (wide open throttle). You'd need to run tests at various partial throttle positions to get all the data you need. – Mark Apr 7 '11 at 19:34

I'm guessing that it would be difficult to install a torque sensor in most cars without disassembling the thing. I imagine that you'd have to place load cells in the most awkward locations.

But... There are devices out there that can measure horsepower, speed, etc by using accelerometers. You can even get an iPhone app to do it. You can't get RPM from that, but connecting to the OBDII port or doing something like what this device does could work.

• This is a 1992 pure-mechanic, no electronics (besides the simple regulator on the alternator) car, no-OBDII, nothing more high-tech then a diode or light bulb is in this car I want to install this measuring device in. As to the accelerometer method, well, that will measure how much power is needed to move the car, but not how much engine power is used - which will be more because of losses in transmission. So I am looking for something to measure how much power the engine generates. Although it might be useful to measure both and calculate the transmission efficiency out of that. – miernik Apr 5 '11 at 18:55
• @miernik Wow, you got me then. The accelerometer method will indeed measure the power transferred to the road, not the "gross" power of the engine. I'm not sure that you could measure the power of the engine separate from everything else (transmission, alternator, fan belt, etc) without removing the engine and putting it on a specific test rig. As for getting the tach, you could get power from the cigarette lighter and looks for engine noise on that. – user3624 Apr 5 '11 at 19:13
• A traditional way to get the tach is to sample the low voltage side of the ignition coil and watch the points open and close. You then divide by the number of spark plugs. Seems like torque has to measure force, and that implies load cells somewhere. Perhaps you can make engine mounts with integrated load cells and sum the load on the mount points? – RBerteig Apr 6 '11 at 7:11

Let's take a look at the constant.

Power in this sense is moment of force times angular velocity. A problem here are RPM.

Let's turn them into angular velocity. So we have number of circles that is made each minute. A single circle is $2 \Pi$ and each minute has 60 seconds, so formula would be $\omega=\frac {2 \Pi RPM} {60}$.

When we calculate that, we get it to be approximately 0.1047195512*RPM = C*RPM. Now lets turn multiplication into division. $C^{-1}\approx 9.54929658548$. And since we're using kilowatts, we divide everything by 1000, so in the end we get divisor of around 9549.29658548

Sandwich some elastomer string gauges into your engine mounts.

That'll measure evolved torque and RPM is easy from and kind of rotation sensor.

If you're budget is very tiny, You might be able to make a suitable sensor (capacitive) from rubber and steel sheets.

body | insulator |steel | rubber | steel | insulator.

the two steel sheets are your capacitor plates.

Now you just need to measure the capacitance. Being a diesel the E-Mag environment is probably nicer than around a petrol engine.