Power meter for car?

[Nicko]

Originally Posted by MetroMPG

...
If the user doesn't attempt to control for external variables as much as possible (eg. speed, grade, road surface, vehicle weight, wind, engine/drivetrain temperature, traffic, other vehicles' effects on your aero, cruise control vs. driver's right foot, etc. etc.) then comparisons between state A & B are suspect, regardless of the measuring tool.

Just saying.

Yes, there are a lot of things that can mess up the quality of the data but there are some things you can do to make the most of the time invested.

Some smart guys (like R Chung and A Haile) have come up with methods to account for accelerations, grades and kinetic energy when regressing power on speed in order to derive Crr and CdA. These methods utilize the fact that a loop has zero net elevation change. With several laps of varying speeds, one can calculate which constant force and which F=k*v2 best describe the work/energy measured by the power meter on a per sample basis

The real enemy is wind. All other factors can be handled with a good testing protocol, a good venue and accurate sensors for mass, torque/force, speed, temperature and humidity. This has been proven many times by many guys with power meters on their bikes.

I see no reason why this wouldn't apply to motor vehicles also, in fact I think the data could be even "cleaner" since the speeds with a car are 3-4 times higher
so any wind would be less influential.

[Nicko]

Originally Posted by RobertSmalls

...
I don't know exactly where the Corolla's motor mounts are, but my Honda has three motor mounts, and all three are six-axis connections, and they all resist driving torque. You should get out a crowbar and see how stiff your mounts are in the relevant directions, and from there decide whether you're comfortable with the assumptions you want to make.

I took a good look at my engine mounts, pushed and twisted them around and realized that my best bet is to let the left and right mounts bear all the weight of the engine/gearbox. My load cell would then be arranged so that it only (as cleanly as possible) sees "torque reaction" around the left/right engine mount "axis".

Can the two side engine mounts take all the load? They seem stiffer overall than the two "fore/aft" mounts (of which one looks "cracked" already).

[RobertSmalls]

Originally Posted by Daox

A g-tech uses accelerometers to measure torque and therefore horsepower. I'm not sure if that is exactly what you are looking for, but its a start.

I don't believe this would work for our purposes. An accelerometer is simply going to read zero if you're cruising along at 60, and it'll read zero as you press a little harder on the throttle to make up for rolling down the windows. If the G-Tech works like I expect it does, you have to do 0-60 sprints before it tells you how much HP+Torque it took to accelerate that fast.

[RobertSmalls]

I look at the equation:

mx"+Crr*x'+0.5*CdA*rho*x'²=F

Of {m; Crr; CdA}, know two and you can calculate the third. It would seem to me, you would be able to tease m, Crr, and CdA out of the test data if you include some driving that highlights each term: Do some quick acceleration at low speeds (where aero is barely in play) to find your mass. Some very low speed cruising will get your Crr. High-speed driving will confirm your CdA.

As I understand it, this instrumentation plus careful analysis of the data *would* eliminate the need for A-B-A testing by producing CdA values that you can compare to other test configurations. Of course, Crr depends on factors extrinsic to the car, including pavement type and temperature, so back to back A-B-A testing is best when you want to work on your Crr. Using the same road each time is second best.

I was hoping for super-accurate bar graphs of energy consumption on the fly. Well, that would be nice, but due to the variable nature of rolling resistance and air density, and the effects of winds, what you display in real-time while commuting wouldn't be as accurate as values you manually comb out of a mountain of data.

[RobertSmalls]

I just had a stupid thought. The G-Tech device can tell you your torque as a function of RPM at full throttle. So, why not drive around at 100% throttle (P&G) while logging your ABS wheel speed sensor and crankshaft position sensor? That'll give you x', x", and torque.

[Nicko]

Finally had some time to work on this power meter idea.

I removed the (cracked) front engine mount today and took a 5min drive to see what would happen...
Nothing

I'll try this for a while now to see if any vibrations or noises will develop over time. If successful, I believe it's possible to replace the two engine mounts with only one. My plan is to replace the front engine mount with a load cell equipped "copy" and at the same time remove the rear mount.

Now, regarding instrumentation, I read up a little bit on the MPGuino project and realized what I want: The MPGuino to measure speed, injector pulsewidth AND propulsion force and keep at least one hour of data sampled at 1Hz in memory for later off-loading to a PC. That would be some 10kB of 8bit samples, within reach of the 32kB of flash memory in the ATmega328P chip (?). This sampling would take place "behind" the normal functionality of the MPGuino.

Does anyone know if this is possible? How would the PC software make contact with the MPGuino and download the data over the USB?

I see great potential in this.
- More accurate Crr and CdA measurements.
- Engine efficiency could be calculated from fuel consumption AND mechanical power production.
- Throttle opening and engine temp are next in line to be measured
- For even more accuracy in the CdA measurements, one could put a load cell equipped "aero dummy" on a boom into clean air. This would be measuring the "0.5 x rho x windspeed^2" part of the equation and be used as a "dynamic pressure reference"
- More nerdy stuff...