We need an "MPG aware acceleration computer"

[stillsearching]

Was reading another thread asking how different rates of acceleration affect mileage, when the total amount of energy to get to X speed should according to physics be constant whether applied over 30 seconds or 3 seconds... see thread for the discussion.


Reading through it all made me aware that maybe hypermilers need even better gauges to get more insight.

I mean we have realtime MPG readouts but that only shows best cruise speed for instance, it does not necessarily show the most efficient way to accelerate necessarily (keeping the figure low may unnecessarily piss off traffic behind you plus the BSFC is less efficient at loads below a certain level) and even if you learn a good habit for say driving on familiar mostly flat roads, if you travel and end up in hills the same strategies may not apply the same trying to accelerate uphill for instance.

I would think a realtime BSFC map would be ideal, especially one taking partial throttle loads into account. (I have to assume at half throttle the figure changes slightly at a given rpm and load, just like other aspects like torque change) For that matter a realtime torque curve display, depending on throttle. Or other graphics that might be useful for helping the "human computer" decide how much throttle to use and when to shift, something a little more complicated than the "shift now" light on most cars. In heavy traffic where I can't drag my rear it could still provide assistance to suggest how to accelerate more efficiently under those conditions. Or maybe i'm towing a small 1000lb trailer and that totally changes the physics of the car I was used to driving in a certain way before. Or I have a couple kayaks on the roof rack throwing off the aero but affecting weight minimally.

There's so many variables that it's really more in the realm of the computer. And since carputers and cheap LCD's are available now it shouldn't be that hard actually to make something. The main questions are to figure out before anything what data is actually the most useful, beyond the realtime MPG display. The realtime MPG display is great for highway cruising under all loads because it accurately represents what you will be using. It does not give terribly useful information for "pulse and glide" calculations, varying highway driving conditions (unable to pass or set a steady state), nor in town. So what could we do to improve it?

[christofoo]

I've thought this many a time.

It doesn't need to be a different form-factor from the existing FE instruments, MPGuino, SGII, UG... it just needs another "gauge" on the UI that displays "Engine Efficiency %".

The only question is how will it map BSFC? Will the firmware have a BSFC map stored in memory for every engine? That's a logistical problem, but it could be done.

Otherwise it needs some additional sensors, i.e. windspeed and incline (or altitude), to be able to calculate on the fly. Even then it wouldn't know (and shouldn't assume) Cd and Crr, so it would need to express the result on the UI some other way than "Engine Efficiency %", for example "MPG-trip-predict" - the MPG you'd expect to get over a round trip at your current efficiency. If this instrument were implemented in a smart-phone, it could use the GPS for altitude (although of course GPS has more error in altitude than it does in latitude/longitude, so the readback would be slow or noisy). Or if it were implemented as a regular FE instrument it could have an accelerometer it could use to get the slope directly (although you'd have to have a method for aligning it, and it would have to be fixed to the panel reasonably well. I also don't think it would work that well considering the vehicle's suspension allows the car to tip, which would create false hill readings during acceleration and deceleration). Arguably it could neglect windspeed. Let the user figure out that wind is hurting/helping their mileage. Actually, I lied about neglecting Cd and Crr. It needs to have some idea how much drag the vehicle has, and you'd need to program the weight of the vehicle along with the other calibrations, so it will know how many miles you can coast at a given speed. But drag is something it could learn on its own, which is best since it will end up factoring engine breaking, which is helpful especially since it may depend on the driver.

I've always thought FE instrumentation could be designed to be far more instructive to new users. MPG is not the most helpful number. Trip average MPG is only a little better.

[Sven7]

Kind of confused here. Why not just use a vacuum gauge to track engine load?

[GRU]

honestly, acceleration is not the most important part, coasting is. you can't constantly look at a screen on your dash to check if you are accelerating corectly, at least safely, but spend time and learn how to conserve momentum and coast as much as possible because that's how you get the good numbers. and if you're on the highway most of the time so you can't coast just go as slow as sefely possible.

on my trips i probably spend about 5 % of the time accelerating so even if i do it completely wrong it won't hurt my milage too bad. go over the trips that you do and find out how much time (in procentage) you need to accelerate, hold speed, coast, idle and focus on improving the ones you spend lots of time doing.

[stillsearching]

Originally Posted by christofoo

It doesn't need to be a different form-factor from the existing FE instruments, MPGuino, SGII, UG... it just needs another "gauge" on the UI that displays "Engine Efficiency %".

The only question is how will it map BSFC?

Otherwise it needs some additional sensors

A gauge may be less useful than some kind of visual representation, if I remember right the 1982 Corvette had a tachometer which left to right was the RPM but also had the torque curve under full throttle mapped out to provide more insight into when to be shifting.

My ideal would be an LCD panel providing a visual representation, until youve seen it so much its pretty much memorized and you can turn the display to something else. Acts like a driver training aid. Under "unusual" conditiosn turn it back on.

Sensors shouldn't matter too much - throttle position itself and vaccuum/boost should provide insight into engine load. Were not predicting, just measuring, and going off whatever the factory or us calculates (or estimates) things to be.

Originally Posted by GRU

honestly, acceleration is not the most important part, coasting is. you can't constantly look at a screen on your dash to check if you are accelerating corectly

It may be the biggest part of it, but different load conditions (say in the mountains, heavy winds, trailer towing, or just plain different conditions of closer traffic) make it matter alot more.

It need be no more of a distraction than the dozen other gauges you see in a sports car, from oil pressure to transmission temp. Having a bright white spot on a false color map can probably be read in your peripheral vision or on a heads up display just fine.