Question regarding Scangauge MPG readout & DIY device

[Gregte]

Question regarding Scangauge MPG readout. I do not have one and have never even seen one so I can only assume what the mpg display is showing. My question is, what does the instantaneous MPG reading look like? I will assume it is fairly constant if driving a constant speed, constant throttle, flat road etc. But if you tromp on the accelerator I assume the readout will drop very sharply, fairly quickly. If you then decelerate (let off the pedal completely) I would expect that the readout would rise very high very quickly. If you then slip it into neutral but while still moving fast I would expect the readout to be extremely high.

So, is this in the ballpark of what the MPG display shows or am I way off base? Also, what might be a typical mpg readout if you floor the gas pedal while moving slowly (30 mph) in high gear?

What would be the readout if you were moving fast (75 mph) and let off the pedal completely, or even shift to neutral?


The reason I ask is that I'm planning to build an FE gauge for my non-OBDII 4 cyl Sonoma. It will take the fuel injector's pulse-width-modulated signal and change that to a voltage which increase with the width of the injector ON time. It will also take the 50% duty cycle square wave of the VSS (will use a 555 timer IC to configure a frequency to voltage converter) which will output a voltage relative to vehicle speed.

Both of these voltages, speed and fuel injector duty cycle, will be input to an op-amp. The speed voltage will go to the non-inverting input and the injector voltage will go to the inverting input. The output of the op-amp will increase as speed increases but it will decrease as injector duty cycle increases. Thus a higher speed will show better mpg but at the same time more fuel going thru the injectors will counter act the speed's increased reading.

What I expect to see will be a constant voltage while cruising, flat road, constant speed. If I increase throttle I expect the voltage to drop, at least until the vehicle speeds up to match the increased fuel/air the engine is now receiving.

I expect that it will take a bit of speeding up and slowing down to find the "highest voltage" on my display, thus signifying optimum MPG for the conditions I'm driving in.

For me the building of the electronics is the fun part of the project even if it does not turn out to be a very useful device in the end, but to the point; does all of this sound reasonable with regard to my assumptions on what the scangauge MPG meter tells you?

[i_am_socket]

You might be interested in checking out the OpenGauge/MPGuino project. Might save a little R&D on your part.

[JohnnyGrey]

Sorry man, but such a design would have been state of the art in the 70's. Integrators and analog circuitry shouldn't really be relied upon for a task like this. Also, it can't really give you trip mpg. You should look into developing for AVR microcontrollers. Check avrfreaks.net for a starting point.

[Gregte]

Originally Posted by JohnnyGrey

Sorry man, but such a design would have been state of the art in the 70's. Integrators and analog circuitry shouldn't really be relied upon for a task like this. Also, it can't really give you trip mpg. You should look into developing for AVR microcontrollers. Check avrfreaks.net for a starting point.

Yes, you are right on all counts. But, I do not know how to write computer code and I also do not want any expense such as buying micro controllers, displays or dedicated laptop computers or PDAs.

Indeed I have looked at the OpenGauge/MPGuino project.

I already have all the components needed for building about anything with analog electronics. I am just looking to build a meter that will tell me if I am at optimum throttle and speed with respect to the wind, hills, temperature, etc.

However, I could reconsider, so your suggestion is not out of my realm of thinking yet.

[Andyman]

You wanted to know how many MPG you get when you floor the accelerator. Try this formula I made. It is for engines running on gasoline.

MPG=74000 x VE/(ER x R x L)

VE: Volumetric efficiency, normally varies from 0.3 to 1, depending on RPM and throttle position. It could exceed 1 if the engine has a turbocharger or supercharger.
ER: Equivalence ratio, equivalent to 14.7/(A/F ratio). Should be close to 1.
R: engine revolutions per mile or revolutions per minute at 60 MPH in this gear
L: Liters of engine displacement

It should be close. Adding ethanol to gasoline would reduce the result because more of it is needed to mix with a certain amount of air.

For the MPG while coasting, try this one.

MPG=MPH/GPH

GPH=L/6 when idling at about 700 RPM

MPG=6 x MPH/L at 700 RPM idle

or

MPG=4200 x MPH/(L x RPM) to calculate at any idle speed

This is only an approxamation. If you have an unusual engine configuration (like a racing camshaft for example) this may not be accurate.

If you have a 2.0 L engine you should expect your coasting MPG to be about three times your speed in MPH.

[JohnnyGrey]

Yes, you are right on all counts. But, I do not know how to write computer code and I also do not want any expense such as buying micro controllers, displays or dedicated laptop computers or PDAs.

Honestly, AVR development is quite cheap. I develop for two chips primarily, the Tiny2313 and the Mega162. The former costs about $2.50 and the latter, about $4. I program them with a parallel port cable I built for under $10 from radioshack parts. You sound like you're good with electronics and have a decent EE background. If you can solder, understand a datasheet, convert between binary, hex and decimal, you can probably program AVR in assembler.

Assembler isn't as hard as people make it out to be. You've got a very limited amount of instructions with a very simple syntax. In fact, I have fewer syntax errors in assembly than I do in C because it's so simple. The difficulty in assembly comes from the fact that you must explicitly tell the device what to do, and how to create complex functions with simple commands. For example, you might have to write your own multiplication or division routine.

Where in C you can say a = b+3; the equivalent AVR code would look like

.EQU A=0x0100 ;Set an alias for the location of A in RAM (for readability purposes)
.EQU B=0x0101 ;Set an alias for the location of B in RAM

LDS R16, B ;Load B into register 16
LDI R17, 3 ;Load 3 into register 17
ADD R16, R17 ;Add register 17 (3) onto register 16 (B) and store in register 16
STS A, R16 ;Store the result (register 16) into the A's RAM location

So you see, it's not terribly difficult, and if you can understand op-amps and read a datasheet, you can surely wrap your mind around this. Development is very cheap and if you already have a computer, no dedicated hardware is needed. I use my bedroom workstation for development. Check youtube for "AVR Project" for some cool ideas. Tell me this isn't crazy:

http://youtube.com/watch?v=jNMvHMVCLlg