mileage computer for a carbureted engine

[diesel_john]

very interesting Omegaman I would like to see how that paddle wheel flow sensor is made. Anyone have a pic?

[gone-ot]

HINT -- lookup the patents for Gilbarco and other gas-pump devices and see how they "measure" gasoline flow for the gas-station pumps...maybe an old gas pump unit could be retro-fitted in-line to the carburetor?

[dcb]

lol, I think archimedes did that first, not exactly a paddlewheel though.

but, these restrictions, it sounds like a simple pressure set up, like a simple manometer and a bead in the fuel line for a reading fuel flow on a pressurized line. Would fuel line and return line signals on two sides of a manometer cancel out the return line and leave the consumed fuel rate, on a graduated tube ? Sorry if this is a dumb question.

[racprops]

I have read just about every post and link to other posts on this subject.

And I can see it is a old thread.

Perhaps I can get some interest and some help??

I think some of the suggested flow meters will work or I will see about making a amultivane pattle wheel version, that way to get the most pauses per second of slow.

BUT a major problem remains for us working on Fuel Injected cars.

Current tech is devices like Scan Gauge II which work off the cars computer and thus is unable to show real MPG changes unless the cars computer "sees" them.

I have a couple of old Zemco Mileage Computers which work off a fuel flow sensor and off a sensor that works on the drive shaft to get speed. The computer all ready compares speed and distance with fuel flow to calculate MPG, miles traveled etc.

Problem is with my 2000 Ford Fuel injection there is a return line returning fuel to the tank.

So my flow sensor will "count" all flow, to the engine and back to the tank. This will not work and what is returned to the tank will change due what the engine uses.

My solution is to run two flow sensors and then compare the two counts and then output the difference. ( now that I have read all of this thread I am censured about the flow sensors..)

I think it will take a small computer chip to sample say ½ a second of both signals and then output the difference, so it will be ½ second behind. No problem.

The output pauses are from a Photocell driven by a light bulb, the light is interrupted by a ball bearing running around a track so that once in ever revolution it blocks the light from the Photocell. I would really like to improve this with LEDs.

The output and thus input to the Zemco computer has to be a negative signal to ground.

Rich


PS or putting it another way someone said:

Old cars with carbs only need 1 flowmeter. newer cars and diesels will need 2 with a little converter box to subtract return line pulses from the fuel supply pulses. The converter box would run the calculated pulses to the MPGuino. MPGuino already has some very talented programmers who I am positive would have NO trouble whatsoever converting the software to figure the pulses per gallon. The cool part would be that you should only need one number for all vehicles.

[diesel_john]

racprops,
My diesel heats the fuel going thur the injector pump so temperature compensation maybe in order. I have taken my Zemco setup off. If I try to measure again I am going to have a graduated cylinder of fuel in the car and two 2 way valves so that my co-pilot can switch off the car tank and switch on the graduated cylinder over a measured distance. Compensate for temperature and eye-ball the fuel used. Crude but empirical. Kids do not try this at home... your mileage may vary.
On my '91 F250 I simply reduce the voltage to the injectors to lean out the mixture. The computer tries but can't compensate enough to richen the mixture but any approach measuring injector pulse width would be invalid. So i know what your mean about limitations of measuring methods.

[mozul]

I see this thread is a bit old, but maybe there is life here yet.

I have begun assembling the parts for an MPGuino for my ancient carb car. I got two flow sensors designed for a boat from ebay here: 2 Standard Communications Fuel Flow Sensors Model ft 100 Fuel Transducer | eBay

and I figured out the pins for that. It puts out a nice square wave when plugged in to 12V and I blow in it to get the impeller to spin.

I am following this article to make my VSS.
Installing a Universal Magnetic Pickup Speedometer

I already have the Arduino Uno and will gather the rest of the parts once my VSS is working.

My main question is: I noticed that some have said that the code may have to be changed for a carb car. How would it need to differ exactly?

[wyatt]

Well, there are only 4 connections to the MPGuiono. The VSS, the Injector pulse, the +12V, and the Ground. You are building a VSS that (I think) should have a consistent signal with what our cars are seeing, so that should be good. The +12V and Ground are easy, so no problem there.
The question will lie in how the Injector pulse is different. It's been a few years since I built mine, and a couple years since I have used it, so my memory of how it worked and such may be a bit foggy. I think the injector was sending a +5V signal (may have been +12V?), so this may need changed. It was also a pulse, as in a square "wave" input, where we were able to adjust (post process) for a triangular front and back (since you will never be able to get a true square input). I don't know how yours works, but if you get a nice square output from your flow sensors, without a variable peak height or anything like that, you may be good to just use the code as is, and just modify the same inputs that everyone else had to modify to get the MPGuino to work with their individual automobiles. It would be easy enough to just try it out and see how it goes.

[mozul]

The signal coming from the flow sensor is a square wave. Here is the image of it on the Oscilloscope. It gets wider as the flow slows, until it stops.

[wyatt]

I think that the INJ signal that my car would send would have been square getting narrower until flow stopped, so a "1-INJ" type term may be in order. Like I say, I don't have a really clear memory on all of this, so if someone with authority (or a better memory) chimes in, that would be good. Otherwise, it seems like you have the know how to get things done, it may just take a bit of playing around. If you need old MPGuino code, I may have some of the original code that ran on the 128's somewhere, but I haven't messed around with it in years.

[AndrzejM]

If you're using flow meters there are two things that you have to focus on. First is counting pulses, because one pulse is a fixed amount of fuel. So if you know pulses count you'll know the amount of fuel going through flow meter. If you subtract pulses from flow meter on a return line from pulses count on a feeding line you'll know the amount of fuel used by the engine. If you want to have instant FE reading though you have to count time between pulses, then you'll know the exact fuel flow.
So you'll need to connect two separate lines to arduino from both flow meters and a VSS signal.