I have a fuel injected vehicle that I want to tie the fuel usage & gps into dynamic "mpg" reading. So I would need two sensors, one to read fuel going to the injectors and one coming back to the tank and take the difference of them.
There is no reading from the ECU on fuel usage, nor is there a good tank level sensor.
Looking around I found some gasoline safe sensors (this with this).
Is there a better way to do it?
Or is there a less expensive way? $100 bys gas; and not sure saving s would pay me back for it...
Many people measure the frequency and/or duration of fuel injector pulses to determine fuel usage. If you can get an oscilloscope connected to your injectors you can tell if the ECU uses varying frequency of fixed-length pulses or fixed-frequency pulses of varying duration to control fuel flow.
The injectors are mapped to throttle position and rpm -- Using the injectors would require me to duplicate the map; and I don't think the Arduino would keep up with the changes... Maybe if I had a dedicated one to just the injectors...
Thank you for the idea; but I don't think that one will work ...
Is there a better way to do it?
Or is there a less expensive way? $100 bys gas; and not sure saving s would pay me back for it...
I've seen that Atlas device before and I don't understand what it has to offer. If you want to save $23 before you start, not buying it might be a good idea. Arduino can measure flow rate and accumulated flow without it. Further to that, I doubt you would need any sensors over what you already have. There is nothing new about these displays and, even if your car doesn't have one, there is probably a version of it that does and some research therein should reveal all. I bet it doesn't have any extra sensors and it's just a case of wiring. My current humble Ford Falcon is my second car with this facility. They both do the same things, I don't know how they work, but there is no evidence of anything special on either of them.
In short, I believe you can do this with a basic arduino, a 16x2 display, a couple of buttons, and some 12v to 5v level shifting circuitry.
I don't think you should expect to save money from this unless you have constant and extreme variations in driving conditions. They are mostly there for entertainment or, at best, the intellectual exercise. That goes double if you want gps involved.
What? How on earth would you be able to know how much fuel is being sent out of the tank? Moreover, if you know how much is leaving, you need to know how much is returning to the tank.
~~Flow sensors have paddle wheels in them an a hall effect pickup to pulse the flow. ~~ The only chance of knowing w/o a flow sensor, would be to pull each injector and have it tested for flow rate (they are rarely, if ever, exactly as advertised, and rarely, if ever, matched across all injectors); then set up a pickup on the trigger for each injector and calculate it's usage based on time the injector is energized. (as suggested above) -- this method would require substantial power from the arduino cpu, when dealing with 8 injectors firing 2 at a time at a rate of 10,000-14,000 per minute (rpm.)
-- edit --
@Nick_Pyner - Sorry, I misunderstood you, you're saying the pulse board is not needed. You are probably right; I probably could do it w/o that and just the actual flow sensor. Perhaps I will give it a go w/ just the flow sensors & arduino.. Sorry for misunderstanding...
@johnwasser -- Thank you; I read a few pages, and have printed it off -- will read it on the ferry -- if anything, I'm sure I'll learn something from it. Thanks.
jAssing:
I misunderstood you, you're saying the pulse board is not needed. You are probably right; I probably could do it w/o that and just the actual flow sensor. Perhaps I will give it a go w/ just the flow sensors & arduino.. Sorry for misunderstanding...
I don't actually know what that Atlas board does that is special. I have three arduinos with Hall effect flow sensors that work fine without it, or anything remotely like it. More to the point: I don't think you need the flow sensors either, thereby saving a few bucks and the installation hassles, and I'm sure the Ford Motor Company agrees with me. I say this because I know what a Hall effect flow meter does and looks like but I don't see any sign of one, let alone two, of them on the current car or the previous. I don't see any surprise in that.
Without bothering to go into all the technicalities, the job of the electronic control unit is to measure the fuel and deliver it as required. All the display does is be a bit more forthcoming with the details, and the signals enabling it to do that are either the same as, or just a byproduct from, the command process. There is surely no need for any pickup devices either, all they can do is tell you what you should already know.
Further, I had occasion to quite rigorously check the petrol consumption against the display on my 1990 dual-fuel Fairlane, and it was extraordinarily accurate. This is something not so conveniently done on the current car.
Having said all that, if you want to go on the Hall effect turbine path, it can't be hard. I guess the Jaycar/Freetronics flow meter software might suffice. All (!?) you need then is to get a feed off the tailshaft encoder and you can probably use the same software for that.
I have notorious bad experience with flow sensors, especially for low flow rates. The paper mentioned by johnwasser mentions this fact. They settled for a custom designed differntial pressure sensor.
Thank you -- while this is not a ford, the only thin the ecu measures is "time the injector is open". This does not equate to fuel used, unless you know the injectors flow rate . I will look into alternative ways to measure it.
I have not read the paper (yet, it's printed and will be commute reading today). I hope it let's me know what a low rate is. I can consume 5 gallons in less than an hour.
Further more, I would like to also test rates on a carbonated truck.
I have a FloScan 9000 meter on our boat, which uses a hall effect sensor (turbine/finned) and is dead on accurate, after 60 gallons of flow "shown" the remaining fuel is exactly correct, telling me the meter is spot on accurate.
If that's all you need, a turbine should be fine. The situation could be quite different with continuous rate of flow readout, even in the face of the cumulative result being OK, and I guess that is what nilton61 is referring to. I am using far high flow rates than any car engine and I don't think it is something confined to low flow rates. The size of the turbine is selected for the size of the job, and the accuracy of the accumulated flow suggests that the problem with the rate is inherent in what they are.
The 1st use of the FloScan was off; but it was easy enough to calibrate; but at the price (and that's w/o a gps) they are prohibitively expensive (not to mention doing it myself is far more enjoyable) for a hobby/project such as this... What they do in hardware, I suspect I can do in software (calibration and calculations).
The tank holds 5 gallons (no gas gas gauge) and I want to get an alert / reading so I can see when it's almost empty to avoid the pump running dry... I can make that 5 gallons last 2 hours or sub 1 hour; so it is variable on use, load, speed, terrain etc. I don't need to know if there's "3 mm of fuel left" if, in the end, it thinks I'm empty with 4.5 gallons consumed, I can work with that....
That was a very good read, and very informative - thanks!
Having the manifold made is trivial; but they didn't list any part #'s -- and all I can find from the manufacturers listed don't list "gasoline safe"; nor the specs for the sensor itself; so it's hard to figure out.
however, Their problem with flow/paddle wheel type sensors was not finding one to work slow enough (I was going to use a jet/small orifice to get the original wheel spinning at low flows, but as they pointed out, that didn't work well)
After spending a day / night (re)searching differential sensors, I found a sensor that is capable of flows close to the Porter/Zheng project (Projected needed minimum:35ml/min and maximum: 510 ml/min) FLOWFUEL30L0 capable of Minimum: 33 ml/min and maximum: 500ml/min For a lot less money -- so I think I'm going to try that one 1st, if it fails, then I'll circle back to the pressure differential.
Curiously (to me anyway)is that they discarded the ultrasonic sensor for "not metallic fluid" -- but I found (cost prohibitive) ultrasonic sensor that was very accurate at low or high flows. (url not saved, sorry)
If anyone is interested, I'll post my results once it's all together -- will do bench testing this winter and field testing in the spring/summer.
Since you've read the report, it might pay to examine what exactly Porter's problem was, and how it affects you. The fact that they had a problem may go quite some way to explain why Ford doesn't use them and Porter might have saved himself a lot of trouble by asking the taxi driver about it on the way to the lab.
The sensor you refer to doesn't appear to be fundamentally different from any other. It is just sized to the job and comes with 1/4" barbs for fuel lines i.e. you may not be solving the problem with it, just moving the problem, but incurring less expense to find out that you've still got it. Most notable in the spec, apart from the accuracy, is that it delivers 200 pulses/litre, about the same as its sibling.
You might check the Swissflow SF800, which is also reasonably priced. I believe it only claims to go down to 0.2 l/min but it is an axial flow turbine and delivers 6000 pulses/litre.
Thanks; but either way I'm at an experimentation stage.
The differential pressure sensors I've found don't seem to be fore corrosive/flammable fluids; w/o part #'s I'm taking a guess and trying -- at a higher cost, as I need to make a manifold....
I know a certain amount of r&d will go into it; and part of the R&D will be worthless (or in-appropriate) hardware.
the SF800 looks nice, but it's slow rate requirement is too high...1/2 liter per minute when we need down to 30ml/min....
I will re-read tomorrow; but w/o a part # I'm left guessing at the differential pressure approach;and sizing of the manifold (I haven't run the calculations yet; perhaps that will lead me to the manifold size...) I ordered one of the less expensive differential pressure 4 wire sensors -- not sure if gasoline will negatively impact the plastic...
It is possible I missed the actual part # (differential pressure sensor) used; free and honey both make many sensors.