More accurate gas tank readings

-
FISHBREATH

FISHBREATH

Well-Known Member
Joined
May 21, 2008
Messages
21,384
Reaction score
73
One thing that has always bothered me about automotive gas tanks and gages is the innacuracy concerning actual volume in the tanks. I know that the tanks are made to fit into various locations, thus having unusual shapes. The standard float sending units cannot differentiate between wide and narrow parts of the tank.

Has anybody ever considered this: when I worked in wind tunnel testing years ago, we used various instruments to sense the physical locations of various components. Most were simple units constructed similarly to linear rheostats. The analog data was fed to a computer through analog/digital convertors. We would calibrate the units by running the machinery through the ranges, taking data points along the way. The computer then gave us an accurate digital read-out of the mechanical position on a monitor.

Here is the question: could the electrical signal from the tank sending unit be sent through an A/D convertor and into a simple programmable chip and then into a gage? The unit could be simply calibrated by pouring in a gallon of gas at a time and taking a data point, continuing the process until the tank is full.

It seems to me that this would provide a very accurate, gallon-by-gallon, read-out on the gas gage (probably an aftermarket gage or digital read-out).

As I am an "electron moron," do you guys think that this is feasible?
 
Lots of ways to do this. Modern cars monitor fuel flow already, most actually keep track of how much the injectors are flowing. Couple that with a weight sensor on the tank, or another flow sensor in the neck of the tank and you'd know exactly how much fuel you had. Heck just a weight sensor on the tank would get you closer.

But I'm not sure I understand the point. So you can literally run the tank down to the absolute drop of fuel? People would still run out of gas, thinking that last couple of drops would get them to the gas station. It would just make them even more irritable if they got caught at a light, knowing they only had 3 ounces of fuel in the tank.

I have a pretty good idea of exactly how low I can run my tank, and I still almost always fill up at a 1/4 tank. No worries that way. Once you've put a couple tanks of gas in a car, you should have a decent idea of how the gauge works.
 
The problem is not the A/D converter. The problem starts with the sender itself, and ends at the dash cluster with the gauge, and voltage limiter for the gauges. From one end to the other, this is an inaccuracy waiting to happen.

So the first thing you'd have to do is come up with a quality sender (which I guess no longer exists)

There's lots of "talk" around the www about replacement senders.

On mine, I recently discovered that if it gets down to 1/4 tank on the gauge, you'd better be at the stop light across from the gas pump. I know this because..........
 
I put a 8 ohm resistor inline in the sender wire and now it drops all the way to empty on the guage before it runs out of gas.
Drawback is that the guage will not go all the way up now, but to 3/4.
I don't care about that though, I care more that it runs out of gas when the guage says it should.
I carried a couple of gallons around in the trunk for a few weeks just in case while testing.


67Dart273 said:
The problem is not the A/D converter. The problem starts with the sender itself, and ends at the dash cluster with the gauge, and voltage limiter for the gauges. From one end to the other, this is an inaccuracy waiting to happen.

So the first thing you'd have to do is come up with a quality sender (which I guess no longer exists)

There's lots of "talk" around the www about replacement senders.

On mine, I recently discovered that if it gets down to 1/4 tank on the gauge, you'd better be at the stop light across from the gas pump. I know this because..........
Click to expand...
 
My dad's sand car has a digital dash in it. We did some thing kind of like that to programe the electronic sender and dash. The dash came with a program and softare. We had a tube we filled up with gas a little at a time not the tank though. His can be changed to read either Gal or a % of gas left.
 
In F/A-18's, and probably most other aircraft, there are several fuel sending probes in each tank. For example, each wing has 3 probes; an inner, middle, and outer. The signals from all three are sent to a Signal Data Computer (or Signal Data Converter, depending on what Lot aircraft) and an average of the 3 is used to determine the fuel load in lbs. There are probes similar to those that are used in racing fuel cells. In a racing fuel cell, they are accurate because the cells are uniform in shape from top to bottom.
I don't really think there is a way to accurately in most gas tanks because they are irregularly shaped, unless there was one probe that read top to middle and another probe to read middle to bottom, each calibrated to read it's own half of the tank, but that's getting way beyond what I could ever figure out how to do.
 
'Member one time I was goin' somewhere in my 70RR. Those have the filler low, behind the plate. I dove off I5N, filled up, and the ramp back N was nice and straight, so I got right with it. Ran up the road about a mile, and here we have less than a 1/2 tank. Had the trunk open, checking the oil, and had laid the cap in the trunk, then forgotten it. Boy it didn't take long to sheeeitt a lot of gas right out the rear.
 
I find this interesting ... I have Autometer gauges with stock mopar sending unit. I never get a accurate reading. When it reads a half a tank that means empty actually. Another words if I fill up it takes a long time for it to come off "full".
 
The suggestion would work and should be fairly easy in a new car with a vehicle computer. Indeed, I expect they already do that if the fuel display is digital (suspect many analog gages today are actually a computer output). Many have long suspected that auto manufacturers design so that the fuel gage drops slower at first. Their benefit is that when people take a long test drive or rent a new model, they get the impression they are getting great mileage ("only used 1/4 tank"). John Q Public has no concept of non-linear response. He thinks if you double something, you always get double the result, that is simply "common sense".

I mainly care about "exactly where is empty". To determine mileage, I use the meter on the fuel pump. It probably doesn't shut off at the exact same volume when full, but that averages out over many repeats. Just write down the mileage on the gas receipt each time, or just reset your trip counter if you have one. If I fillup 19 gal and know I have an 18 gal tank, I report the station to the State's Measures and Weight Department.
 
BillGrissom said:
If I fillup 19 gal and know I have an 18 gal tank, I report the station to the State's Measures and Weight Department.
Click to expand...


I've done that before when the pump registered that I took a gallon more than my tank held. I got a letter in the mail about a month later telling me that they recieved my complaint and were investigating it. Never heard anything after that, though.
 
FISHBREATH said:
One thing that has always bothered me about automotive gas tanks and gages is the innacuracy concerning actual volume in the tanks. I know that the tanks are made to fit into various locations, thus having unusual shapes. The standard float sending units cannot differentiate between wide and narrow parts of the tank.

Has anybody ever considered this: when I worked in wind tunnel testing years ago, we used various instruments to sense the physical locations of various components. Most were simple units constructed similarly to linear rheostats. The analog data was fed to a computer through analog/digital convertors. We would calibrate the units by running the machinery through the ranges, taking data points along the way. The computer then gave us an accurate digital read-out of the mechanical position on a monitor.

Here is the question: could the electrical signal from the tank sending unit be sent through an A/D convertor and into a simple programmable chip and then into a gage? The unit could be simply calibrated by pouring in a gallon of gas at a time and taking a data point, continuing the process until the tank is full.

It seems to me that this would provide a very accurate, gallon-by-gallon, read-out on the gas gage (probably an aftermarket gage or digital read-out).

As I am an "electron moron," do you guys think that this is feasible?
Click to expand...


thats great but why? why re-invent the wheel and make it more complicated when it rolls just fine?

when my car gets to around 1/4 tank I fill it, as long as the reading is somewhat right and not going to leave me stranded I don't care
 
I like the inaccuracy in my Caravan. I know I can get at least 4km when it reads 0 lol..

Grant
 
The reason for the question is road trips. I plan on configuring my '67 New Yorker for interstate driving. There are some areas of the country where gas stations are few and far between. I would like to be able to calculate estimated range based on fuel remaining in the tank.
 
Hell I just use a stick in my tractor.
 
I often just use the odometer as a fuel gauge on the bike or truck (truck gauge didn't work for years) I know my average MPG, zero out the trip odometer when I fill up,
MPG X gallons in tank = max range, 1/2 of max range = 1/2 tank of fuel,
I find it is more accurate than a working fuel gauge
 
You can do it. Just need a gauge and sender that works off of amps instead of ohms.
Need a digital gauge with a program though. You put a gal in your tank then check the amps. The put another gal i and check it ect till the tank is full. Then you you punch in all the readings in to the program and it does the rest.
 
Burntorange70 said:
You can do it. Just need a gauge and sender that works off of amps instead of ohms.
Need a digital gauge with a program though. You put a gal in your tank then check the amps. The put another gal i and check it ect till the tank is full. Then you you punch in all the readings in to the program and it does the rest.
Click to expand...

That's exactly what I was thinking about.
 
You must log in or register to reply here.
-

Back
Top