I've mentioned this before and I'm still kicking it around, just so many things going on I'd like some opinions before I start investing in a few parts to attempt it as well as if it's actually feasible enough to even bother with exploring.

I'm trying to figure out a way to determine how full (or empty) a tank is by hitting it with a sonic pulse and looking at the resultant reflection that returns.

The real world: Most motorcycle tanks don't have any sort of gas gauge on them and due to their design of straddling the frame it isn't really feasible to introduce any sort of sensor down the filler neck from the top. On top of that they're of an irregular shape.

Let's say I've got a metal tank that holds four gallons and it's half full.
Obviously I've got 2 gallons of liquid in it and two gallons of free air space.

If I attach some sort of piezoelectric device to the outer metal of the tank, hit it with some sort of pulse then look at the (hopeful) reflection is there going to be enough of some sort of variation returned that would indicate what percentage/how much of the full volume is occupied by liquid and how much is occupied by free air space?

 

The safest thing you can do on a motorcycle is to run out of gas.

How about using fuel flow? That method is now used routinely in aircraft and other systems in which a typical fuel gauge doesn't work well. Even some cars offer real-time fuel flow as a bit of marketing pizazz.

John

 

How about a thermal based sensor? Run some current through a resistor for a while and measure the before and after temperatures. The rise will be higher if it's air on the other side.

 

Maybe you could ping the side of the tank periodically, and do an FFT on the resulting audio transient. You could then use a lookup table to display the remaining gas.
It would probably work best on a metal tank.

 

I don't ride this bike as my main vehicle, it's a restoration project that saw all of 100 miles last year and none yet this year. I used to ride all the time but that was in a college town where they looked out for two wheel vehicles. Sadly anyone with a motorcycle is at a distinct disadvantage nowadays what with soccer moms driving SUVs, talking on their cell phones and helping the kids operate the DVD player in the back seat at the same time they're trying to put on makeup. People tend to drive on autopilot and could really care less when they're driving a tank that has 88 air bags.

The question of coming up with a sensor method is perhaps to benefit others as well as myself. Some variations of this old bike (such as mine) only have a 1.6 gallon tank and you can't see down into the darn thing to know where you're at. Counting miles, even when you know you average around 50 mpg, just isn't my idea of knowing where I'm at, I like to know a general idea of that so if I'm by a station that sells good gas I can fill it up if need be.

Fuel flow is not exactly that easy or inexpensive to calculate, and as far as a resistor I don't want to intrude into the tank much less have electricity going into it.

This is actually a general idea that could be used for many things, if I can come up with some sort of sonic method that works then it could be adapted to monitor the water level in a couple of tanks that are part of the HVAC system at work.

I've got a great scope and power supply but when it comes to a signal generator I'm limited to an antique HP rackmount unit that weighs a ton. Extremely accurate and stable for its day but not all that portable. I suppose I can easily whip together something more appropriate for that.

It's my guess that the amount of air space is going to cause the tank to exhibit some sort of resonant frequency peak if properly excited. At least this is where I was going the last time I thought about it.

 

Fuel flow is not exactly that easy or inexpensive to calculate, and as far as a resistor I don't want to intrude into the tank much less have electricity going into it.

I didn't mean for the resistor to be inside the tank, that would be an accident waiting to happen.
It would just need to be on the outside of the tank with a good thermal contact.

 

Still a couple of problems - I'd rather it not be visible so it has to mount to the bottom or inside area of what goes around the hump.

I'm just going to have to come up with something to ping the tank with and look at what comes back on a scope.

 

What is the minimum accuracy you are looking to achieve?

hgmjr

 

probably 10%

 

Here is an interesting patent on the subject: http://www.freepatentsonline.com/7287425.html

Note that it measures from the bottom of the tank to the surface, i.e., it measures fuel depth, not empty space and relies simply on TOF, not analysis of a reflected waveforms.

Here is a link to a commercial unit for a somewhat larger tank than is on your motorcycle that apparently works similarly.

http://www.scribd.com/doc/9641764/Measuring-Diesel-Fuel-Tank-Level-122808

Back to the patent....it seems simpler than waveform analysis. You mentioned earlier that the tank has two bulges on the bottom separated by the frame. Are those bulges connected? If not, do you control cross-feed?

John

 

You mentioned earlier that the tank has two bulges on the bottom separated by the frame. Are those bulges connected? If not, do you control cross-feed?

Yes, it's just one single tank, the "equalizing" area would be towards the rear of it such that if you lean to the right the fluid flows towards the right half of the tank.

Thanks for the links, I'll go through them this evening when I get home.

The only drawback to measuring the level of fluid is that the tank would have to be level at the time to know where you're at, something that could sense presence/absence of content would work regardless.

 

I didn't mean for the resistor to be inside the tank, that would be an accident waiting to happen.
It would just need to be on the outside of the tank with a good thermal contact.

You guys don't realize that the fuel gage sender on your car has a OPEN wire wound potentiometer inside it.

Also most fuel injected cars have the fuel pump inside the tank with a unsealed brush style electric motor!

 

Consider putting a pressure gauge on the fuel tube out of the bottom of the tank. Measuring the head would let you calculate how much fuel is in the tank (calibrate it while the bike is vertical). You'd have to average out noise.

If it was me, I'd make myself a glass sight tube; no electronics required. Of course, it would have to be mechanically robust to avoid breakage. I'd machine the thing out of brass bar stock and epoxy in some heavy glass tubing, then connect the thing to the tank with two rubber fuel hoses. Make sure the epoxy is compatible with the fuel and include shut-off valves so the gauge can be isolated from the tank. The return line at the top might require you to drill a hole in the tank to attach a fitting, so the usual cautions about working with gas tanks apply.