Date: Thu, 27 Apr 2000 04:28:33 -0400
Reply-To: David Beierl <dbeierl@IBM.NET>
Sender: Vanagon Mailing List <vanagon@gerry.vanagon.com>
From: David Beierl <dbeierl@IBM.NET>
Subject: Re: LP tank level detection proposal, its Friday..
In-Reply-To: <8e.442bbf5.26392b8f@aol.com>
Content-Type: text/plain; charset="us-ascii"; format=flowed
At 01:35 4/27/2000, S Sittservl wrote:
>The standard add-on "barbecue" tank gauge is a pressure
>gauge, which measures the pressure of the gaseous propane
>that's above the liquid in the tank, on the theory that the pressure
>goes down as the liquid level drops.
I don't think that's a very good theory...I would expect the tank pressure
to vary directly with temperature regardless of the amount of liquid in the
tank. If you withdraw vapor at a given rate then a full tank will cool and
reduce pressure slower than a nearly empty one. Once you've gotten all the
liquid into vapor, the pressure will drop as you continue to withdraw
vapor...ok, I've done a bit of looking. Commercial propane has a vapor
pressure of roughly zero at -40F, 53 psi(gauge) at freezing, 110 psig at
70, 200 psig at 110. BBQ tanks, btw, are rated at either 240 or 300 psig
working pressure. As I recall, a propane torch running wide open at -20
will go out in a few minutes as the liquid cools (can't remember was it -20
or -30...had to use a hotplate instead).
The three-gallon tank holds about 0.4 cubic feet, which would be about 1000
BTU at 60F or 2000 BTU at 105F. The reefer burns something less than 650
BTU/hr, but the stove burners are rated at 5200 -- so I'd expect you'd get
10 to 20 minutes of burner in that range of temps, after the pressure
starts dropping. The reefer, OTOH, might run three or more hours. Either
way, not much warning.
Now, at a given consumption rate the tank will eventually cool to a stable
temperature -- hence pressure -- takes about 785 BTU/gallon to evaporate
the stuff. Someone a lot more energetic than I could work up a set of
curves for differential pressure or temp...
Two other things occur to me -- first, from experience, trying to fill the
bloody thing on a steep driveway in Scotland, I've observed that the 80%
gauge fitting is well aft on the tank. A little calculus should provide
fill level vs. angle so you could jack up the front and see if you had a
half-tank or not (I'm pretty sure that it took 2/3 tank or less from empty
in Scotland). That's eminently practical even if an awful nuisance.
The other would be to install bleed valves at half and quarter points in
the tank. Manchester might be willing to do it if you got a tank from
them. If you do it yourself you void the tank certification, I'm sure.
Oh yeah, you could always put a gas meter in the line and measure
consumption. But they're kinda bulky, and tend to be calibrated in
hundreds of cubic feet -- the tank holds somewhere around 100.
Ah Ha! Got it. Hang the tank on a load cell (S-shaped member with attached
strain gauges to feed an electronic scale) and just weigh it. Probably do
it for under $1k with commercial gear, lots less if you make the electronic
part. The commercial stuff has to deal with long stretches of wire between
the load cell and the electronics, so it has to have amazing common-mode
interference rejection $$$$ -- I bet this could be done for less than $50
in electronic parts, plus whatever load cells cost. It's a bridge circuit
driving an op amp. You could calibrate the output of that to drive the
water-level LEDs... Come to that, I've got a $40 bathroom scale with a
load cell in it. And it's set up to weigh in this sort of range -- it's
got a lever system to reduce people-weight. Unfortunately it self-zeroes
when you turn it on. But it's got the basic circuitry.
The mechanical details of getting the load cell up there will be a good
deal harder, I think.
And of course a few pounds of mud would give a false reading, but hey...
david
David Beierl - Providence, RI
http://pws.prserv.net/synergy/Vanagon/
'84 Westy "Dutiful Passage"
'85 GL "Poor Relation"