The Maxbiotics Ultrasonic Range sensor does a good job of measuring short distances. BSR has a how-to for using it to measure water softner level. It can be interfaced via RS-232 or with something like DS2760/DS2450/DS2438 to do A/D conversion on the 1-wire. They sell in the $25 range. If you are looking for resolutions of under 1" then it will not be a good choice.
I decided to give this Maxbotix sensor a try. Some interesting discoveries along the way in case anyone else is interested in experimenting.
I bought the sensor from "Sparkfun", about USD25. I found a really easy way to connect it to 1-Wire was to canibalise a spare hobby-boards solar sensor. I cut off the photodiode to use as a spare, then connected the Maxbotix sensor to the three pin-holes on the board designed for the (non-existent) humidity sensor. Three wires for GND, +5v, and "Analog". It steals power from the regulator on the solar board. That's it ! It works a treat. The VAD field in the DS2438 has a resolution of 10mV which is almost exactly the same as the Maxbotix sensor which reads 9.8mV per inch. I can read distances from 6 inches to 256 inches in 1 inch steps. To check the calibration I connected the Maxbotix serial output to a PC and it came out at 9.7mV per inch, which is only 1% off the supposed specification. Close enough for me, I correct this and convert to mm in software. The Maxbotix sensor itself is amazingly accurate according to my faithful tape-measure.
So far so good. But it can't be that easy ...
Problem 1 is that the Maxbotix sensor is very sensitive to any close up objects interfering with the beam. I had planned to mount the sensor in a PVC tube sticking out the top of the water tank, however the "beam width" is such that the tube would need to be a very large diameter. I ended up with it only 50mm or so above the top of the tank. That wouldn't matter except for a side effect of problem 2 ...
Problem 2 is that the sensor could not cope with the high humidity. I tried all sorts of things but could not find a way to keep the humidity down so in the end a "cheap" sensor had to be replaced with the "expensive" weatherproof version , this time about USD99. Damn! This one does indeed work perfectly in the wet conditions, but has a slightly different range (12-256 inches instead of 6-256). The minimum 12in range is unfortunate because of problem (1). In my case if the tank is full the water comes to 9 inches from the sensor, and I can't easily pull back the sensor because of the beam with issue. Always one for accuracy it annoys the heck out of me that I can't distinguish between "full" (at 9 inches) and "nearly full" (at 12 inches)! Below 12 inches it works perfectly.
So in conclusion, the sensor and solar-board makes an excellent "proximity sensor". I could imagine all sorts of uses but the tank-meter is maybe not the best application. My 1st tank has been working fine for a while now, but I have 3 more to go ... and at $99 a time I'm working on a plan-B based on the capacitance meter for the others. Knowing me it will take a few months, but if I remember I'll report back.