I'm laying out PCBs next week. I plan to get 10 PCBs made. If anybody wants a coupla PCBs let me know. These will be non-SS/ non-SM, but 2-sided. Not very tight layout. All SMT. The tiny sensor board will have 5 points in a row so that the PIC10F200 can be programmed after tacking on the wires, and it will need to be "sawed away" from the control PCB. BTW: I have located cheap sources for everything, even the relays (ZETTLER AZ971, app $1.50) and the washing machine filler control solenoid (about $4). --Bob Bob Axtell wrote: > OK, here is the design: > > At the Fountain: > > 1. At the fountain is a white plumbing pipe, very thin wall. This one has a > water connection (ribbed stub) on the side and has an open top and bottom. > It is the water filling method, and doubles as the water sensor as well. > This looks to be a standard plumbing fixture, cost about $1 USD. > > 2. The outside of the thin wall pipe has 4 "plates" of copper adhesive foil, > rectangles, two above the other two. These plates do NOT contact each other > and are 2"x W by 2" high. There should be a 1/10" gap between the > plates. Small > insulated wires are soldered to the plates, going up on the outside of > the plates > to the top of the pipe. > > 3. Craft stores here in the Colonies named Michael's sell a two-part > clear epoxy > that hardens rock hard. A batch should be made up and the lower part of > the pipe, > including all of the copper, should be dipped several times until it has > a coating of > about 1-1.5mm of plastic. Allow this to harden for 24 hrs. You can now > run tests > on the water sensor, if you like. NO conductor touches the water, EVER. > > 4. The water sensor is composed of the 4 plates of the water pipe. The > lower two > plates are to be tied (1) to a PIC input and (2) to PIC GND . The upper > two plates > are tied to another PIC input and also to PIC GND. The upper plates are > the reference > capacitor and the lower plates are the sensor capacitor. > > 5. Water in the fountain floods the lower set of plates , INSIDE the > pipe. This dramatically > increases the capacitance of the lower "capacitance". The water does NOT > reach the upper > plates, which act as a reference; because the "PIC" input capacitor > measuring method is subject > to temperature variants, the fact that both caps are measured allows us > to make a more reliable > judgment. The ratio of ACTIVE count to REF count is about 4:1 with water > detected. Our water > is pretty ho-hum, has a little magnesium and a little chlorine. > > 6. At the top of the pipe is a PIC10F200 on a 1/4" x 1/2" PCB encased in > plastic epoxy. The > PIC, when powered, makes several capacitor measurements by HOW MANY TEST > LOOPS > are needed before the capacitor no longer holds a charge. The PIC will > make 16 passes for > each capacitor, then send back both averages to the main control unit. > The data is sent back > using Manchester code, with a bit time of approx 5mS/bit. The data > stream is two '1' start > bits, followed by 8 REFERENCE data bits, followed by 8 SENSOR data bits, > followed by a > dummy STOP bit. This keeps repeating as long as it is powered. So the > decision about the > water level is decided AWAY from the water. > > The Controller to Fountain Cable: > > 1. The water sensor itself needs 5V, GND, and an output signal, so a > cable of 4 conductors is > used, GND is doubled and the cable is shielded. This will be about 25' > (7m) long. > > 2. The water pump needs AC mains (here 115VAC 60Hz 30w). > > These two cables will be enclosed into a waterproof outdoor conduit. > This will be buried 12" > down. > > Note: there will be a water hose as well; I am using flexible hose > enclosed by neoprene split > hose, and also buried at 12". The neoprene will prevent the other hose > from being closed by > the weight of the soil above it. The hose can be quite small; I am using > 1/4" hose. This keeps > the fountain filled with water. > > Controller: > > 1. Mounted on the side of the house is a small waterproof NEMA box > containing electronics > and controls. A small waterproof PB switch, GREEN/RED LED, and > photodiode is mounted > on the outside. > > 2. Two 30A 12V relays operate the PUMP and the WATER SOLENOID. While > neither of > these draw 30A, a larger contact area in general means a longer life. > Internally the 115VAC > mains is protected by an internal 1A "slowblow" fuse. > > 3. A internal PCB contains the following: a PIC12F675 DIP (socketed) , > the two 12V relays, > a 12V & 5V power supply for the PCB and fountain sensor, a photodiode > (for sensing daylight), > and a couple of cables & connectors. > > 4. Action: (A) the fountain does NOT pump during the night, as the > photodiode detects no daylight. > But during the night, every 15 minutes, the water sensor turns on and > the water level is determined. > If water is needed, the sensor is turned off and the water flows for > about one minute. This cycle repeats > all night. > > 5. Action: (B) during the day, the fountain pumps. Every 5 minutes, the > water sensor turns on and the > water level is determined; if water is needed, the sensor is turned off, > water is turned on for one minute. > The reason to do this more often is that during the day, evaporation is > MUCH higher. > > 6. The RED LED will wink if both sensors appear to be identical; this > might indicate that the water has > run totally dry; this will also cause the pump to shut off. > > 7. Normally operation is automatic; if the switch is pressed, whatever > is happening will stop for one hour. > For example, if the pump was running normally, it will now STOP. If it > was previously stopped manually, > it will now start running again. This is normally used to allow the user > to clean bird feathers out of the > fountain. > > - - - > > Flames expected. > > --Bob > > > > Vasile Surducan wrote: > >> On 4/9/06, Bob Axtell wrote: >> >> >>> My wife bought a ceramic fountain. It has a pump that draws water >>> from the bottom and pumps it to the top, so the birds can bathe in it >>> and it srips down to the bottom again. Looks neat, works nice, kills >>> some of the road noise. >>> >>> The problem is that in So Arizona, the water evaporates so fast, I >>> can't discipline myself to keep it filled. So... I am gonna design a >>> slick way to sense when the water level is low so it can kick on a >>> solenoid and pump some back into it. I don't want to use anything >>> ugly, or big (like a float valve). I had in mind a non-contact sensor, >>> perhaps a capacitor. The whole sensor/solenoid deal could then >>> be automatic. >>> >>> What would work reliably? I am leaning toward a probe with two >>> insulated contacts, that I can make part of a tuned circuit; when >>> the water is missing between the probes, the frequency is at a >>> certain range, with the water between the probes, the frequency >>> is detuned. Another idea was to mount a tiny magnetic float in >>> a plastic tube, and when the float is high enough, it triggers a >>> hall device. >>> >>> Any other ideas? >>> >>> >> A fountain automation it can't be without a PIC. Telemetry, ultrasonic >> range finder, laser beams...using top DSP or ARM cores... >> :) >> Maybe a simple mercury sealed switch (floating on the water surface) >> connected directly (or using a relay) into your low power pump circuit >> ? No PIC, allowing a long distance to pump and safe histeresys. The >> smallest I've seen looks like a watch bulb. >> >> >> >>> >>> >> >> > > -- http://www.piclist.com PIC/SX FAQ & list archive View/change your membership options at http://mailman.mit.edu/mailman/listinfo/piclist