Sealing the cable end took some effort. I first dipped the end in "liquid tape" several times, allowing it to dry between coats - but the cable effectively shorted after only a few hours at depth. Either the liquid seal is not truly waterproof or, I think more likely, it does not adhere well to the cable's PVC insulation and allows infiltration under the ~45psi of pressure at ~-90' of depth. I finally solved that by pulling on the cable insulation end so that the wire ends retracted into the insulation sufficiently to then melt and crimp the insulation with flame-heated needle-nosed pliers. For good measure, I dipped that in liquid tape and covered it with heatshrink. So far, that seems to be effective.
I collected some initial data overnight, enough to see an apparent cyclic change - probably tidal. I only formatted the depth data to tenths so the curve is crude but here's what it saw: http://rightime.com/well_level.gif
I suspect the depth offset (I measured -18' but this data is centered on ~-12') might be due to condensation on the "dry" top cable section; the humidity in the closed well is probably 100%, and I have noticed dampness on that section when I've pulled it up. Recalibration necessary.
Capacitive water tank level determination
Well (pun intended), this has become a more complicated project - but a friend found a paper that reinforces the initial results I found - that the well level is very dynamic, affected by nearby ocean tide (Gulf, in my case), lunar tidal influence, geological mass squeeze (by lunar tide and seismic activity) and atmospheric pressure (which is itself also affected by lunar tide). For those interested, the effects are shown on Page 7 and later in this PDF:
http://www.google.com/url?sa=t&rct=j&q= ... edAVl41x7w Notably, the word drought does not appear in that paper.
I've encountered some persistent difficulty measuring the period of the '555 I use to infer depth by the effective capacitance of the 100' cable that hangs in the hole. I'm about to conclude that the irregularly I see is, at least in part, due to the sheer length of the cable which, despite being mostly submerged, might exhibit antenna effect. There appears to be not only a 60Hz component (not a surprise), but also lower-frequency cycling that I can't account for. I'm not sure how to proceed at the moment. Filtering is not obviously feasible since capacitance variation is what I'm detecting - by using it in a '555 RC. I'm hesitant to try using the ZX24's RCTime function, which would directly expose the processor to the same antenna effect - and we get some hefty lightning here.
Further, the cable seems ever-so-slightly hygroscopic. In the week it's been submerged the cable capacitance has slowly and continually increased as, I suspect, water permeates the PVC jacket. I've just pulled it out of the well and allowed it to dry in sunlight and, despite feeling dry to the hand, the capacitance continues to decrease.
So, I'm thinking of boring methods like a laser rangefinder - but the drop pipe and the wellhead seal makes that problematic. More thought required.
FWIW.
http://www.google.com/url?sa=t&rct=j&q= ... edAVl41x7w Notably, the word drought does not appear in that paper.
I've encountered some persistent difficulty measuring the period of the '555 I use to infer depth by the effective capacitance of the 100' cable that hangs in the hole. I'm about to conclude that the irregularly I see is, at least in part, due to the sheer length of the cable which, despite being mostly submerged, might exhibit antenna effect. There appears to be not only a 60Hz component (not a surprise), but also lower-frequency cycling that I can't account for. I'm not sure how to proceed at the moment. Filtering is not obviously feasible since capacitance variation is what I'm detecting - by using it in a '555 RC. I'm hesitant to try using the ZX24's RCTime function, which would directly expose the processor to the same antenna effect - and we get some hefty lightning here.
Further, the cable seems ever-so-slightly hygroscopic. In the week it's been submerged the cable capacitance has slowly and continually increased as, I suspect, water permeates the PVC jacket. I've just pulled it out of the well and allowed it to dry in sunlight and, despite feeling dry to the hand, the capacitance continues to decrease.
So, I'm thinking of boring methods like a laser rangefinder - but the drop pipe and the wellhead seal makes that problematic. More thought required.
FWIW.
Tom
Capacitive water tank level determination
Hi Tom,
I know you like to experiment with different ways to measure things, so I have some ideas you could try.
1. Wouldn't averaging a number of readings that you are making now remove the 60Hz and the other lower frequency component? (I assume you have already tried this. Wouldn't help with the hygroscopic effect.)
2. Put a small compressible bladder (like the squeese-bulb used when measuring blood pressure) on the end of a small (1/8") nylon tube long enough to be under the water and attach an electronic pressure sensor at the top end. Differential pressure sensor if you are worried about atmospheric pressure.
3. Dangle a small stainless steel cable down the well and measure the resistance between it and the well casing (if metallic) or a ground rod. Or two cables, one inside of a small plastic pipe (open ended) and the other tie-wrapped to the outside of the pipe. I only say stainless because you may not want copper. The water pipes in my house are copper so I suppose it would be OK.
4. On the twin-lead you are using now, remove some (1/8" to 1/4") insulation from both wires every 4" or so. Now the change in capacitance should change in measurable steps every 4" (or whatever resolution you need). This should lessen the effects of the hygroscopic problem.
5. Hang an open ended hard nylon tube (1/8" or so) in the water and use a pressure sensor at the top. The water level should change the pressure measured.
All I could think of right now. I haven't explored any of these ideas to any depth (pun intended), they just came to mind.
Tom W.
On 8/3/2012 5:56 PM, General wrote:
I know you like to experiment with different ways to measure things, so I have some ideas you could try.
1. Wouldn't averaging a number of readings that you are making now remove the 60Hz and the other lower frequency component? (I assume you have already tried this. Wouldn't help with the hygroscopic effect.)
2. Put a small compressible bladder (like the squeese-bulb used when measuring blood pressure) on the end of a small (1/8") nylon tube long enough to be under the water and attach an electronic pressure sensor at the top end. Differential pressure sensor if you are worried about atmospheric pressure.
3. Dangle a small stainless steel cable down the well and measure the resistance between it and the well casing (if metallic) or a ground rod. Or two cables, one inside of a small plastic pipe (open ended) and the other tie-wrapped to the outside of the pipe. I only say stainless because you may not want copper. The water pipes in my house are copper so I suppose it would be OK.
4. On the twin-lead you are using now, remove some (1/8" to 1/4") insulation from both wires every 4" or so. Now the change in capacitance should change in measurable steps every 4" (or whatever resolution you need). This should lessen the effects of the hygroscopic problem.
5. Hang an open ended hard nylon tube (1/8" or so) in the water and use a pressure sensor at the top. The water level should change the pressure measured.
All I could think of right now. I haven't explored any of these ideas to any depth (pun intended), they just came to mind.
Tom W.
On 8/3/2012 5:56 PM, General wrote:
Well (pun intended), this has become a more complicated project - but a friend found a paper that reinforces the initial results I found - that the well level is very dynamic, affected by nearby ocean tide (Gulf, in my case), lunar tidal influence, geological mass squeeze (by lunar tide and seismic activity) and atmospheric pressure (which is itself also affected by lunar tide). For those interested, the effects are shown on Page 7 and later in this PDF:
http://www.google.com/url?sa=t&rct=j&q=&esrc=s&source=web&cd=1&ved=0CEsQFjAA&url=http%3A%2F%2Fpubs.usgs.gov%2Fwri%2Fwri034267%2Fwri03_4267.pdf&ei=6tkVUL-QK4Kk8QTI2IHYDA&usg=AFQjCNEsoXJ4CaDeEI5eeU_yedAVl41x7w
I've encountered some persistent difficulty measuring the period of the '555 I use to infer depth by the effective capacitance of the 100' cable that hangs in the hole. I'm about to conclude that the irregularly I see is, at least in part, due to the sheer length of the cable which, despite being mostly submerged, might exhibit antenna effect. There appears to be not only a 60Hz component (not a surprise), but also lower-frequency cycling that I can't account for. I'm not sure how to proceed at the moment. Filtering is not obviously feasible since capacitance variation is what I'm detecting - by using it in a '555 RC. I'm hesitant to try using the ZX24's RCTime function, which would directly expose the processor to the same antenna effect - and we get some hefty lightning here.
Further, the cable seems ever-so-slightly hygroscopic. In the week it's been submerged the cable capacitance has slowly and continually increased as, I suspect, water permeates the PVC jacket. I've just pulled it out of the well and allowed it to dry in sunlight and, despite feeling dry to the hand, the capacitance continues to decrease.
So, I'm thinking of boring methods like a laser rangefinder - but the drop pipe and the wellhead seal makes that problematic. More thought required.
FWIW.
Tom
http://www.ustream.tv/channel/bowcam
http://www.ustream.tv/channel/cape-coral-marine-radio VHF
http://67.207.143.181/vlf9.m3u Lightning, spherics
All good ideas, Tom.
Yes, I've massaged the data to death - literally - using an integral function, not a true average of many stored samples, to the point where the result no longer resembles the measured data. Whether using a true average of stored data or an integral of deltas, at some point the result takes a random walk.
The root problem is that the measurement noise is not Gaussian but has a distinct bias. The 60Hz stuff, if a sine, will smooth out but odd large and irregular spikes pull the result toward one polarity, and they seem to come in bursts. If enough conditioning is applied to effectively eliminate them, the result drifts away from reality. Windowing the measurements - outright discarding outliers in either direction - might reduce that.
While a lot of work, your stepped-by-perforation cable idea is intriguing, but I doubt it would work for capacitance (direct water contact looks like a short at Hi-Z, so many of them submerged might become indeterminable); maybe it would work for resistance. A drill bit or a burr could perhaps expose a repeatable area of copper at each step.
Resistance of a single wire against Earth ground should work, too, although I expect it would be pH-sensitive which, I know from previous work, varies. The well casing is heavy PVC so the conductive path would be out the bottom of the casing, something like 200' in this 360' hole. Stainless is probably wise; we have relatively corrosive water so a copper wire might become oxide-coated. That sounds easy to try, though.
A captured air volume in a dip tube has a long-term problem: the air slowly dissolves into the water, reducing the head over time. One way to counter that is to feed the tube with slow sufficiently-pressurized air to emit bubbles at the bottom - so the tube is always filled with air whose pressure is a function of depth. That would require about 50psi at 100', albeit with little required volume.
Offhand, I can't think of a problem with a bulb at the bottom - except that it would want to float so it would need a ballast to keep it there. I suspect it would also need to be pressurized to prevent its collapse at depth, and the tube might need to be rigid since it's a closed system.
Good thoughts.
Yes, I've massaged the data to death - literally - using an integral function, not a true average of many stored samples, to the point where the result no longer resembles the measured data. Whether using a true average of stored data or an integral of deltas, at some point the result takes a random walk.
The root problem is that the measurement noise is not Gaussian but has a distinct bias. The 60Hz stuff, if a sine, will smooth out but odd large and irregular spikes pull the result toward one polarity, and they seem to come in bursts. If enough conditioning is applied to effectively eliminate them, the result drifts away from reality. Windowing the measurements - outright discarding outliers in either direction - might reduce that.
While a lot of work, your stepped-by-perforation cable idea is intriguing, but I doubt it would work for capacitance (direct water contact looks like a short at Hi-Z, so many of them submerged might become indeterminable); maybe it would work for resistance. A drill bit or a burr could perhaps expose a repeatable area of copper at each step.
Resistance of a single wire against Earth ground should work, too, although I expect it would be pH-sensitive which, I know from previous work, varies. The well casing is heavy PVC so the conductive path would be out the bottom of the casing, something like 200' in this 360' hole. Stainless is probably wise; we have relatively corrosive water so a copper wire might become oxide-coated. That sounds easy to try, though.
A captured air volume in a dip tube has a long-term problem: the air slowly dissolves into the water, reducing the head over time. One way to counter that is to feed the tube with slow sufficiently-pressurized air to emit bubbles at the bottom - so the tube is always filled with air whose pressure is a function of depth. That would require about 50psi at 100', albeit with little required volume.
Offhand, I can't think of a problem with a bulb at the bottom - except that it would want to float so it would need a ballast to keep it there. I suspect it would also need to be pressurized to prevent its collapse at depth, and the tube might need to be rigid since it's a closed system.
Good thoughts.
Tom
I had a conversation with an engineer at one of Cape Coral's Reverse Osmosis water treatment plants, finally - and got an invitation to tour the facility (the largest high-pressure RO plant in operation, apparently) - after 10 years of no public tours, post-9/11.
He says there is no evidence that the deep aquifers, at least in Florida, are affected by drought. He reiterated that the recharging water source is rainwater from north Florida and Georgia, and we draw water that fell as rain there 10,000 years ago.
The volume of available water here is enormous. Still, what is drawn from the aquifers is controlled by the Florida Department of Natural Resources, the EPA and several other government entities. Even farmers and grove fruit growers have meters - and lockouts - on their private wells, and are regulated.
More on topic, he said that they use absolute pressure gauges hanging on the end of a cable to measure monitor well heads - and that he isn't aware of a capacitive method. Maybe that's for good reason.
He says there is no evidence that the deep aquifers, at least in Florida, are affected by drought. He reiterated that the recharging water source is rainwater from north Florida and Georgia, and we draw water that fell as rain there 10,000 years ago.
The volume of available water here is enormous. Still, what is drawn from the aquifers is controlled by the Florida Department of Natural Resources, the EPA and several other government entities. Even farmers and grove fruit growers have meters - and lockouts - on their private wells, and are regulated.
More on topic, he said that they use absolute pressure gauges hanging on the end of a cable to measure monitor well heads - and that he isn't aware of a capacitive method. Maybe that's for good reason.
Tom
10K Yrs! Wow. So I guess flushing a dye marker in a toilet in Georgia and watching for it to show up in your well just isn't going to happen!
Thanks for posting the update.
I would think with water being essentially incompressible, (in its liquid state), measuring pressure may not really tell one much information without radar(?) acquired maps of the aquifer's size, shape, volume, deapth, etc.
You could certainly track it, and plot it on nice graphs, but how does one relate it to whether or not the well is going to run dry tomorrow, next week, next decade, ?
JC
Thanks for posting the update.
I would think with water being essentially incompressible, (in its liquid state), measuring pressure may not really tell one much information without radar(?) acquired maps of the aquifer's size, shape, volume, deapth, etc.
You could certainly track it, and plot it on nice graphs, but how does one relate it to whether or not the well is going to run dry tomorrow, next week, next decade, ?
JC
Not in our lifetimes, anyway. They apparently estimate the rate of flow from well-to-well but not over a 300-mile baseline. A radioactive marker might not even survive a one-mile path; at a few inches per day, say three, that would take about 57 years. One year would cover 91 feet, or about eight feet per month. Pretty slow, despite most of south Florida civilization's water demand of many millions of gallons per day!DocJC wrote:... I guess flushing a dye marker in a toilet in Georgia and watching for it to show up in your well just isn't going to happen...
I learned also that some utilities store excess groundwater in wet months, mostly treated river water, in deep aquifers via ASR wells, some 3000 feet deep but the practice has lost favor after the revelation that oxygenated and chlorinated water - drinking quality, as required by the State - that is injected into deep wells dissolves normally-insoluble arsenic from the strata, so when later recovered the water is unable to pass the State's quality test and can't be legally used for drinking. As an unintended consequence, treated potable water poisons itself - and the natural aquifer water - when stored in a deep Florida aquifer. Way OT, but ironically interesting, I thought.
Tom