Borehole well water level sensing

Jeremy Harris

Senior Member
Wow, that's quite saga! I've had good experience except when a new replacement pump failed after 2 hours and they had to pull all 80 meters twice in the same day. Yes, 10% or more of US households have a well, it's probably closer to 50% where I live.

Since we've already strayed more than an little, may I ask what you are using for iron filtering? I have BIRM followed by a potassium salt softener to tame mine.

Nice job on the camera :)
Thanks for the kind words about the camera, it was the only thing I could do to prove, beyond any doubt, that the bottom of the well was capped at 53m. It was a bit of work to put together, but I've since built a very much smaller one that I am keeping in case I ever want to look down there again (it's smaller so that it can go down past the pump with plenty of room to spare).

I started off with a filter similar to your BIRM filter, but using a mix of Aquamandix and sand, with a renewable air pocket at the top, to make sure the oxidation was effective. The air pocket is replenished every time the Clack valve goes through a backwash cycle. The water still had a slight smell and taste, which a carbon block filter would remove, but this meant changing filters. So, after some analysis I decided to add a home made oxidation tank ahead of the system (i.e. between the pump and the pressure tanks). I made an ozone injector, and a small injection venturi (very similar to the Clack aerator) with the Dole valve on the bypass. When the pump comes on it turns on the Ozone generator, and the pressure drop across the Dole valve ensures that the injector venturi always works OK. This sucks in enough ozone to both oxidise out pretty much everything in the water and kill all the bugs. The aeration tank has a float valve to keep an air/ozone pocket at the top, but discharge excess air/ozone outside. In practice the ozone is all dissolved in the incoming water, so all I get out of the air outlet is air, with not significant ozone smell. I have a solenoid valve on the aeration tank bottom outlet, that opens for about 20 seconds every time the pump cycles, and blows any crud in the bottom of the tank to the waste drain.

It all seems to work very well. Adding ozone right at the front of the system means that the water is always very well oxygenated and so the Aquamandix/sand filter doesn't have much to do. Knowing what I know no I'd never have bought the big Aquamandix/sand filter, but would have fitted a much smaller backwashing Filter Ag filter, just to catch any particles of ferric oxide. Aquamandix is very similar to BIRM, essentially it's manganese dioxide, but it's more tolerant of a bit of hydrogen sulphide, I was told.

I too have a softener, as the water is a bit hard, and I suspect that the softener helps remove anything left in the water.

I've been impressed by the effect of adding ozone, not just for getting rid of the iron, but also because it makes the water taste so much better. We have really nice water now, good enough for making a really good cup of tea.............
 

Jeremy Harris

Senior Member
Water is about 12 metres below ground level according to the government's on-line water atlas. The (permanent) outer casing is about 125mm. I think it's that size to fit into the temporary 150mm/6" casing that is used while drilling the hole. The assembled 125mm gear is lowered into the outer casing before the outer case is recovered. The ground is quite sandy.
Sounds very similar to our borehole. In our case they drilled with a 200mm drill and 200mm diameter steel temporary casing, then lowered in the 125mm OD PVC liner (slotted at the base and fitted with an end cap) surrounding this with crushed glass as a filter. They pulled the steel liner out and filled around the remaining PVC liner with Bentonite clay, to seal it to the surrounding ground and stop rain water run off getting in.

We're lucky, in that our well apparently has the potential to deliver around 4000 litres per hour maximum, as that's the calculated recharge rate from the aquifer that's around 45m down. Water comes into our well under some pressure from below, as the resting level is usually only around 4m down, so there's about a 40m head of pressure in the aquifer. There are a dozen or so boreholes and wells around us, and one on the opposite side of valley is set with a slightly lower ground level and is artesian most of the year around. Sadly it's not in use, as there's mains water that side of the valley, and the 70 or 80 year old steel casing has corroded, so we have a "spring" that runs across the lane on the other side of the stream from us. It's a nuisance, as on mornings like today (-6 deg C first thing) the lane is like a skating rink from the ice caused by the freezing well water. I keep meaning to find the land owner and go and offer to plug the well properly for him, just to get rid of the nuisance.
 

Jeremy Harris

Senior Member
I've been looking around the web to try and see if there are commercial units that operate acoustically, and if so, try and get an insight into how they work.

It seems that none (at least none that I could find) use the resonant frequency of the borehole; all seem to use acoustic time-of-flight measurement. I've found a model that uses 200Hz pulses, and another (that claims to be "better") that uses 50Hz pulses. It's hard to determine exactly how these things work, but it seems that they transmit a short burst at a set frequency (and I'd guess the burst is a few whole cycles long) and then listen for the first echo at that frequency.

None seem to use ultrasonic sensors, because of the multiple echoes from the wall of the pipe. It seems that as soon as the sound wavelength gets to be significantly longer than the diameter of the pipe, the pipe just ducts the sound up and down, and even around bends (very similar to the old "speaking tube" that was fitted between cockpits on early aircraft (and on ships between the bridge and engine room, I believe).

The bore of our well is 119mm, which is equivalent to the frequency of sound at about 2800Hz. Using the general rule that about 1/10th of this is probably OK, then it looks like a pulse at a frequency of 200 to 300Hz should be OK. I'm guessing that the unit that uses 50Hz pulses does so in order to work more effectively with larger diameter wells.

Clearly there are digital ways to do this, but my thoughts are that it may well be easier to build the transmitter and receiver as separate analogue parts, then use a Picaxe to to the time-of-flight measurement. If I use 200Hz, and I want to be sure that at the shallowest depth the pulse is always at least half the time-of-flight duration, then I need to keep the number of cycles down to a fairly low figure. I reckon 10 cycles should be enough. I remember using some phase locked loop tone detectors years ago, and probably still have some around, the NE567. If set to have a fairly wide bandwidth (probably OK in the relative quiet of the borehole) then one of those should just about work with 10 cycles. I could increase this to about 15 cycles and still be OK, I think.

This now looks to be a bit more manageable. A Picaxe could send a burst of 200Hz pulses, through a filter and amplifier to a small waterproof speaker (one of the ones with a Mylar diaphragm should be OK). The speaker would probably also work fine as a microphone, so an amplifier (with input limiting to avoid overload on transmit) followed by some filtering and then an NE567 should detect the return tone burst. I can't see an easy way to get the Picaxe to also work as a tone detector, as 200Hz seems just too high a frequency for several consecutive measurements to be made.

The commercial instruments have a small "sound tube" that can be fitted into a port on the well top plate, so I shall try and do something similar, to fit to the 1 1/2" threaded port I already have.

Time to get some parts and start experimenting to see how it works.
 

techElder

Well-known member
Not trying to discourage you, but ultrasound at 40 meters distance will need more amplification and filtering. You should also think about gating the return echoes so you will ignore everything except the interesting echoes.
 

Jeremy Harris

Senior Member
Not trying to discourage you, but ultrasound at 40 meters distance will need more amplification and filtering. You should also think about gating the return echoes so you will ignore everything except the interesting echoes.
Yes, ultrasound won't work, as I mentioned, as the wavelength is far too short. The acoustic systems I've been able to find data on use much lower sound frequencies, as I mentioned, 200Hz in one case and 50Hz in another.

By using such low frequencies the sound gets ducted down the borehole, not reflected of every small obstruction and the borehole walls as would be the case with ultrasound. The reason is as I mentioned above, the wavelength of the sound. Ultrasonic transducers are around 40kHz have a wavelength that is shorter than the bore of the well, around 83mm at 0 deg C. Decreasing the frequency increases the wavelength, and as long as the wavelength is significantly longer than the internal diameter of the borehole (in my case 119mm, or about 2.8KHz) then the sound doesn't reflect off the walls to any appreciable degree (well, it does, but there's a lot of cancellation and absorption).

So, if I opt to use 200 to 300Hz sound pulses, rather than the 40kHz sound pulses from an ultrasonic unit, then most of the sound will go down the well and be reflected back from the water surface at the bottom.

I could use two ways of doing the measurement, rather similar in some ways to two types of radar. I could use a burst of, say, 200Hz, around 10 to 15 cycles long, and listen for the specific 200Hz return echo, or I could use a continuous wave technique, with cycle counting. Of the two, I think the pulsed low frequency sonar type will be the easiest to implement. I doubt that the transmitted signal needs to be anything more complex than a set number of pulses at 200Hz from a Picaxe output, passed through a low pass filter to a small speaker. The same speaker should also work well enough as a receiver, I think. I've dug out a Mylar speaker from an old car alarm unit, around 70mm in diameter. I'm thinking of sticking it to the end of a bit of pipe, with some adapters to form an acoustic filter, and I could then drop the shorter pipe into the 1 1/2" capped hole I have on the well top plate. Having around 700 to 800mm of pipe sticking down from the top wouldn't be a problem at all, and it should improve the signal to noise ratio by filtering out everything outside the tuned frequency of the pipe, with luck.

I could probably fit the speaker and all the electronics, plus the pipe reducer, into a sealed plastic box that just screws on to the 1 1/2" diameter well head stub pipe that currently has a threaded cap on it. It would then just be a matter of powering it and getting data back, and luckily I fitted an extra length of water pipe underground, that at the moment just pokes into the well head chamber. I reckon I could use this as a cable duct to run a power and data cable from the water treatment room out to the well head in the garden.

This all assumes I can get it to work!
 

Jeremy Harris

Senior Member
FWIW the Maxbotix website has a discouraging article about using ultrasonic sensors in pipes.
Yes, I tried it some time ago when I was building a four sensor ultrasonic doppler wind speed and direction measurement system. I thought tucking the sensors back in the pipes would protect them, but instead I created problems with reflections.

I'll say again, I'm not looking at using high frequency sound here, but sound at a very much lower frequency. I'm basing this on what I've been able to glean about the operating method these two acoustic well water depth sensors use:

http://www.globalw.com/products/wl600.html

http://www.enoscientific.com/well-sounder-PRO.htm

Both use low frequency sound, 50Hz to 200Hz as far as I've been able to find out.
 

techElder

Well-known member
OT I've built ultrasonic pipe length gauges that used a reflection at the "interface" at the end of a tube. I can't see a way to use this phenomenon in Jeremy's app.
 

Jeremy Harris

Senior Member
Is it possible to measure power used by the pump for a known quantity of water ?
Yes, measuring the flow rate and time is pretty easy. The problem is, I'm trying to monitor the depth of water in the borehole, and that varies with the seasons as well as with the pumped rate. There's not a fixed relationship between the pump flow rate and the well level, as it depend on the well recharge rate, and that depends on the state of the aquifer that the well is drawing from, plus any sand clogging that may occur.
 

geezer88

Senior Member
Actually, a pump needs more input to lift water higher, so there is, at least theoretically, a relationship between water level and pump power input. Unfortunately, if you look at pump operational curves, the power to deliver water curve is fairly flat in the area most domestic pumps operate. This would make the strategy hard to implement. I like the concept, though, and if a particular pump had a steep curve, it might be a way to do your water level monitoring.

tom
 

Pongo

Senior Member
Actually, a pump needs more input to lift water higher, so there is, at least theoretically, a relationship between water level and pump power input. Unfortunately, if you look at pump operational curves, the power to deliver water curve is fairly flat in the area most domestic pumps operate. This would make the strategy hard to implement. I like the concept, though, and if a particular pump had a steep curve, it might be a way to do your water level monitoring.

tom
Years ago, before I was a well owner, I met an engineer who was logging the pump current to track water depth. He was smart guy (Intel) and was convinced that it worked, but I don't know anything about well depth or pump type. Based on where he lived I think it had to be a fairly deep one.
 

BESQUEUT

Senior Member
Years ago, before I was a well owner, I met an engineer who was logging the pump current to track water depth. He was smart guy (Intel) and was convinced that it worked, but I don't know anything about well depth or pump type. Based on where he lived I think it had to be a fairly deep one.
pump current is not enough : you also have to track water flow.
1 kWh=3.6 x 10^6 J
1J = 1 N x 1 m
1 J = M(kg) x 9.81 x H (m)
Of course
1 m^3 (water) = 9810 N

Efficiency for centrifugal pump :
0.4 to 0.7
this number is quite constant for a known pump, so Can be measured.

With others words :
H (meters)= K x Current / Water flow

H from water level to ground

Water level = K x Water flow / pump current
 
Last edited:

Pongo

Senior Member
In a typical well/pump/pressure tank system, as used here, I suspect that the flow vs time curve at any particular head pressure is quite predictable allowing the pump current vs time to be used as a "indication" of head/water level. Anyway, other than panic and drill a deeper well ($20K minimum), I'm not really sure what I would do with the data so pretty soon after becoming a well owner I decided not to worry much about it.

@Jeremy Harris: The taste of my water is acceptable, although it's not exactly Malvern water. Aeration/ozone would be a good pre-treatment but it would be a huge increase in complexity and cost for a minor improvement - I'm too old to hassle with something like that now lol.
 

premelec

Senior Member
FWIW there are commercial units that monitor pumps and watch for faults - One my system uses is "Pump Saver" - gives error codes - flashing lights - when various faults or unusual operation occurs [suck out, dead head etc] - this implies that a bunch of information can be surmised from the pump current... ;-0
 

BESQUEUT

Senior Member
Actually, a pump needs more input to lift water higher, so there is, at least theoretically, a relationship between water level and pump power input.
This is not only theory.
Yes, I know : some are thinking that science is not true. Probably they are also thinking that Picaxes are God driven....

If not :
To be exact, power is H * water flow + heating
if heating is not a constant, there is probably a malfunction...
But it is probably a good idea to also track water temperature.

also
H is not exactly from water level to ground
H is from water level to free air level. If a tank is used, this is probably a constant.
 

Jeremy Harris

Senior Member
One problem I have with using pump current and flow is that I have an artificial restriction on the pump flow that might create a problem, or might not. My system is very close to a constant flow rate one, as I have a Dole valve to control the flow rate. It's there because if I pump the well too hard it draws in running sand; keep the pump rate down and it doesn't draw any sand at all.

I've dug out the old Mylar cone speaker and will have a go at the weekend making up an experimental rig, just to transmit tone pulses and see what I get back. I have a length of left over drain pipe that I can use for testing - saves going out in the cold and working at the well head! The aim initially is to transmit pulses of a dozen or so cycles at 200Hz down the pipe and see what comes back with a 'scope on the speaker. If it looks like I can get something that can be cleaned up well enough to be fed directly into a Picaxe for a pulsin measurement, then I think the problem will be cracked. I'll probably add a DS18B20 temperature sensor as well, just to be sure the speed of sound is corrected properly, but this may be overkill. My well seems to be pretty close to 8 deg C all year around, so I don't think there will be any temperature change to be worth worrying about.

I already have a 12V isolated signal that is active when the pump is running, so I can use that to log pump run times. I'm thinking of just logging water depth, temperature and pump run time, as I can derive water volume used from the run time, because of the Dole valve. I'm really just curious to see how the water depth changes, both seasonally and with use. Every time I've dropped my home made beeper tape down the well the resting level has been the same, around 4m from ground level, so I don't think it changes much, but it'll be interesting to see for sure.
 

techElder

Well-known member
Every time I've dropped my home made beeper tape down the well the resting level has been the same, around 4m from ground level, so I don't think it changes much
Better start digging a 10 meter hole or a high platform so you can test for resolution of the measuring device you come up with. :D
 

SAborn

Senior Member
A rather simple passive method is a small plastic tube filled with water and plugged at the bottom, then suspended down the well.
The top end of the tube is attached to a weight measuring device like a load cell or strain gauge of some sort, you simply monitor the weight of the tube.
As the water level drops the tube gets heavier.
 

geezer88

Senior Member
Nice idea! Shouldn't be too hard to model that with a spreadsheet to get an idea of change in weight of the water filled tube as the well level changes.

tom
 

Jeremy Harris

Senior Member
A rather simple passive method is a small plastic tube filled with water and plugged at the bottom, then suspended down the well.
The top end of the tube is attached to a weight measuring device like a load cell or strain gauge of some sort, you simply monitor the weight of the tube.
As the water level drops the tube gets heavier.
That is a neat idea.

I've used the HX711 bridge amplifier/load cell A to D converter modules with a Picaxe, with a bridge-type pressure gauge. Wiring one up to a cheap load cell should be easy enough, and the popular auction site is awash with cheap load cells, as used in scales etc.

I could probably get the whole thing to fit to the threaded cap I already have on the well top plate, in the same way as I was thinking of doing with the acoustic sensor. I probably need about 12m at the most of thin pipe, which would be well clear of the pump which is 25m down. I have several metres of both 10mm and 8mm OD HDPE pipe that might well do the job, I'll have to have a look around and see what I can get in the way of load cells. A 1kg one should do the job OK, perhaps like this one: http://www.ebay.co.uk/itm/Electronic-Balance-Weighing-Load-Cell-Sensor-0-1Kg-S/182204380208?_trksid=p2047675.c100005.m1851&_trkparms=aid=222007&algo=SIC.MBE&ao=2&asc=38661&meid=1827af06a4a2437ebbcc770dcac8d670&pid=100005&rk=3&rkt=6&sd=272011056209

At only £1.41, including postage, it's got to be worth buying one to try. I'm tempted to buy a few of different ratings, just so I can experiment a bit. At that price it's not really going to break the bank.
 

Steve2381

Senior Member
Maybe I am missing something, but could you not use a laser measuring device. I have a very accurate one here that does 80m easily.
Not sure if the pipe would add reflections to the beam, but it might work
 

Jeremy Harris

Senior Member
Maybe I am missing something, but could you not use a laser measuring device. I have a very accurate one here that does 80m easily.
Not sure if the pipe would add reflections to the beam, but it might work
Much earlier in this thread we discussed laser solutions, but they seem prohibitively expensive. The cheap ones are proximity-only, and to get accurate ranging seems to need a pretty expensive unit. The relevant discussion starts at the beginning, here: http://www.picaxeforum.co.uk/showthread.php?29471-Borehole-well-water-level-sensing&p=304395&viewfull=1#post304395
 

Camerart

Member
Hi,
Here's another idea for you to peruse.
I've read that some altimeter chips are quite accurate. If you had on at the top of the well and another floating, subtract the difference.
What do you think?
Camerart.
 

Jeremy Harris

Senior Member
Hi,
Here's another idea for you to peruse.
I've read that some altimeter chips are quite accurate. If you had on at the top of the well and another floating, subtract the difference.
What do you think?
Camerart.
A pressure sensor was on the list of potential solutions, but there is a slight entanglement risk with putting more wires down the hole. As mentioned earlier in this thread, I have a 100psi gauge pressure transducer, that I could fit into a watertight housing and drop to well below the lowest water level I expect to get, connecting it to the surface with a tube, that could also contain the power and data wires, as well as act as the surface air pressure reference. In essence, this would be the same as your suggestion, but using a single gauge, rather than absolute, pressure sensor, to give the differential pressure directly.

I think my favourite approach would be a completely non-contact one, like a low frequency acoustic time-of-flight or laser approach, fitted at the well head. This avoids the need to put yet more stuff down the well, which would be a good thing. Following that, the idea of a length of thin, water-filled pipe (or just end ballasted and sealed) dropped down just low enough to be in the water at its lowest depth and connected to a weighing system is the next favourite, as a relatively short, smooth, pipe isn't as likely to get tangled up as something like wires or a longer pipe. Last night I though of a variation on this idea, by just fitting a load cell inline with the pump safety wire, and pulling that wire tight, so it was taking most of the pump and MDPE pipe weight. The level of water in the well would vary the effective weight of the pump and pipe, and so might be good enough as a way to determine water depth. As the safety wire, pump and pump cable are all attached to a bolted-on top plate, I could even just arrange to put a load-sensing ring around under the top plate, and measure the change in load that way.
 

techElder

Well-known member
As the safety wire, pump and pump cable are all attached to a bolted-on top plate, I could even just arrange to put a load-sensing ring around under the top plate, and measure the change in load that way.
Thumbs up on not interfering with the basic function of the "safety" wire.

With this approach you might even be able to add some leverage to the load-sensing mechanics.
 

Jeremy Harris

Senior Member
Thumbs up on not interfering with the basic function of the "safety" wire.

With this approach you might even be able to add some leverage to the load-sensing mechanics.
All it really needs is a way to include load sensing between the top plate and the flange that it's bolted to. The top plate doesn't really need to be bolted down, it just has to be located in place and seal the top of the well. I remember years ago using annular load cells when we wanted to measure nut and bolt pre-load, so something similar, but cheaper, should do the job OK. There would be a bit of mechanical stuff involved, but I have a lathe and a couple of milling machines, and machined up the top plate that's on the well at the moment, as I couldn't buy one here. I'll give some thought as to whether I can come up with a new top plate that incorporates load sensing, as that should be a near-universal solution. In my case, the right angled stainless pipe fitting that comes out of the well top plate is connected to a stainless braided flexible hose, that's then connected to the underground water feed pipe. I did it this way so that I didn't put any strain on any of the fixed pipes, but it would also mean that the pipe would have little impact on the measured load variation with water depth.
 

Pongo

Senior Member
You might want to do the math for this first. If your pump is anything like mine it's not exactly "buoyant" and my guess is that the change in measured weight of the whole pump/cable/pipe/top plate assembly for the small change water in level will be extremely small - but I could be wrong.
 

Camerart

Member
Hi J,
Have you tried:

1/A version of a barometer or pressure tube, e,g, a tube to the bottom of the well, open at the bottom, and the pressure measured at the top

2/ A reel of string (Reverse spring loaded) with a floating weight, and count the turns on the reel
I'm now out of ideas.
Cheers, C.
 
Last edited:

Willie...

New Member
Use audible ECHO for well depth

Just a quick comment... I, too, have a well, but it is quite deep! At about 200 feet (71M), when the water level is low, the "acoustic resonant frequency" would be well below audible! (About 2Hz)

With the well cover off, I can clearly hear a "slap echo" from the water below. I have noticed the significant difference between when it is full and when it is empty. My idea would be to use a speaker to make a quick, audible CLICK. (A microphone would detect the echo.) The the CPU would start a timer when the click was sent, and stop when the echo was detected. Obviously, a sensitive audio amp and comparator would be needed to convert the microphone signal into a logic level for the CPU.

The click needs to be in the normal audible range, not ultrasonic, as high frequencies are very clearly attenuated in my well. A clap of the hands above the well, echoes back as a dull thud.

This has been an interesting topic, for sure. :)
 
Last edited:

premelec

Senior Member
Acoustic method seems good - AND the speaker can also serve as microphone listening for back pulse - however not easily done in sealed wellhead installations [like mine] though perhaps there is some way to hear pump start vs electric activation or analysis of running pump sound vs depth with DSP etc... so many clever concepts... a fun thread...
 

Willie...

New Member
I've given some thought to the dual-purpose use of the speaker, but speakers are NOTORIOUSLY poor microphones. Since the click would need to be a fairly high-level signal (a 5v pulse thru a resistor) it would likely overwhelm the mic preamp circuit, and throw-off the timing. I still haven't gotten to the point where I've figured out how to get the signal wires from the house to the well, tho. ;) I definitely don't want to run them thru the conduit with the 240V mains power!
 

Pongo

Senior Member
Since we are kicking around ideas, if it's a metal casing it would make a great waveguide...
 

Steve2381

Senior Member
I found this...

http://www.hydrokit.co.uk/137/level-sense-encoder.htm

Could you not make something similar? Find a rotary encoder and basically count the rotations/pulses as it lowers the end weight.
How it detects contact with the water, I am not sure. I am thinking a very simple RF transmitter in the receiver weight (dirt cheap on Ebay), that transmits a signal upon water contact to the top base unit.

Or get really clever and compare the known weight of the lowered rig per metre (using a strain gauge on the rotary encoder) until it becomes light again when it floats upon water contact (and stops the lowering mechanism).
Bit complicated however... anything with moving parts will probably go screwy at some point.

Yes, possibility of tangling upon operation, but then again... anything going up/down will suffer that. Maybe try and design the end sensor weight as streamlined as possible.

Interesting thread as others have mentioned. The sound/listening idea is great, but I would think very hard to set up and get repeatable results?

I am assuming that a pressure sensor would not work? I have one in my RC submarine to detect the depth and keep the model at periscope depth. Its pretty damn sensitive.

If you dropped the sensor down to a point well below where you know the water never passes (or the bottom), would it not detect the increasing pressure as the water got deeper above it?

Maybe something like this... http://uk.rs-online.com/web/p/pressure-sensors/7396777/.
This sensor is expensive (its RS... they only do expensive).
 

Jeremy Harris

Senior Member
I've given some thought to the dual-purpose use of the speaker, but speakers are NOTORIOUSLY poor microphones. Since the click would need to be a fairly high-level signal (a 5v pulse thru a resistor) it would likely overwhelm the mic preamp circuit, and throw-off the timing. I still haven't gotten to the point where I've figured out how to get the signal wires from the house to the well, tho. ;) I definitely don't want to run them thru the conduit with the 240V mains power!
So far I've done some very crude experiments with a small, Mylar cone, speaker (one from an outdoor speaker). It's 85mm in diameter, and when I send ten pulses at 300Hz down a 3m length of drainpipe I get back a return echo with an amplitude of around 2mV p-p. That's usable, and easily amplified with a narrow band amplifier, then rectified and fed to a comparator.

Protecting the mic preamp turns out to be easy. The speaker has a very low impedance, so can be fed into a resistor and a pair of back to back diodes to clip the transmit signal to a low level. Using the same tuned amplifier and comparator for the transmit pulse and the receive pulse means that any phase shift from the tuned amp will be cancelled out. I should be able to clock the two pulses into into a bistable and then feed the resulting pulse into a Picaxe to measure the pulse width, which should be the time of flight of sound down to the water and back.

Since we are kicking around ideas, if it's a metal casing it would make a great waveguide...
Even thin wall plastic waste pipe seems to make a great waveguide at 300Hz. 300Hz seems fine for the diameter of my liner, which is 119mm internal diameter, 125mm external (so around 4 11/16" bore). My test pipe is only 64mm bore and that seems to work very well, too.

One problem is overcoming reverberation, I've found. A single burst of 300Hz, even at low amplitude, causes the pipe to reverberate such that several returns are received and detected. Only the first one is the true one, so I think the answer is to just make infrequent measurements and allow time for the echoes to die away between measurements.

The next stage is to look at building a suitable box to house the speaker, that can be acoustically coupled to the 1 1/2" threaded port I have in the well top plate. The idea I have at the moment is to use a length of tuned pipe that will fit inside the 1 1/2" threaded port, down into the well. If that pipe is tuned to 300Hz (so around 280mm long) it will act as an effective bandpass filter at 300Hz, reducing the effect of noise.

I can't see any challenges in building a 300Hz bandwidth hi gain amplifier, rectifier, comparator and bistable, which should then give a pulse width proportional to time-of-flight of sound. I may as well add a DS18B20 to measure the temperature in the air in the well at the end of the resonance tube, too, so I can just feed that into the depth calculation (not sure it's really needed - I could just assume the speed of sound is going to be fairly constant at around 336m/S at the normal well temperature). The intention is to just transmit serial data, that is the depth of the water in metres, back to the room where the pressure tanks, filtration gear etc is, so I can either log depth over time or just have a water depth display.
 

techElder

Well-known member
In ultrasonics we used a BLANKING gate at the initial pulse. This can also be used as an adjustable START signal or ZERO.

Also, consider measuring time between the first return and the second return. Divide by two for the travel time.
 

Jeremy Harris

Senior Member
In ultrasonics we used a BLANKING gate at the initial pulse. This can also be used as an adjustable START signal or ZERO.

Also, consider measuring time between the first return and the second return. Divide by two for the travel time.
That's a good idea to look at using the second echo, as it removes all the potential issues from the massive difference between the transmitted signal and the received signal. Instead I could, as you say, just turn the receiver off, in effect, for the transmitted pulse, then just look for the first and second echoes.

I'm impressed by the clean signal I seem to get back, and that's just transmitting 10 pulses, with some very crude low pass filtering before the speaker and no acoustic filter, just the 'scope hooked up to the speaker terminals. I've acquired a couple of small sealed (IP66) ABS boxes and am looking at bonding a bit of PVC sheet inside to mount the speaker and form a sealed chamber either side of it. I reckon I can then use glued PVC pipe fittings to both fit a female thread to suit the 1 1/2" stub I have coming out of the well top plate, and glue a 280mm long tube that will poke into the top of the well and act as a bandpass filter at 300Hz. I can fit the electronics on the "dry" side of the speaker support sheet, with a cable gland to get the power and cable out of the box.

This would make the thing easy to fit and remove, to allow getting at the well top plate if any work needs doing.

I have an electret mic around somewhere, and may try that instead of using the speaker as a mic, but given I can get a reasonably good signal from the speaker I'm not sure it's worth the extra hassle. I'm not sure how well a mic would withstand high humidity all the time, whereas I know the speaker will, as it was one that came out of an outside extension for a ships radio.
 

techElder

Well-known member
Well, the reason a separate mic would be needed is if you damped the speaker down to reduce ringing. You might find it advantageous to only transmit one or two cycles of sound. Consider how much one cycle of sound is in distance to decide about resolution etc.
 

Pongo

Senior Member
Even thin wall plastic waste pipe seems to make a great waveguide at 300Hz. 300Hz seems fine for the diameter of my liner, which is 119mm internal diameter, 125mm external (so around 4 11/16" bore). My test pipe is only 64mm bore and that seems to work very well, too.
I was thinking of 1 GHz and up for my 7.5 inch casing :)
 

Willie...

New Member
I was thinking of 1 GHz and up for my 7.5 inch casing :)
Then, you are calculating/counting NANOSECONDS! (Or relying on delicate/dodgy phase discriminators.)

At least with an acoustic system, it's in much more manageable MILLISECONDS. :)

I also would recommend keeping the audio burst as SHORT as possible. At 300 Hz, 10 cycles would be 33 milliseconds! That's quite a bit, and could easily "muddy" the return echo. (A single cycle of 300Hz is 3.3 milliseconds.) I'm planning to use just a DC pulse ON then OFF. A pulse 1ms wide is approximately 1Khz. That will make detecting the echo pretty easy in a "shallow" well. HOPEFULLY, it will work "well" in my DEEP well. ;)
 
Top