Author Topic: Measuring water in a water tank  (Read 19528 times)

0 Members and 1 Guest are viewing this topic.

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #25 on: January 06, 2013, 11:01:15 am »
One traditional method, much used in the industry, is to measure the fluid column pressure using a simple air pipe. Just run a hose or tube to the bottom of the tank - open from the bottom, sealed from the top. Attach one of these http://www.digikey.com/product-search/en/sensors-transducers/pressure/1966259?k=pressure%20sensor to the sealed top and process the result as you need.
No moving parts, measures anything from pure water to septic waste, virtually foolproof. The only precaution would be to have some kind of air valve, maybe a car tire type, for flushing the pipe in case of a leak.

I had considered this as well, but the major drawback is during the day, when the sun heats the air in the pipe, the air expands and exerts pressure on the sensor causing inaccurate readings.

don't put the pipe in the sun?

Much appreciated - replace the word "sun" with QLD summertime ambient temperature.
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #26 on: January 06, 2013, 11:04:27 am »
OK. the pipe is open from the bottom so any expanding air will escape and the reading won't change because of that. But the situation changes when the system cools down and the air contracts. Then you do get a measuremnent error because the pipe is not completely air filled any more. What they do in this case is to set up a bubbler that slowly pushes air into the pipe, letting it escape from the bottom. That way the system is not dependent on temp variations. Could still be the cheapest solution provided you can find a suitable pump... If the accuracy is sufficient, i would say this merits serious consideration due to its simplicity. There must be a reason why it is/was so common in the industry.

The bubbler method is another industry standard measurement technique but that I have researched, but I did not want the complexity and failure point of a pump in the mix.
 

Offline Kremmen

  • Super Contributor
  • ***
  • Posts: 1283
  • Country: fi
Re: Measuring water in a water tank
« Reply #27 on: January 06, 2013, 11:07:30 am »
OK, just a suggestion. Hope you find a solution that works for you.
Nothing sings like a kilovolt.
Dr W. Bishop
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #28 on: January 06, 2013, 11:09:04 am »
OK, just a suggestion. Hope you find a solution that works for you.
No probs - I appreciate the input ! I have looked at doing this for about 2 years and cannot come up with a reliable inexpensive method. The best would be a submersible temp compensated differential pressure transducer but these run at about $500 AUS.
« Last Edit: January 06, 2013, 11:12:44 am by ilium007 »
 

Offline ftransform

  • Frequent Contributor
  • **
  • Posts: 729
  • Country: 00
Re: Measuring water in a water tank
« Reply #29 on: January 06, 2013, 11:14:11 am »
I think a magnetic float in a PVC pipe with a bunch of hall effect sensors could be really cheap and reliable.
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #30 on: January 06, 2013, 11:15:38 am »
I think a magnetic float in a PVC pipe with a bunch of hall effect sensors could be really cheap and reliable.

That would be awesome and thats what brought me to the site to post - but I dont know how these would be strung together nor how to read multiple sensors to get an intermediate position reading.
 

Offline ftransform

  • Frequent Contributor
  • **
  • Posts: 729
  • Country: 00
Re: Measuring water in a water tank
« Reply #31 on: January 06, 2013, 11:20:29 am »
I think a magnetic float in a PVC pipe with a bunch of hall effect sensors could be really cheap and reliable.

That would be awesome and thats what brought me to the site to post - but I dont know how these would be strung together nor how to read multiple sensors to get an intermediate position reading.

me either lol I have hall effect sensors coming from hong kong for a similar application!  :)]

How are you going to fit this in your tank? Is this like a water tower tank? It sounds big.
Are you a farmer?
« Last Edit: January 06, 2013, 11:22:33 am by ftransform »
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #32 on: January 06, 2013, 11:22:59 am »
I think a magnetic float in a PVC pipe with a bunch of hall effect sensors could be really cheap and reliable.

That would be awesome and thats what brought me to the site to post - but I dont know how these would be strung together nor how to read multiple sensors to get an intermediate position reading.

me either lol I have hall effect sensors coming from hong kong for a similar application!  :)]

Well I would like to keep an eye on your project !

How are you going to fit this in your tank? Is this like a water tower tank? It sounds big.
Are you a farmer?
 

Offline mamalala

  • Supporter
  • ****
  • Posts: 777
  • Country: de
Re: Measuring water in a water tank
« Reply #33 on: January 06, 2013, 11:37:48 am »
Two isolated metal strips or rods, parallel aligned and at a fixed distance, form a capacitor. Water between them will change the capacitance. Use a fixed frequency oscillator and a MC1496. Split up the osc output, one going directly into the 1496, the other through a L/C network for phase shifting, then into the other input of the 1496. The C in that network includes the sensor. At the 1496's output you then get a voltage that corresponds to the amount the phase is shifted. Read that with a decent ADC.

No movable parts, continuous output.

Greetings,

Chris

 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #34 on: January 06, 2013, 11:40:35 am »
Two isolated metal strips or rods, parallel aligned and at a fixed distance, form a capacitor. Water between them will change the capacitance. Use a fixed frequency oscillator and a MC1496. Split up the osc output, one going directly into the 1496, the other through a L/C network for phase shifting, then into the other input of the 1496. The C in that network includes the sensor. At the 1496's output you then get a voltage that corresponds to the amount the phase is shifted. Read that with a decent ADC.

No movable parts, continuous output.

Greetings,

Chris

I did see a design for a capacitance based gauge based on an insulated wire and the result of his testign that it was inaccurate at low levels in the tank. Do you know of any sampel circuits I could study as I do like the idea of a capacitance based gauge - but again, I know nothing about oscillators  :-//
 

Offline johnwa

  • Regular Contributor
  • *
  • Posts: 237
  • Country: au
    • loopgain.net - a few of my projects
Re: Measuring water in a water tank
« Reply #35 on: January 06, 2013, 11:57:56 am »
I have been working on a similar setup for a while now. My first attempt was with a capacitive sensor, using the circuit from here: http://njhurst.com/electronics/watersensor/. This however had some serious drift issues, and I also found that it was susceptible to radio interference - possibly because I did not have the circuit close enough to the sensor.

I therefore decided to use a pressure based approach. This was the same method used in a Silicon Chip project a few years ago. They used this sensor: http://au.element14.com/jsp/search/productdetail.jsp?sku=1608910, which I also elected to use. I see Farnell also have some cheaper sensors, which may be suitable. The Silicon Chip article suggested two sensing methods: using an air-filled dip tube (which will have issues with temperature compensation), or submerging the entire sensor assembly in the tank. The second method will have greater accuracy, but requires extra work in waterproofing the sensor.

Instead of these methods, I decided to mount the sensor externally, connected through a sensing port at the bottom of the tank. This gives a proper pressure reading without needing to waterproof the sensor. Note that the outlet used should be dedicated to the sensor, as any inflow or outflow through the port used will upset the reading.

While the datasheet for the sensor says that it is only suitable for sensing air, the sensor die is potted in silicone, so I can't see how clean water would damage it. I left a small amount of air in the pipe to isolate the sensor from the water, though this may dissolve eventually. This small volume of air should not affect accuracy with temperature variations, provided that there is a reasonable length of pipe running horizontally out of the sensor.

My installation has only been running for a couple of weeks though, so I cannot comment on the long-term reliability of this arrangement.

One other good level sensing solution is a magnetostrictive linear position sensor, with a toroidal floating magnet around the sensing rod. Unfortunately, these are quite expensive.
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #36 on: January 06, 2013, 12:02:01 pm »
That is the exact capacitance sensor I was referring to above. I also pawed over the silicon chip design as well as the freetronics design. I still dont like the idea of water coming in contact with the sensor and will be interested to see how yours turns out. Is your build on a blog ?? I am certain I have seen your username elsewhere - I commented on a blog about a year ago where a guy was building exactly what you describe.

This is the site with the silicon chip version but a little modified:
http://kayno.net/2010/02/11/wireless-picaxe-based-water-tank-level-sensor/
« Last Edit: January 06, 2013, 12:04:05 pm by ilium007 »
 

Offline mamalala

  • Supporter
  • ****
  • Posts: 777
  • Country: de
Re: Measuring water in a water tank
« Reply #37 on: January 06, 2013, 12:12:16 pm »
I did see a design for a capacitance based gauge based on an insulated wire and the result of his testign that it was inaccurate at low levels in the tank. Do you know of any sampel circuits I could study as I do like the idea of a capacitance based gauge - but again, I know nothing about oscillators  :-//

Attached are three circuit snippets that should give you the idea. I can't really give out the complete thing, or too many details about this specific one, since it is part of a project i did for a customer.

In sens_1 you see the sensor circuitry. At the top right a sine-wave is fed in. It goes through R6/C15 and leaves there as CARRIER signal. But it also goes through the L/C circuitry into the OpAmp IC2 to generate the SIGNAL output. SENS_T and SENS_B are the two metal strips that form a capacitor. Feeding a voltage into TUNE allows to offset/calibrate the output signal. The adjustable capacitor in the middle-right of the circuit is to adjust it's center frequency. The sensing element is, as said, just two metal strips in parallel, with a gap between them.

In sens_2 you can see the sinewave generation. The osc is just a regular oscillator, you feed it +5V and it generates a square-wave output of the fiven frequency. This gets filtered in this snippet through the L/C, to form a sine-wave, and amplified by the OpAmp.

the part sens_3 is the actual signal measurement. the two signals, carrier and signal, from the osc and L/C circuitry enter the MC1496. That chip in turn generates a proportional output at OUT- and OUT+. Feed that into another OpAmp circuit to lowpass filter it and to get an unbalanced output signal that you can feed into a ADC.

Of course the values in the given snippets, especially the filter and L/C part, are pretty much useless for you, since your sensor element would be much longer. But the general idea is the same. All it does to modifiy the signal through the L/C circuit, which is detuned by the variable sensor capacity. The MC then detect that difference and generates an output voltage.

Greetings,

Chris
 

Offline mamalala

  • Supporter
  • ****
  • Posts: 777
  • Country: de
Re: Measuring water in a water tank
« Reply #38 on: January 06, 2013, 12:16:56 pm »
Forgot to add:

You can put the two metal strips inside a metal tube (that tube of course open at both ends) and connect it to GND, to give a better shielding against external RF/noise. Just make sure that you have a rather big diameter tube. The sensor strips should have more distance to the tube wall than to each other.

Greetings,

Chris

EDIT: And the wavelength of the operating frequency should be much longer than the length of the sensor strips.
« Last Edit: January 06, 2013, 12:18:38 pm by mamalala »
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #39 on: January 06, 2013, 12:23:01 pm »
Thanks for the extensive reply. I will start studying those circuits but I think it is above my level of knowledge. I have never done any signal generation / interpretation so this is a whole new league for me.
 

Offline mamalala

  • Supporter
  • ****
  • Posts: 777
  • Country: de
Re: Measuring water in a water tank
« Reply #40 on: January 06, 2013, 12:26:42 pm »
Thanks for the extensive reply. I will start studying those circuits but I think it is above my level of knowledge. I have never done any signal generation / interpretation so this is a whole new league for me.

Well, knowledge can be extended ;)

That circuit is very stable and gives repeatable readings/results. It is originally used in test equipment to test material samples, that requires a rather high accuracy and sensitivity.

Greetings,

Chris
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #41 on: January 06, 2013, 12:30:00 pm »
Thanks for the extensive reply. I will start studying those circuits but I think it is above my level of knowledge. I have never done any signal generation / interpretation so this is a whole new league for me.

Well, knowledge can be extended ;)

That circuit is very stable and gives repeatable readings/results. It is originally used in test equipment to test material samples, that requires a rather high accuracy and sensitivity.

Greetings,

Chris

Much appreciated.
 

Offline johnwa

  • Regular Contributor
  • *
  • Posts: 237
  • Country: au
    • loopgain.net - a few of my projects
Re: Measuring water in a water tank
« Reply #42 on: January 06, 2013, 12:37:05 pm »
That is the exact capacitance sensor I was referring to above. I also pawed over the silicon chip design as well as the freetronics design. I still dont like the idea of water coming in contact with the sensor and will be interested to see how yours turns out. Is your build on a blog ?? I am certain I have seen your username elsewhere - I commented on a blog about a year ago where a guy was building exactly what you describe.

This is the site with the silicon chip version but a little modified:
http://kayno.net/2010/02/11/wireless-picaxe-based-water-tank-level-sensor/

Hi ilium007, I don't have anything on the web at the moment, though I am working on getting a site together. I hadn't seen that guy's implementation, though it is basically exactly the same as what I have done. Note that the sensor is teed off from the tank outlet in the photo, this may cause fluctuations in the reading when water is being drawn off, though I don't know how bad this would be. For my installation, I made up a barbed fitting with an O-ring and an M6 thread, and drilled and tapped a hole for it in the wall of the plastic tank. I have another concrete tank that I want to monitor, I am planning to put a tee in the outlet, but feed some small (~3-5mm) metal pipe through the tee, through the valve, and about 500mm into the tank, to avoid the high velocity flow at the outlet.

In theory the sensor shouldn't get wet, but the air pocket that protects it will probably dissolve in the water eventually. I will have to post an update once it has been running for a while.

Temperature stability looks good so far, there have been some quite hot days the past week, and the reading is within a couple of percent of where it was originally. (see graph)

 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #43 on: January 06, 2013, 12:46:23 pm »
That is the exact capacitance sensor I was referring to above. I also pawed over the silicon chip design as well as the freetronics design. I still dont like the idea of water coming in contact with the sensor and will be interested to see how yours turns out. Is your build on a blog ?? I am certain I have seen your username elsewhere - I commented on a blog about a year ago where a guy was building exactly what you describe.

This is the site with the silicon chip version but a little modified:
http://kayno.net/2010/02/11/wireless-picaxe-based-water-tank-level-sensor/

Hi ilium007, I don't have anything on the web at the moment, though I am working on getting a site together. I hadn't seen that guy's implementation, though it is basically exactly the same as what I have done. Note that the sensor is teed off from the tank outlet in the photo, this may cause fluctuations in the reading when water is being drawn off, though I don't know how bad this would be. For my installation, I made up a barbed fitting with an O-ring and an M6 thread, and drilled and tapped a hole for it in the wall of the plastic tank. I have another concrete tank that I want to monitor, I am planning to put a tee in the outlet, but feed some small (~3-5mm) metal pipe through the tee, through the valve, and about 500mm into the tank, to avoid the high velocity flow at the outlet.

In theory the sensor shouldn't get wet, but the air pocket that protects it will probably dissolve in the water eventually. I will have to post an update once it has been running for a while.

Temperature stability looks good so far, there have been some quite hot days the past week, and the reading is within a couple of percent of where it was originally. (see graph)

Ok - that looks promising ! Looks like you are in QLD as well ! I am still keen to look at this floating magnetic / hall effect array though. Given your efforts do you agree that information on doing this in the net as very hard to find ??!!

**edit given your username and time difference I assume you are in WA !
« Last Edit: January 06, 2013, 12:50:16 pm by ilium007 »
 

Offline phatcenter77

  • Contributor
  • Posts: 6
  • Country: us
Re: Measuring water in a water tank
« Reply #44 on: January 07, 2013, 03:50:04 am »
First, place a conductor at the lowest point in the tank, anything really would work, but if it's corrosion resistant even better.  Then drop a "strip" of conductors spaced at a distance that will provide you the resolution required.  I imagine this as something as simple as a bundle of wires with each conductor exposed at the distance noted above, see ascii drawing below:

l l     l l     l l     l l
l l     l l     l l     l l
l l     l l     l l      l
l l     l l      l
l l      l
 l

You can then measure resistance from the conductor at the bottom to each of these exposed conductors hanging in the tank and determine where the water level is.  This will not drift over temperature, and can even be fairly immune to environmental concerns if you scan the points and look for a sharp change in whatever you measure.  This could be very cheap and use a few muxes or something else clever to allow enough measurement heights to provide plenty water volume resolution.

I hope this makes sense and helps.
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #45 on: January 07, 2013, 06:40:46 am »
Sorry but that won't work long term - amongst a few reasons the fact that electrolysis exists will preclude this design. But thanks for the reply.
 

Offline phatcenter77

  • Contributor
  • Posts: 6
  • Country: us
Re: Measuring water in a water tank
« Reply #46 on: January 07, 2013, 06:59:08 am »
I disagree that electrolysis will be a large factor.  With low voltage, low duty cycle (measurement to take much less than a second repeated every hour, and proper material choice it becomes a non-issue over any projected lifetime.  What other shortcomings do you see, because I believe this to be one of the only suggestions that meets all original criteria laid out and does not require a mechanical system.  With regards to KISS I don't think it gets any better.  I understand if it won't work for your specific use, but I'd like to know for my own knowledge.

I do, however, really like the one linked to above much more...

Sorry but that won't work long term - amongst a few reasons the fact that electrolysis exists will preclude this design. But thanks for the reply.
« Last Edit: January 07, 2013, 07:09:15 am by phatcenter77 »
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #47 on: January 07, 2013, 07:29:36 am »
I disagree that electrolysis will be a large factor.  With low voltage, low duty cycle (measurement to take much less than a second repeated every hour, and proper material choice it becomes a non-issue over any projected lifetime.  What other shortcomings do you see, because I believe this to be one of the only suggestions that meets all original criteria laid out and does not require a mechanical system.  With regards to KISS I don't think it gets any better.  I understand if it won't work for your specific use, but I'd like to know for my own knowledge.

I do, however, really like the one linked to above much more...

Sorry but that won't work long term - amongst a few reasons the fact that electrolysis exists will preclude this design. But thanks for the reply.

My requirement is as much academic learning as it is getting a result. The fact that electrolysis is an issue precludes it form my options. On your point above - I believe that if you use AC the electrolysis issue goes away.
 

Offline ilium007

  • Regular Contributor
  • *
  • Posts: 210
  • Country: au
Re: Measuring water in a water tank
« Reply #48 on: January 07, 2013, 07:43:04 am »
So back to the ultrasonic method....

This is the Maxbotix ultrasonic sensor I am going to trial. It is water resistant and is design especially to filter out small objects in the path of the beam:





The idea is as follows:

150mm PVC pipe on outside of tank connected at the bottom - this will reflect the true level of water but wont have all the turbulence of the water inside the tanks during a bit downpour. I want to test if the beam width at the bottom will be obstructed by the diameter of the PVC pipe. Maxbotix cant tell me so it is up to me to test !



I would be happy for any criticism or otherwise !
 

Offline Kremmen

  • Super Contributor
  • ***
  • Posts: 1283
  • Country: fi
Re: Measuring water in a water tank
« Reply #49 on: January 07, 2013, 09:35:15 am »
Actually not a bad idea at all. Regarding sound propagation, only one potential issue comes to mind: multiple signal paths caused by reflections from the pipe sidewalls. That can cause the echo to spread in time so that a naive detection may give inaccurate readings. Having said that, it however should be the case that the direct reflection from the water surface should be the shortest path and arrive first. Surely it can be made to work with a bit of labbing.

BTW the sensor looks interesting. Does DK or Mouser source them? I might be interested myself in that case.
Nothing sings like a kilovolt.
Dr W. Bishop
 


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf