Absolute pressure sensor

[IconicPCB]

I have been asked to look into a pressure sensing issue.

Consider a rigid fluid container exposed to the atmosphere with an absolute pressure sensor connected to the bottom level of tank fluid via a side tube.

As the fluid level changes so the pressure  inside the connecting tube changes and as the the sensor senses absolute pressure it will register level changes.

But as the sensor IS an absolute pressure sensor i will also be affected by changes in atmospheric pressure. In case of water one atmosphere is approximately 10.33 metre high column of water.

Note atmospheric pressure variations are in the order of  1037.4hPa and 986.8 hPa  ( 1057.8cm of water and 1006.2cm of water ).

How will this variation in atmospheric pressure be reflected in the output of the absolute pressure sensor?

[joeqsmith]

It must be late as I am not sure what you are asking.

There's going to be an offset in the sensor's output based on atmospheric pressure but I don't think that's what you are asking. 

[pmbrunelle]

At the bottom of the tank, the absolute pressure is equal to the atmospheric pressure plus the pressure from the height of the liquid.

If you want to measure the height of the liquid, use a differential pressure measurement. Then the atmospheric pressure won't matter.

[JacquesBBB]

Yes,
You need two pressure sensors . One in the bottom of the tank, the other outside for atmospheric pressure.
You get the level of water with the difference

Pbottom - Pout

For a precise measure, you need first to calibrate your sensors so when they are bouth outside the  diff is zero. Just record the diff at zero liquid level.

[mikerj]

Note atmospheric pressure variations are in the order of  1037.4hPa and 986.8 hPa  ( 1057.8cm of water and 1006.2cm of water ).

How will this variation in atmospheric pressure be reflected in the output of the absolute pressure sensor?

Not quite sure what you are asking here, but without correcting for atmospheric pressure you will have an error of up to +-51.6cm in the measured fluid height, depending on the atmospheric pressure when you calibrated the system.  A differential pressure sensor would be the best solution to minimise the error, otherwise a second sensor for atmospheric pressure could be used.

[IconicPCB]

Thank You for confirming the error due to atmospheric variations.

It has been my contention there would be a need for a differential measurement in a discussion with a potential client.

This may call for a rethink.

[dwpatter53]

To measure accurately, level in a closed tank you need a differential pressure sensor.
Connect the atmospheric port on the absolute pressure sesor to the top of the tank.

LP connected to top of tank and HP connected to bottom.
You therefore are only measuring the head from the liquid.

You may have to fill the HP leg with liquid and zero it's head pressure out.

[IconicPCB]

Thanks Davo.

[IanB]

It seems odd to be "reinventing the wheel" like this. If the application is to sense the level in a tank full of liquid, the reliable and logical solution is to obtain a purpose designed level sensor from a qualified instrumentation vendor.

Does the client know all the technical details about how to design and calibrate such a sensor for long term and reliable operation? Instrumentation vendors have the experience and application knowledge to do the job right.

Another thread somewhere mentioned the Dunning-Kruger effect. That rather comes to mind here.

[dwpatter53]

It is re-inventing the wheel; but what ever.

Your standard level transmitters available from any vendor like Rosemount (Emerson) or E&H would be;
-ultrasonic
-microwave
-displacement
-differential pressure
-float
-conductivity

For a closed tank or vessel under pressure dp is the most common.
Microwave works if it can handle the temperature and pressure; versions that will almost handle 900# steam are available.

http://www2.emersonprocess.com/en-us/brands/rosemount/pressure/pressure-transmitters/3051-pressure-transmitters/pages/index.aspx
http://www.ca.endress.com/en

Good luck

[TheDirty]

Getting on the liquid level sensing train here as I'd like some help with this too.  I'd like to measure the level in my water tank in the trailer.  I was planning on using this diff sensor:
http://www.digikey.com/product-search/en?keywords=MP3V5010DP

I know nothing about these pressure sensors and am not an EE.  Just making sure this doesn't have any gotcha's that I'm not seeing.  There will not be more than 1 meter of water in the tank so I'm assuming the 1.45psi working pressure is okay.  The tank is open vented at the top so will not be pressurized.  It should be as easy as plugging it into the drain hose at the bottom and measure the output.  Is there anything I'm not seeing, do I need a higher working pressure?

[retrolefty]

Getting on the liquid level sensing train here as I'd like some help with this too.  I'd like to measure the level in my water tank in the trailer.  I was planning on using this diff sensor:
http://www.digikey.com/product-search/en?keywords=MP3V5010DP

I know nothing about these pressure sensors and am not an EE.  Just making sure this doesn't have any gotcha's that I'm not seeing.  There will not be more than 1 meter of water in the tank so I'm assuming the 1.45psi working pressure is okay.  The tank is open vented at the top so will not be pressurized.  It should be as easy as plugging it into the drain hose at the bottom and measure the output.  Is there anything I'm not seeing, do I need a higher working pressure?

 Many inexpensive pressure sensors are only rated for dry air service, not liquid service. Matching a pressure sensor to a given service/application is most a matter of chemical media compatibility and rated min/max temperature service. Industrial pressure sensors will have a solid metal barrier diaphragm, usually stainless steel, or other material for acid service, etc.

[TheDirty]

Many inexpensive pressure sensors are only rated for dry air service, not liquid service. Matching a pressure sensor to a given service/application is most a matter of chemical media compatibility and rated min/max temperature service. Industrial pressure sensors will have a solid metal barrier diaphragm, usually stainless steel, or other material for acid service, etc.

Thanks.  In this case it is not rated for liquid use.  Had to go deep into the datasheet to even see a mention of it.
Looked for something else and I think the MPX5010DP looks okay.  The datasheet specifies liquid use and has a specific application sheet for water level measurement in washing machines.  I certainly want as cheap as I can get, this is not a precision measurement and I have time to calibrate it myself.

[Brent C]

There are several manufacturers of level transmitters similar to the ones shown in the links below.
These transmitters use a ventilation tube that is ran within the cabling between the device and the atmosphere. This allows for a differential pressure measurement to be achieved.

http://www.wika.ca/il_10_en_co.WIKA

Spec Sheet:
http://www.wika.ca/upload/DS_PEIL10_en_us_16678.pdf

[JimRemington]

For a small water tank, an external tube and a couple of strips of foil make a simple but accurate, DIY capacitive level sensor. It was described in Nuts-n-Volts a long time ago, and I marveled at its simplicity. Reprint here http://engineering.nyu.edu/mechatronics/Control_Lab/ME3484/Readings/NV27-Measuring%20Water%20Level.pdf

[SeanB]

If you want a cheap reasonably easy to use sensor that will do a 1m or so head then try a washing machine sensor. The modern ones are a 2 wire LC circuit that uses a CD4069 as an oscillator, giving a frequency proportional to level, and are mounted with a thin tube leading to the bottom of the tank so the air pressure in the tube is proportional to water depth. Gives a frequency output for feeding into a MCU pin which is then going to be used internally in a counter.

[John Heath]

If you want a cheap reasonably easy to use sensor that will do a 1m or so head then try a washing machine sensor. The modern ones are a 2 wire LC circuit that uses a CD4069 as an oscillator, giving a frequency proportional to level, and are mounted with a thin tube leading to the bottom of the tank so the air pressure in the tube is proportional to water depth. Gives a frequency output for feeding into a MCU pin which is then going to be used internally in a counter.

I remember having a similar problem in a chemical plant trying to computerize storage tanks to prevent overfill. The problem was an inconsistency of the liquid weight so pressure sensors could not be relied on for 50% or 95 % full alarm. It ended up being brute force with 3 pressure senors at 50% , 90% and 100% full by drilling holes in the side of the tank. Not proud of it but hey it worked. The pressure difference caused by liquid is a hundred times greater than a weeny atmosphere difference so it was not given much thought. Point 0f interest is the computer was a 100 buss Z80 with a smoken 1 MHz clock rate and 16 K byte ram , you heard me right "K" as in 16 1000 bytes of memory with a CPM operating system no less. We could conncor the world with this kind of computer power.   

[retrolefty]

If you want a cheap reasonably easy to use sensor that will do a 1m or so head then try a washing machine sensor. The modern ones are a 2 wire LC circuit that uses a CD4069 as an oscillator, giving a frequency proportional to level, and are mounted with a thin tube leading to the bottom of the tank so the air pressure in the tube is proportional to water depth. Gives a frequency output for feeding into a MCU pin which is then going to be used internally in a counter.

I remember having a similar problem in a chemical plant trying to computerize storage tanks to prevent overfill. The problem was an inconsistency of the liquid weight so pressure sensors could not be relied on for 50% or 95 % full alarm. It ended up being brute force with 3 pressure senors at 50% , 90% and 100% full by drilling holes in the side of the tank. Not proud of it but hey it worked. The pressure difference caused by liquid is a hundred times greater than a weeny atmosphere difference so it was not given much thought. Point 0f interest is the computer was a 100 buss Z80 with a smoken 1 MHz clock rate and 16 K byte ram , you heard me right "K" as in 16 1000 bytes of memory with a CPM operating system no less. We could conncor the world with this kind of computer power. 

 Good story. Takes more then simple pressure measurements to gauge liquid high level if the specific gravity of the contents is not known or changing. Most such facilities also required independent height measurement and high alarm level sensing.

"conncor the world"  Is that something you all got from the French?   

[TheDirty]

For a small water tank, an external tube and a couple of strips of foil make a simple but accurate, DIY capacitive level sensor. It was described in Nuts-n-Volts a long time ago, and I marveled at its simplicity. Reprint here http://engineering.nyu.edu/mechatronics/Control_Lab/ME3484/Readings/NV27-Measuring%20Water%20Level.pdf

This is interesting and I'll give this a try.  I can just T off my drain hose, and run a vertical hose up beside the tank.

[dwpatter53]

Getting on the liquid level sensing train here as I'd like some help with this too.  I'd like to measure the level in my water tank in the trailer.  I was planning on using this diff sensor:
http://www.digikey.com/product-search/en?keywords=MP3V5010DP

I know nothing about these pressure sensors and am not an EE.  Just making sure this doesn't have any gotcha's that I'm not seeing.  There will not be more than 1 meter of water in the tank so I'm assuming the 1.45psi working pressure is okay.  The tank is open vented at the top so will not be pressurized.  It should be as easy as plugging it into the drain hose at the bottom and measure the output.  Is there anything I'm not seeing, do I need a higher working pressure?

The datasheet shows a silicone seal on the pressure sensor so even though its mostly for gas service clean water is probably OK.
If you are running at full scale with the tank full, doesn't matter.
You want to know when it's getting empty.
Max pressure is 75 kPa (10.9 psi) well above your working pressure.

Looks good

Good luck

[TheDirty]

The datasheet shows a silicone seal on the pressure sensor so even though its mostly for gas service clean water is probably OK.
If you are running at full scale with the tank full, doesn't matter.
You want to know when it's getting empty.
Max pressure is 75 kPa (10.9 psi) well above your working pressure.

Looks good

Good luck

Thanks for checking this.  With a sealed tube and a u-bend I could make an air pocket and measure that air pressure.  Not certain how much better that is because of the humidity, but as you said, there is a silicon seal.  I have put the MPX5010 on my next Digikey order.

I am going to try out the capacitive measurement method as well.  An unsealed tube with copper tape up the sides.

[meeder]

Why would you use a absolute cell? A relative cell isn't affected by the atmospheric pressure and it is very easy to measure the level with it.
There is however one thing to take into account and that is the density of the fluid. The pressure can be calculated by P=rho*g*h.

You could also mount a guided radar probe in the tube.

Differential pressure isn't necessary since the tank is open to the atmosphere, otherwise you would need a delta-p measurement.

A decent pressure transmitter will have options to enter a linearization table so you can calculate the volume directly and scale that to the output.

In my job as a service engineer I deal with applications like this almost on a daily basis.

What kind of liquid is there in the tank and how big is the tank?

[IconicPCB]

Meeder,

Can you suggest a sensor ?

By the by a sensor which is exposed to the atmosphere will provide the differential pressure reading( between atmosphere and say water column plus atmosphere.

[meeder]

Meeder,

Can you suggest a sensor ?

By the by a sensor which is exposed to the atmosphere will provide the differential pressure reading( between atmosphere and say water column plus atmosphere.

For the exact sensor type you would need more information. I think it is wise to contact a suplier to get the exact type you need.

What you call a differential sensor is what we call a relative pressure sensor.
Basically there are three types of sensor.

Relative, the side opposite to the process connection is open to the atmosphere.

Absolute, the opposite side is closed off.

Differential, this sensor typically has two process connections and is either connected through impuls lines connected directly to the process or capillaries with secondary membranes connected to the process.


https://youtu.be/QrT8VWxdxwk

[Circlotron]

If you want a cheap reasonably easy to use sensor that will do a 1m or so head then try a washing machine sensor. The modern ones are a 2 wire LC circuit that uses a CD4069 as an oscillator, giving a frequency proportional to level, and are mounted with a thin tube leading to the bottom of the tank so the air pressure in the tube is proportional to water depth. Gives a frequency output for feeding into a MCU pin which is then going to be used internally in a counter.

And don't remove and replace the tube to the sensor while there is water partway up the tank because this partway will then become the new zero. Result = Mrs Circlotron not happy (again) and bans all further experiments on washing machine.