weird differential pressure sensor behaviour

[cheerio]

Hi guys.
I got one of those(MPXV5050DP):
http://www.farnell.com/datasheets/673750.pdf

i want to use it to measure the level of water in a watertank. to do this i use a smallpipe that stands inside the tank and is attached by a silicone hose to this sensor. Because it is differential i can ignore the changes in pressure of the environment.

here is the weird stuff. after weeks of debugging i found out that pressure escapes THROUGH the sensor. Maybe it could not handle the humidity or smth like that and got corroded or whatever. But this behaviour should never ever happen right?

i think i will use 2 absolute pressure sensors. one to measure the pipe and one to measure the surrounding pressure.

[Sir N]

But this behaviour should never ever happen right?

Right!

Did you notice this part:

The MPX5050/MPXV5050G series pressure sensor operating characteristics, and internal reliability and
qualification tests are based on use of dry air as the pressure media. Media, other than dry air, may have adverse effects on
sensor performance and long-term reliability. Contact the factory for information regarding media compatibility in your
application.

(My emphasis)

Perhaps the bonding of the membrane has got a very small leak?
Did you try shooting Motorola an email?
Fresh or saline water?

i think i will use 2 absolute pressure sensors. one to measure the pipe and one to measure the surrounding pressure.

If you're not getting different makes, you may get into the same trouble, or perhaps you were just unlucky with the present one - adding something that's not needed, just means that there's more to fail.

Not knowing your purpose, did you consider other methods for measuring the water level?

[cheerio]

It is fresh water. The water has never direct contact to the sensor. However, the humidity could be quite high.

the first method i tried was detecting the waterlevel by capacitance changes.
http://pcb.tennoio.de/index.php?topic=1252.msg6212#msg6212

that was ok, but it was not a linear function. The really bad thing was the change of the function when the water composition changed. so i ditched it. To be able to measure the waterlevel w/o having to worry about the composition i had the idea to measure the pressure. 1m waterlevel = 0.1bar
this should work very well but the stupid sensor screws me. I put a nail in the silicone hose on the differential side of the sensor. now i should have different readings when the atmospheric pressure changes, but it is a first work around to try...

I gotta test the current setup for a while and then i will think about what happens next.

[kam]

Some many years ago when i used to work in ship repairs, we had differential sensors for extreme pressures and temperatures inside the turbine boilers. They were used to detect the water level under many PSI and temperatures above 400oC. So you can definitely find one to fit your needs
Such systems are widely used.

[George]

there is a whole thread on detecting water level in a storage tank over at EEVBLOG, I'll link it here to save you searching

http://www.eevblog.com/forum/projects/water-tank-level-measurement-what-sensor-to-use/

[cheerio]

thx george. unfortunately the ultrasonic won't work very well in my application and i want a seemless reading(not those hanging electrodes). but i found a differential pressure sensor made by honeywell that should do the job very good. it is my favorite sensor manufactor anyway ^^

http://sensing.honeywell.com/product%20page?pr_id=5326 check it out.

kam do those systems have a special name? and are they cheap?

[kam]

We used pressures for steam turbines, 120psi 550^oC, i do not know much about these small systems. We call them "pressure transmitters"

Check this out:
http://uk.farnell.com/freescale-semiconductor/mpxm2010gs/ic-pressure-sensor/dp/1555611

It has a nice application note:
www.freescale.com/files/sensors/doc/app_note/AN1516.pdf

More sensors here:
http://uk.farnell.com/sensors_pressure

[cheerio]

ah that is very nice and they are pretty cheap too 
those tanks were sealed i guess? in my application the tank is not sealed, so a change in athmospheric pressure will change the level reading. that is why i need the sensor to be differential

i will check out the app note. it looks really interesting. thx big time






edit: did not calculate kpa to bar right -_-

[Sir N]

The water has never direct contact to the sensor. However, the humidity could be quite high.

If you can get hold of the type of RTV silicone sealant that is almost self leveling (selbstnivellierenden in German), you could use that for a conformal coating of the membrane on the "water side" of the sensor. That would seal it AND make it humidity resistant.
Just make sure to get the RTV with a base (acid/alcohol/whatever) that won't harm the existing membrane.

Another way of sealing it could be a thin coating of Araldite (or similar epoxy glue) thinned with a bit of white spirit (to make it semi-fluid). The leak is likely to be found at or near the rim.

The above assume that the tube you connect it to is sealing properly of course, as this is perhaps a more likely place for a tiny leak (RTV of any kind would help, if that's where the problem is).

[kam]

those tanks were sealed i guess?

we're talking about 120psi!!! Of course sealed. Open the boiler when under pressure and the ship turns into airplane, and then submarine. The differential sensor is used like this: One side measures the steam pressure, the other side measures the water pressure. The sensor then subtracts mechanically the steam pressure from the water pressure. This way the water level is measured.

[cheerio]

ok
that's basicly the same way i want it to work. just with less pressure