Weird data from Analog Input

Hi all,
I have a problem with Analog Input on my Arduino Uno. I have A0 port to AD620N in-amp (powered from Arduino - +5V from USB), which amplifies signal from MPX2010DP sensor (powered from external 12V power supply). Readings I get from A0 port are far from I expect - they tend to oscillate from around 400mV to 4000mV (with the same frequency as sine signal I get when A0 is not connected to anything), when, standard, simple voltometer shows "steady" 1700+-10mV - and I don't know why.

I've tried connecting AD* to the same power supply as MPX* (power supply has +5V and +12V rails with two ground lines), disconnecting external supply, changing AD* - but no success.

I would be very thankful for any help/ideas why this may be happening.

Thanks in advance.

What gain resistor are you using on the IA?

Do you have an oscilloscope handy?

I am using 47 Ohm resistor (for very high gain, since changes in pressure I'm going to detect are very small) - but changing it does not change the reading form A0 - it still oscillates.

Unfortunately, I do not have an oscilloscope, only very simple voltometer.

If you have a second Arduino you could make a (simple) scope - Xoscillo, Cheap Arduino Based Oscilloscope | Arduino Blog -

I do not have second Arduino either. I have tried connecting IA to other analog inputs, but with the same results - oscillating input (with nothing connected, it is simple sine wave).

If you short -IN and +IN on the AD620 do you get 0 volts out?

Are you using long wires between the pressure sensor and the AD620? They might act as antenna and shielded twisted-pair cable might be a better choice.

Are you sure the 12V to the sensor is well filtered?

Can you photograph or sketch your whole circuit? We're guessing otherwise.

The AD620 will struggle work from a single 5V supply - its inputs both need to be very close to +2.5V (about 2V to 3.5V) or it'll saturate.

I'll post the sketch as soon as I get home. What baffles me is that readout from voltometer works (or looks like it is) OK- it changes from 1700mV to 1600mV (or 1900mV) as I change pressure on sensor inputs.

I can try and connect AD620 to 12V power, but I don't know if Arduino AIn can handle its output.

Little selfbump, here is my board, +5V comes from Arduino and AD620 is grounded by Arduino.

I have shorted IN+ an IN- and I do not get 0 volts, I get around 1.6V. Cables are quite short (max 5cm).

There's your problem! You are hooking one side os a differential output to both sides of the amp!

Shouldn't you connect Vout+ (Pin 2) to +IN and Vout- (Pin 4) to -IN?

OK, i forgot so specify that MPX image is wrong here - in database I found it has only one output. I have VOut+ connected to VIn+ and VOut- to VIn-. My bad ^^'' But, nevertheless, when I shor the AD620 i do not get 0 volts...

You don't seem to be providing a reference voltage on the REF pin of the IA. Have you tried grounding that pin?

I have grounded that pin, output still oscillates but with different pattern (i get "peaks" from time to time), and when shorted i still get 0.5V volts on IA. Voltometer reports changing pressure better, and "peaks" change in amplitude when I change pressure)

Is there something plugged into the AOUT jack? If not it appears to be wired to short DOUT to Ground.

This is the datasheet at Analog Devices.
http://www.analog.com/static/imported-files/data_sheets/AD620.pdf

Note carefully in the specifications the output swing (page 4) and the input voltage range (page 3). Neither goes to the ground rail.

edit: I think that IC was designed for a dual power supply. Maybe there is a better op-amp for this?

AOUT is completely disconnected at the moment, so this should not be a problem. I have to return the project tomorrow, so I'll just leave it with analog part for now, with IA connected to +12V rail. I'll also look around for another IA - can anyone point me to something that works well with +5V power supply? Or maybe a better A-D converter?