Load Cell Application - An ADC decision

[carl]

What about the AMP04 - this allows true 0V for a single 5V Supply.
Would this be better?

http://www.analog.com/static/imported-files/data_sheets/AMP04.pdf
Page 6

Carl

[carl]

Hmm,

I am trying this out so far on a bread board. I Knew that this was probably not the best idea - due to the extra capacitance and noise that it presents.

I also had an ADC connected (AD7776) and a 4MHz oscillator. I was using the 2V reference from the ADC to bias the IA to 2V. The gain gave out a 4.096V span.

I decided to take everything off the bread board except the load cell and IA. I changed the gain to give out a 2V span.

The lineararity has improved immensly. It must have been interference on the bread board - especially the Oscillator.

I understand that I have not set up any offsets - and that if the ref is at 0V then I cannot get the signal into that scale.

But setting the ref to 2V via a potentail divider seems to have solved the problem.

So now at 0Kg - I get 2V
at 5Kg - I get 2.97V
at 10Kg - I get 4V

So the process is working. So the inaccurate results were because of the set-up (bread board with other components on) interferring with the Load cell and IA.

I already had decoupling capacitors on the board.

My question is now - how can I avoid this interferrence when I continue with connecting up the ADC (which has yet too be finalised) and associated parts.


What I want (ideally) is to be able to see the Load sensor, IA and ADC all working correctly first on bread board - without it being connected to the PIC. That is why I am considering using a Parallel output ADC - so that I will be able to visually see the output (10 bit word connected to LED's).

Any help would be appreciated.

Carl

[mkuang]

You are trying to measure millivolt level signals on a breadboard, with no ground planes, a mix of digital and analog signals running all over the place. You expect accurate results? You should make a prototype board with a ground and power plane at the very least. On your prototype board try to keep the digital and analog separate, use surface mount parts as much as possible to avoid lead inductance and stray capacitances, and retest your circuit.

[carl]

Point taken.

Thanks
Carl

[gammi,grt]

Hi, I am working with a load cell. I know the analog voltage of load cell can be converted into digital value using any adc. But using load cell, I am trying to calculate the weight of anything. So I want to know how could we denote the digital value in terms of weight (kg) ?
And I am using a positive voltage regulator and negative voltage regulator because I am providing a +12V and -12V to the circuit and one amplifier ic. Please help me. Thanks.

[johanpret]

I Used the AD7794. It is a very nice chip to connect load cells to a micro. Use SPI.

http://www.analog.com/en/analog-to-digital-converters/ad-converters/ad7794/products/product.html
_________________
Johan Pretorius

[sseidman]

Keep in mind that the input impedance on the PIC ADC's can be fairly low, about 10K IIRC. This might be a problem, depending on the circuitry of your bridge.

[Gabriel]

Hi,

I did a similar project a few years back, it was for a GE Druck Pressure sensor.

it had diferential outputs aswell and was bridge type sensor .

I used a single LM324 OP AMP IC to do all the buffering, amplifiying, ground referencing etc... and then read everything with the PIC ADC.

worked perfectly... 1cm water level resolution...


i dont know how much swing your get from your your outputs but _MAYBE_ the code i posted for the LM35 on the library can work for you, since it works with a "diferential" input ...and its 10 bit...
you will probably have to modify the ADC constant from 0x31 to 0x1E8, to get more "resolution", and place a LM358 as line buffers for impedance reasons.

worth a try..

G.
_________________
CCS PCM 5.078 & CCS PCH 5.093