Ac bridge

[vinke]

What exactly is the Sensor element and what parameter are you measuring?

The sensor is a gypsum block.

"A gypsum block is not like a resistor; it is an electrochemical cell with a saturated (weak) solution of calcium sulphate, also known as gypsum, forming the electrolyte.

If you try and measure it with a normal digital voltmeter, which uses DC excitation, then what you will notice is that the reading drifts continuously due to polarization of the block (movement of ions towards the electrodes). Its true - you can then have any reading you want, provided you wait long enough! But not good measurement practise...

In all workable gypsum block resistance measurements, you need AC excitation to prevent this ionisation occurring (if you keep changing the direction of the electric field, average ionic movement is zero).

The second factor is voltage level - too high a voltage, and you "gas" the gypsum block. This changes the blocks resistance. We use around about 1V peak to peak

The third factor is frequency - if you change the frequency, you change the resistance reading. (I don't know why!). We stick with 1kHz. "
Ive taken the paragraphs in bold from this site:**broken link removed**
These gypsum blocks come with a digital meter which cost $250.
But me im using only the blocks, and i want to interface them with PIC16F877A in a stand-alone system. I already have the curve of the gypsum block impedance against relative soil moisture content,and already design the bridge.Now i need to design a suitable AC voltage source for the bridge.You have had a look at the DC/AC converter using a 555 timer and complementary transistors?The link is in my previous posts above. I was thinking to design an AC source using the same school of thought.

 

[ericgibbs]

hi,
Looked thru your links, they recommend a 1 VppK signal.
The 555 timer driving the push/pull pair into an inductor/cap giving a 'sinewave' would be OK.
I would suggest a small audio transformer would give you sufficient drive for the Sensor ie; 1 Vppk.
Power the 555 from a regulated supply, say +5V, keeping the bridge excitation constant.

You will have to adjust the value of the inductor/cap to suit 1KHZ, rather than 50/60HZ,
to give something 'like' a sinusoidal excitation waveform at 1KHz.

Use a 'ac' coupled OPA, with a bandpass filter for 1KHz, the output into the FWPrec OPA circuit.
Design for 0v thru +5Vdc into the PIC's ADC, for the maximum range of the Sensor bridge.

 

[vinke]

hi,
Looked thru your links, they recommend a 1 VppK signal.
The 555 timer driving the push/pull pair into an inductor/cap giving a 'sinewave' would be OK.
I would suggest a small audio transformer would give you sufficient drive for the Sensor ie; 1 Vppk.
Power the 555 from a regulated supply, say +5V, keeping the bridge excitation constant.

You will have to adjust the value of the inductor/cap to suit 1KHZ, rather than 50/60HZ,
to give something 'like' a sinusoidal excitation waveform at 1KHz.

Use a 'ac' coupled OPA, with a bandpass filter for 1KHz, the output into the FWPrec OPA circuit.
Design for 0v thru +5Vdc into the PIC's ADC, for the maximum range of the Sensor bridge.

hi ericgibbs,
thanks,its what i was thinking to do and im reassured now.
Is the filter really necessary?
I was thinking to connect an instrumentation amp to the output of the ac bridge,then a precision rectifier at the output of the instrumentation amp.

 

[ericgibbs]

hi ericgibbs,
thanks,its what i was thinking to do and im reassured now.
Is the filter really necessary?
I was thinking to connect an instrumentation amp to the output of the ac bridge,then a precision rectifier at the output of the instrumentation amp.

hi,
The reason I am suggesting a filter is to try to reduce 'mains hum' and other em noise, as its an 'ac' amplifier.
You stated 60/100mV signals, so 'ac' gain required about *50 for +5Vdc into ADC,
I would also consider a 5.1V clamp on the input to the ADC.

 

[vinke]

hi,
The reason I am suggesting a filter is to try to reduce 'mains hum' and other em noise, as its an 'ac' amplifier.
You stated 60/100mV signals, so 'ac' gain required about *50 for +5Vdc into ADC,
I would also consider a 5.1V clamp on the input to the ADC.

Thanks ericgibbs, you greatly helped.