I am fairly inexperienced with load cells + instrumentation amplifiers. I'm looking to create the following circuit so that my load cell signal is amplified into a proper range for my ADC using a single supply. This is very similar to the situations described in figure 5 and 15 in the AD621 data sheet:
I have the gain set to 100 based on my load cell output of 2 mV/V, an excitation of 10 V from a single supply, and my ADC reading a range of +/- 2.048 V (ignore the resistance values in the bridge, they are not accurate and just a representation of my load cell). I chose the AD621 based on its low error and gain of 100.
My questions are the following: How should i connect the REF of the AD621? Should it be tethered to the ground of the single 10 V supply, or do i require some sort of voltage divider like what is shown in the AD621 datasheet circuit below? I'm not worried about using low current and thought i could do away with AD705. If needed, i can use a dual supply (+/-5 V), but it would be easier to utilize a single supply (10 V).
I was also hoping to utilize the differential capabilities of my ADC at A0 and A1 to eliminate any noise in the signal from ground loops. I believe this would be similar to the REF and IN pins on the ADC shown above by finding the difference between the two voltages. The ADC is powered via another 5 V "digital" circuit that also powers my Raspberry Pi, and i'm not sure if there will be any resultant noise or how to eliminate the floating voltage across the circuits. Is it necessary to tie the grounds of either circuit? The ADC available to me does not have a AGND pin. There is no connection between the power of the ADC and the analog circuit i've constructed.
Please advise on whether my theory/circuit is correct, any help is appreciated.
What ADC part number are you looking at?
In the AD621 Figure 5 circuit, a 3V voltage, divided down from the Instrumentation Amplifier 5V supply is fed to the ADC REF pin. The AD621 REF pin (pin 5) is driven from a low impedance 2V source which is generated by the AD705. So, the ADC analog input has a nominal / no-signal voltage of 2V at the IN pin. The AGND pin of the ADC is what the ADC IN input is referenced to. So, with AD621 at 2V, and AGND also at 2V, you would register a 0V input signal. The maximum allowed IN voltage will be 3V which is the voltage the ADC REF input is tied to.
I don't think it'd be wise to eliminate the AD705 because you'll cause errors in the AD621 if its REF pin is not driven from a stiff source like the AD705 output.
If the ADC you're using has a differential or pseudo differential input, you could do something similar to AD621 Figure 5. An example of pseudo differential ADC is LTC2367-16 where IN- pin is equivalent to the AGND pin. If your ADC is single ended input, you would have to devise a digital method where 2V at AD621 would "correspond" to a 0 digital code (that's because the AD621 output swings both above and below 2V). There may be other tricks you could play by using +/-5V dual supplies, but I'd have to know the ADC that you've chosen to comment.
There may be other methods of tackling this but this is what I could think of off the top of my head.
Hope this helps.
I'm sorry. I misspoke.
The maximum voltage on the LTC2367-16 IN- input is only +/-100mV. So you could not use that ADC and tie its IN- input to 2V (as shown in AD621 Figure 5). The AD621 Figure 5 requires a true differential input ADC (not a pseudo differential ADC):
"IN– (Pin 5): Analog Ground Sense. IN– has an input range of ±100mV with respect to GND and must be tied to the ground plane or a remote ground sense."
So, it really depends on what ADC you're trying to drive.
Thanks for the reply! I know this deviates from the original post for the AD621, but I did some research and found AD7705, which seems to integrate everything i was trying to build into one chip. By following the exact setup described in fig. 13 and utilizing the 5 V output of my micro controller, i should be able to achieve my goal of reading the load cell, with a reference of 2.5 V:
Is the 5V output of the Rpi reliable enough to use as VDD, or should i utilize a external power supply? And if i use the external supply, should I somehow tie the grounds of the RPi and the Power supply to ensure theres no floating voltages?
Thanks again for your help
I don't have much experience with the Raspberry pi (RPi) that you're working with to tell you whether its 5V VDD is "clean" enough to use on the AD7705 or not?
If I were you, I'd put in the hooks to use a clean / LDO type VDD source for AD7705 if the RPi source proves to be too noisy. That can usually be done fairly easily by using jumpers and 0ohm resistors on your board. I think you can run the AD7705 off of 3V supply if you end up having to use an LDO on the RPi 5V to regulate down to a clean 3V.
And Yes, if you use an LDO on the RPi 5V line, you normally should tie its GND to the RPi ground at one point.