Instrumentation amplifier configuration

[nabla101]

Hi, I want to use an instrumentation amplifier as the first stage of an ultrasonic reciever circuit.

I noticed there are two ways instrumentation amplifiers are used:
-With the REF pin tied to ground, and one pin of the sensor tied to ground
-With the sensor connected across the instrumentation amplifier inputs, and the REF pin supplied with bias voltage (e.g. half the supply).

I was wondering what is the difference between the two, and when would one be used over the other?

Thanks

 

[FvM]

Firstly, if you don't need a differential signal amplification, why using an instrumentation amplifier?

Secondly, important points to consider are:
- single or dual supply voltage
- input bias method
- intended output bias level
Knowing the answer to this questions, you'll easily find out the right connection.

 

[nabla101]

Firstly, if you don't need a differential signal amplification, why using an instrumentation amplifier?

Secondly, important points to consider are:
- single or dual supply voltage
- input bias method
- intended output bias level
Knowing the answer to this questions, you'll easily find out the right connection.

Hi, I want to detect a pulse on the sensor, so I would pass the output of the amp through a filter, an envelope detector, and then to a comparator, to give me a digital pulse when an ultrasonic pulse is received in the passband of the filter.

I have been told that if I use a fully integrated instrumentation amplifier, and tie an input to ground (Making the transducer signal single-ended), then I can acheive much better performance in terms of SNR than with using discreet op-amps and/or transistor amplifiers.

Also, an instrumentation amplifier is gain-programmable with a single resistor, and I want to be able to easily vary the value of the gain using this resitance.

I guess in answer to my oiwn question, I am just trying to measure the voltage across the transducer, so if one transducer terminal is grounded, I have a single-ended output, and so I could use gnd for the IA REF pin, and the other IA input.

edit: I'm just unsure about whether I could do this, or whether I would need a bias voltage on the IA?

 

[FvM]

With single supply (you didn't actually tell, I just presume), inputs as well as reference need to be biased to mid supply or similar.

Assuming a capacitive sensor (you also didn't tell) input bias resistors are required.

Also, an instrumentation amplifier is gain-programmable with a single resistor, and I want to be able to easily vary the value of the gain using this resitance.

Single resistor gain programming is also implemented with an inverting or non-inverting OP configuration.

I have been told that if I use a fully integrated instrumentation amplifier, and tie an input to ground (Making the transducer signal single-ended), then I can acheive much better performance in terms of SNR than with using discreet op-amps and/or transistor amplifiers.

The sensor connection will hardly affect amplifier SNR. Technically, a basic transistor amplifier promises the best SNR ratio, also over OPs, except for very low frequencies where thermal fluctuations are a dominant noise source. It should be added, that although the differential input stage adds 3 dB of noise, the drawback is often compensated by using low noise transistors and optimal bias circuits.