Single ended signal to differential line driver

[Misterbenn]

Hi there, I could use a hand getting my head around a design for a single ended signal to differential input&output amplifier.

I'm using an AC current sensor (LTSR-6NP, **broken link removed**) which gives a single ended voltage output referenced to ground with the zero current point at mid supply rail.

This particular sensor also has a voltage reference input to adjust the zero current midpoint.

I want to make this signal differential so its more noise immune, to do this i found a differential line driver ADA4049-1 (https://www.analog.com/media/en/technical-documentation/data-sheets/ADA4940-1_4940-2.pdf)
This has a fully differential input and output as well as a output common mode voltage set point, Vocm.

This is the kind of thing i'm thinking of at the moment. I'm not sure if i need the Rg and Rf on the -IN input, or if i should just connect the 2.5V mid point reference directly to the negative input?

or this

 

[ronsimpson]

I did not look up your current sensor but; the ones I used before have 1/2 supply=0 current.

You want two outputs. Do you want 0v=0 current or do you want 1/2 supply = 0 current? (is your supply +5V? or +/-5V)

 

[Misterbenn]

Hi Ron,

yes current sensor gives 1/2 supply at zero current.

supply is just 5V, from an LDO.

I want differential outputs, such that:
input = 2.5V: outputs = 2.5V and 2.5V giving 0V differential
input = 5V: outputs = 5V and 0V giving 5V differential
input = 0V: outputs = 0V and 5V giving -5V differential

hope that make sense

 

[Misterbenn]

So I've attached the schematic that I've settled on in the end,

It turns out my single end to differential signal circuit can only have a gain of <=1 so I added a second differential amplifier stage. both amps have an output common mode Vocm of 2.5V from a precision source/sink reference.
The current sensor being 2.5V +/-2.5V@full range.
so when the current is at + full range my sensor output is 5V and my differential signal output is Vdiff = 5V
when the current is a - full range my sensor output is 0V and my differential signal output is Vdiff = -5V

One thing I could use some expert help on is my first stage gain setting. I've accounted for the RC filter impedance of 100Ohms by using an Rg of 1k in the non-inverting leg and Rg of 1k1 in the inverting leg. Is this correct, or should I be ignoring the filter impedance as I want to measure the node voltage at the output of the RC filter?

Thanks

 

[Misterbenn]

small update,

It was pointed out that the RC filter matching I've done is only valid for DC signals not AC.
So i've added a buffer to the RC filter output to give a low impedance input to the first amplifier stage. all the Rg and Rf resistances on that first stage are now 1k to give a balanced gain of 1 on the inverting and non-inverting input over the frequency range.