# Are there fully differential feedback amplifiers with series inputs?

There are several examples of the 4 feedback topologies in single-ended circuits.

In the case of series inputs (i.e. series-shunt and series-series topologies), the input is connected directly to an input terminal of the forward amplifier (e.g. non-inverting input of an op-amp or base of a transistor). Examples of single-ended series-shunt and series-series circuits are shown below (snipped from Design with Operational Amplifiers by Sergio Franco):

Are there fully differential feedback circuits with series inputs?

An example of a fully differential feedback circuit is shown below (snipped from Gray and Meyer). This example, however, is a shunt-shunt topology after a Norton transformation is applied to the input.

In the case of a fully differential amplifier, I think it would look something like connecting the input voltage between the feedback network and the amplifier input (potentially wrong example shown below).

simulate this circuit – Schematic created using CircuitLab

• Your question is not sufficiently clear. Perhaps do a quick sketch of what you are thinking of. Either use schematic editor available on this site or even a photo of a sketch on a napkin. Commented Mar 1 at 16:35
• @DwayneReid is it more clear now? Commented Mar 1 at 18:20
• To apply series feedback to a fully differential amplifier, it would need 4 inputs. I remember seeing a chip like that... Commented Mar 1 at 18:23
• @bobflux that is the answer I am looking for. Commented Mar 1 at 18:49

Apparently it exists...

(source)

It's a FDDA "Fully Differential Difference Amplifier", searching this term yields articles and schematics.

In other words, it's an opamp where both inputs and the output are differential signals.

Since a differential signal has two "halves", it has double the number of pins.

It has the same structure as a differential amplifier, but it has two gm input stages (source).

There is a small subtlety: in the usual differential amplifier, as long as feedback is operating and it's not clipped, both inputs sit at the same potential, which is between the common mode of input and output, and differential voltage between the two inputs is very small.

However, the FDDA has two differential inputs, which means each has its own common mode. And in the first pic of this post, voltage between Vin1 and Vin2 is the input signal, which means it's not small, which means these "+" and "-" don't belong to the same input LTP. Otherwise, it would clip. Or it would need to be something else than a LTP.

A quick fix is to swap one of the polarities from one LTP to the other... but this only works when input and output common mode are the same, otherwise the input LTPs get too much differential voltage...

• Maybe spell out "long tailed pair" one time before using LTP. Commented Mar 1 at 20:51
• Interesting that the feedback and input signals are summed as currents at the drains of the input pairs. So maybe true series inputs, where input and feedback voltages are summed, only exist for single-ended amplifiers. Kind of an over-detailed point though, I think this is a series input fully differential feedback circuit. Commented Mar 1 at 20:53

The second image of a differential amplifier is very common with high speed diff amps. Something like the LTC6363 (randomly chosen from Analog Devices web site) uses that topology.

For precision instrumentation, an instrumentation amplifier is a common configuration. An oldie (40+ years old), but goodie, is the INA101.