Feedback resistors values in fully-differential amplifiers

[Junus2012]

Dear friends,

The feedback resistors in fully-differential amplifier have a loading effect on the amplifier core. However, I mostly see from data sheets how they connect small resistor values like the circuit shown below

1. The value of 150 Ohm is heavy load, why not using bigger resistors to relax the amplifier driving capability
2. The low frequency differential input impedance will be 2 X 150 Ohm = 300 Ohm. So now we are loading the previous stage as well. Additionally it will make voltage divider with the output resistance from the previous stage that make loss in gain or altering the closed loop gain of the fully-differential amplifier

I would like to have your discussion on these points

Thank you in advance for your help and best regards

 

[KlausST]

Hi,

* the amplifier is designed for that low ohmic loads.
* the input impedance of that high frequency signals should be low, too. (50 ... 120 Ohms)
* higher value resistors add noise

Klaus

 

[FvM]

You are asking about an amplifier with about 2 GHz closed loop bandwidth for the shown G=1 configuration. You forget the role of parasitic capacitance (amplifier input capacitance, resistor parallel capacitance) for circuit behaviour. You should perform some survey simulations to see their effect.

There's a reason why Rf has a parallel capacitor in the application circuit. Omitting it and increasing Rf a bit will cause amplifier instablity by working of the input capacitance.

 

[Junus2012]

Thank you friends for your help,

You covered my inquiry and you are right, larger resistor cause RC input oscillation, I have simulated it with similar circuit and get bad ringing effect.

 

[Junus2012]

Dear friends,

I have a relevant question,

the input impedance of this amplifier is ideally 300 Ohm (2 X 150 Ohm),

They can effect the previous amplifier stage in this way

1. Loading effect
2. Can cause a voltage divider with the output impedance of the previous amplifier especially if the latter one has not small output impedance. Generaly even low output amplifiers tends to have larger value at highr frequency or at high closed loop gain, assuming here talking about voltage feedback amplifiers


I would like to have your discussion on it

Thank you

 

[d123]

Hi,

Low-level design-wise, not engineer-level, I always put a voltage follower as a buffer between op amp stages that have restive networks at input that I believe will cause undesirable interaction, then for such designs I post schematics of here for queries, Klaus tells me half or all of the buffer stages are not necessary. But I think like you (seem to) - Rout and Rin forma divider that must distort thr input signal, even slightly, and maybe enough to be undesirable, depending on goal(s).

If you can be certain of Rout, then Rdiviver calculation shows degree of effect.

 

[KlausST]

Hi,

My answer is already given.

Joke aside.
I don't know for sure. My guess:
This is a high frequency amplifier.
You don't design HF for high impedance, because stray capacitance would kill gain flatness vs frequency.

My idea: if you have a 10 kOhm node just 2pF stray capacitance then you get a cutoff frequency of poor 8MHz
1kOhm, 80MHz
...

Klaus