Terminating to the virtual ground of TIA

Thread Starter

Guttalax

Joined Feb 3, 2018
2
Greetings. Please, consider the following circuit (attached).

The source may create current type signal ranging from DC to 400 MHz. It is then filtered by an LC filter which is damped (matched, whatever) by a 50 Ohm resistor. At this point I came up with the idea that I might try to direct the source current not to the ground but to the virtual ground created by LTC6268-10 (https://www.analog.com/media/en/technical-documentation/data-sheets/626810f.pdf) opamp which is used as transimpedance amplifier in this application. The output of the opamp is expected to be of voltage type with relatively small amplitude of about 150mV (hence, the feedback resistor value is relatively small).

The problem is that I have never tried this IRL and do not know if that will work. The questions are:
1) Is it legal to terminate to a virtual ground?
2) Can TIA work with such a low gain?
3) Anything else I've missed? I feel like I don't know what I don't know. Any comments would be appreciated.

T56d93pIkTU.jpg
 
Last edited:

OBW0549

Joined Mar 2, 2015
3,446
The problem is that I have never tried this IRL and do not know if that will work.
I doubt it will work satisfactorily. This is one of those things that looks good on paper, but once you get down to particulars it falls apart. The devil is in the details.

The questions are:
1) Is it legal to terminate to a virtual ground?
Yes, provided the virtual ground remains a virtual ground at the frequency of interest.

In this case it will not, due to the very limited open-loop gain (≈ 10 db) and high open-loop output impedance (≈ 40Ω) of the op amp at 100 MHz.

2) Can TIA work with such a low gain?
Yes; that's not an issue.

3) Anything else I've missed?
Since you're merely attenuating the input signal, just get rid of the op amp altogether and terminate your filter in a voltage divider consisting of a 32Ω resistor in series with an 18Ω resistor. You'll get better results that way.
 

danadak

Joined Mar 10, 2018
4,057
There is a demo circuit here that you could mod so that you can get a better
handle on Zout for your specific circuit, where G is << 1. You could add a
current source driving the output to make a test case to evaluate Zout.

Or just eval the circuit overall.

https://www.analog.com/en/design-center/evaluation-hardware-and-software/lt-spice-demo-circuits.html

The demo circuit is a scope diff probe, just mod it to see what sim results you get.

How good the sim is at RF questionable, I do not see in the sim any stray C layout
lumped elements, or for that matter PCB trace L. But try it. If it looks poor strays will
only make it worse.


Regards, Dana.
 
Top