I am building light balancing circuit. For higher bandwidth, I need lower feedback resistance (i.e) 100 ohm. When I use feedback resistance of 100 ohm and feedback capacitance of 2 pF. I am facing ringing effect. So, how do we select the capacitance according to the resistor. Will the opamp IC LTC6269-10 works with the feedback resistance 100 ohm. I am attaching my circuit.
Can you provide more details on your requirements?
It seems like your net input current is 0, is that correct?
What is your expected output voltage? What is your target transimpedance gain?
ADI has a tool called photodiode wizard tool in this link and will help you to determine if LTC6269-10 is suitable for your requirements and the tool will also help you with the recommended design values of feedback R and C.
Yes, Net current should be 0. I1 and C1 is the equivalent circuit of Photodiode 1, I2 and C2 is the equivalent circuit of Photodiode 2. The circuit is basically lightt balancing detection circuit. So, if light is equal, Output will be 0 V. If the light coming of the two photodiodes are equal. I need to use Rf of 100 ohm because I want to make this circuit work in higher bandwidth.
The resistor has nothing to do with it, look carefully at this condition:
Cf and Cin capacitors set the noise gain at high frequencies. This condition is not met in your circuit
CIN = CPD + CCM + CDM
CPD = Photodiode capacitance
CCM = Common Mode Capacitance of Amplifier
CDM = Differential Mode Capacitance of Amplifier
CIN = 0.25 pF + 100 fF + 450 fF
= 0.8 pF
CIN/Cf >= 10
0.8 pF / 0.025 pF = 32 which is greater than 10. But I am facing same type of error.
The product owner of LTC6269-10 forwarded me his thoughts on your query. Please read below.
"The LTC6269-10 has a “Gain-Bandwidth-Product” of 4GHz and a minmum Gain-of-10 requirement for stability. That means that in a gain of 10 it will have about 400MHz of bandwidth. There is no way to get 3GHz of bandwidth in it.
I think the only amplifier we have that could support this design is the LTC6409. It has 10GHz of GBWP, and is unity gain stable. However, I tried this in LTspice and the results are not promising. See attached.
You mentioned the input bandwidth is 3GHz. But do you really need 3GHz of output bandwidth? Photodiodes have extremely high reception bandwidth, but you can remove the charge slowly and not suffer losses. If this is a split photodiode app, for positioning purposes, then any mechanical system will be much slower and does not need such high bandwidth. So can the customer reduce the output bandwidth requirements? I once did a control loop for a 1GHz data bandwidth system, where the position control loop bandwidth (on a 4-way PD) was only 2kHz. It used LTC6081’s for their precision."
3GHz TIA attempt.asc