Maker Pro
Maker Pro

Bandpass transimpenance amp simulation with analog switch

Mmm

Sep 11, 2017
4
Joined
Sep 11, 2017
Messages
4
I have a transimpedance amplifier circuit for a photodiode. The circuit is below.







There are other parts in this circuit as well. There is the photodiode, analog switch and another non-inverting gain stage that has a digital pot on the feedback resistor.



The photodiode model is below:





The Analog switch with input and output capacitance, series resistance and on leakage current is below:

Switch P# = SN74LVC1G66DCKR





Gain stage with digital pot is below. Digital pot is circled in red:

Digital Pot P# is MCP4011-103E/SN







Here is the simulation parameters of the photodiode current source:





Here is the output of the simulation:











Questions:

1) If I put a phase plot to the output, it looks like the circuit is unstable since the phase starts at -180deg and there is gain greater than 1 when it goes below -180. Is this circuit really unstable or am I reading it wrong? The pulse response suggests that the circuit is stable.





2) I want to add a high pass filter to the design to filter out low frequencies. I am only interested in pulses between 7uS and 20uS. Where would this best be achieved? Before or after the 2nd gain stage? I would be ok with a -3db point at 10kHz on the low end and a -3db at 1MHz on the high end.



3) Am I modeling the analog switch correctly?

ADIsimPE simulation file is attached
 

Attachments

  • AFE_ADA4500_Stable_Forum.zip
    5.9 KB · Views: 23

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
The left hand axis is showing dB not phase? or am I missing something. Also a high pass filter of type RC or LC works the way you want for sinewaves only, you are putting in a pulse with multiple sinewave content. Are you trying to make a sinewave from this pulse?
Thanks
Adam
 

Mmm

Sep 11, 2017
4
Joined
Sep 11, 2017
Messages
4
It's labeled dB for the green gain curve. The software did not label it with degrees as well. But I do believe the red phase curve is in degrees. It starts a -180 which i'm guessing is because the TIA acts a an inverting amp.

The goal of the high pass filter was to reduce as much low frequency noise as I can. But your right, I will need those frequencies to make a pulse on the output.

Here is the gain phase plot zoomed into where the phase crosses -360 deg. I'm assuming this is where I read the gain margin since the TIA will start it at -180?

attachment.php


When phase is -360deg the gain is around 15dB.

I'm just trying to make sense of the plots. The pulse response suggests a damped stable response with no overshoot. But I want to confirm stability with gain and phase margin. But when I go through it, I keep concluding it's unstable from the plot.
 

Attachments

  • temp.png
    temp.png
    38.7 KB · Views: 101

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
If you have no over shoot then the system will be stable. If you have over shoot then the system may become unstable. You have to be careful driving capacitive loads. Just out of interest put a 4.7 R resistor on the output of X2. Then connect everything else to the other side of the resistor.
Thanks
Adam
 

Mmm

Sep 11, 2017
4
Joined
Sep 11, 2017
Messages
4
The simulation has no overshoot. But once this is on a real PCB it might. That's why I want to make sure I have good gain and phase margin to make sure small changes won't make it unstable. Have any thoughts as to a way to simulate the circuit to get a correct phase and gain plot to get the margins? I can't tell if the plots are correct or not.

I put the 4.7 resistor on the output of X2. The simulation results are the same (Pulse response and gain/phase plot). Is this just to limit the inrush current supplied by X2 when C6 is discharged?
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
The resistor is a RISO resistor used when driving capacitive loads. Thing is unless you have a 2nd order response in the output when you still have gain then the system is stable. There is a way of working out the phase margin.... let me find my resource.
Adam
 
Top