Hi guys -

I'm trying to calculate the bandwidth of my transimpedance amplifier. I was assuming that you can simply divide the GBWP by the feedback resistor you incorporate (and hence the transimpedance gain) to find your GBWP limit, but I don't think this is the way to do it. I would appreciate a more knowledgable insight into this calculation!

Thanks a lot

 

You calculate the amplifier circuit bandwidth by calculating the gain from the non-inverting (+) input to the output and dividing the GBWP by that gain. Thus if you have an inverting gain of -1, the gain from the non-inverting input would be 2 so you divide the GBWP by 2 to get the circuit bandwidth. This assumes there are no other frequency determining components in the op amp circuit.

 

I thought this was the case too crutschow, but transimpedance amplifiers are not governed by the GBWP in the same way normal amplifiers are -

Page 11: http://www.ti.com/lit/ds/symlink/opa846.pdf

"Adding the common-mode and differential-mode input capacitance (1.8 + 2.0)pF to the 50pF diode source capacitance of Figure 3, with a 10kΩ transimpedance gain using the 1750MHz GBP for the OPA846, requires a feedback pole set to 16.1MHz. This requires a 1pF total feedback capacitance. Typical surface-mount resistors have 0.2pF parasitic capacitance leaving a required extrinsic 0.8pF value, as shown in Figure 3. Equation 2 gives the approximate –3dB bandwidth, if CF is set using Equation 1."

Equation 2:

This theory is backed up at the beginning of the OPA656 datasheet:
http://www.ti.com/lit/ds/symlink/opa656.pdf

"Broad transimpedance bandwidths are achievable given the OPA656’s high 230MHz gain bandwidth product. As shown below, a –3dB bandwidth of 1MHz is provided even for a high 1MΩ transimpedance gain from a 47pF source capacitance."

If transimpedance amplifiers worked as normal voltage amps do, a 1MΩ gain would result in a bandwidth of 230Hz, using this specific op amp.

 

Have you seen these analyses?
http://www.ti.com/lit/an/sboa055a/sboa055a.pdf
http://www.maximintegrated.com/app-notes/index.mvp/id/5129
http://www-f9.ijs.si/~margan/Articles/trans_z_amplifier.pdf

 

I thought this was the case too crutschow, but transimpedance amplifiers are not governed by the GBWP in the same way normal amplifiers are -

.....................................

Well, I did give myself an out with my statement "This assumes there are no other frequency determining components in the op amp circuit." Obviously the various input stray capacitances do affect the bandwidth.

 

Ron, thanks for those links - I'm studying them as we speak