# Understanding Op-Amp gain bandwidth product

Below is the schmatics and AC transfer function for transimpedance amplifier from this design.

According to the datasheet, OpAmp has the gain bandwidth product of 20MHz.
But, here the gain is 94.58 dB = 54954 V/V and the bandwidth is 1.464 MHz.

So, Gain * BW = 54954 * 1.464 MHz = 80452 MHz, which is a lot than the value specified in datasheet. Can anyone help me understand how such high GBP is possible here?

The TransImpedanceAmplifier (TIA) has a gain (volts out to current in) equal to the feedback resistor. It is 53.6 kohm and take the log of this and multiply by 20 and you get 94.58 dB.

Gain bandwidth product is all about voltage gain - you don't have any voltage gain in a theoretical TIA circuit unless you are going to perform noise analysis.

If you were to analyse noise you'd realize that without the feedback capacitor (2p7) the amplifier's input's "self noise" is significantly amplified at higher frequencies due to the parasitic capacitance of the photodiode. It basically forms a gain stage from the equivalent input noise in series with the non-inverting input.

The 2p7 seeks to reduce this effect by progressively shunting the 53k6 as frequencies rise.

• The gain in the PDF is in V/A indeed. That's not quite "dB" as we have it in the OP's graph. Commented Oct 30, 2015 at 17:39
• @RespawnedFluff 20log$_{10}$(53600) = 94.58. Not sure what you mean? Commented Oct 30, 2015 at 17:41
• "dB" usually refers to unit-less units; they note it as "Gain (V/A, dB)". Also, the PDF does get to opamp GBW calculations on page 5. That mostly depends on the capacitors added (as I suspected). Commented Oct 30, 2015 at 17:44
• @RespawnedFluff I'm trying to figure out if you are saying my answer is somehow incorrect. Maybe you think I should add something? Commented Oct 30, 2015 at 17:47

You'll have to work through their calculations, but it's mainly the caps that determine the required GBW of the opamp, in order to ensure stability. Intuitively, the caps short out the higher frequencies, in particular C1 shorts out R1, so the [opamp] feedback/gain diminishes at higher frequencies. They came up with 9.653 Mhz GBW needed for the opamp, which the OPA320 satisfies. You have confused that with the gain of the TIA. See pages 5-8 in the appnote for their calculation of the opamp GBW (and opamp selection):