Gain of transconductance amplifier

Pavel Yermolenko

Intellectual 280 points

Hello,

What does exactly mean the term gain for transconductance amplifier where the input is current ?

For example in the datasheet of LTC6268-10 there is the following notion:

"It is a decompensated op amp that is gain-of-10 stable"

The input signal (current) comes from photomultiplier (some kind of photodiode).

There is also feedback resistor Rf.

How gain is calculated ?

Thanks.

over 3 years ago

0 Sean Cashin over 3 years ago

TI__Mastermind 23226 points

Hi Pavel,

They are referring to the "noise-gain", which is the gain at the output for a small signal applied to IN+. An op-amp in a transimpedance circuit has a voltage gain of 1V/V at low frequency and DC. However, the input capacitance from the photodiode and the feedback capacitance across Rf can be used to set a higher gain at high frequency. This higher gain is what needs to be >10V/V in order to make the amplifier circuit stable. The noise gain depends on the circuit itself, not on where the signal is applied.

Here is an article that helps explain the stability of transimpedance amplifiers in greater detail for your reference: /cfs-file/__key/communityserver-discussions-components-files/234/Op-Amp-Bandwidth-for-Transimpedance-Amplifiers.pdf

Best regards,

Sean

0 Pavel Yermolenko over 3 years ago in reply to Sean Cashin

Intellectual 280 points

Hi Sean,

Thanks.

For inverting opamp the gain at DC is -R2/R1.

For non-inverting opamp the DC gain is 1+R2/R1.

In my case there is R2 with known value between output and negative input and let's say "photodiode", connected to negative input. Positive input is at a given bias. How can I check that such circuit will be stable at DC ?

Sincerely

Pavel

0 Pavel Yermolenko over 3 years ago in reply to Pavel Yermolenko

Intellectual 280 points

Hello,

In an article I've found on another forum, the author proposed a criteria to evaluate the stability of a particular circuit, containing opamp.

Here is this criteria: "For unconditional stability, the noise gain curve must intersect the open loop response with a net slope of less than 12dB/octave (20dB per decade)"

I've tried to check my setup using this criteria.

Here is my LTSpice setup (not sure if it's correct):

LTSpice was chosen because of opamp comes from Analog Devices.

1st opamp - from usecase, 2nd - in order to have open loop gain on the same graph.

The photodevice passives are modeled as Rd||Cd (for the moment not sure that the parameters are correct).

Here is simulation:

According to this graph (if setup is correct), the circuits is stable because close loop gain crosses open loop gain with slope less than 20dB/decade. BTW the slope in intersection point is negative, so it seems that the circuit should not have stability problem at all.