OPA2320: A wide-band, high gain, low noise amplifiers system

Hi Michael,

Can you explain why you suggest LMH6229, since there are various high-speed opamps with gain-bandwidth product higher than your suggestion, such as OPA855. Is your concern noise performance?

Hi Yubing,

quite obviously because the LMH6629 is an ultra-low noise OPAamp. Compare the numbers 

Kai

Hi Kai,

Thanks for your reply.

I have a further question. The design target is 1.5GHz close-loop bandwidth and 60dB gain. But the GBP of LMH6229 is only 4GHz, therefore the maximum gain of the first stage is approximately 6dB, which is 54dB shy. What about the rest of the signal chain?

In addition, I know there are amplifiers called 'RF amplifier'. I think they are more appropirate than the suggested high-speed opamps for this application. Maybe I'm wrong, and looking forward to your comment.

Hi Yubing,

it depends on how much gain stability you need. For highest gain stability an OPAmp should be used.

A typical RF amplifier, on the other hand, shows a big dependency of temperature, frequency, supply voltage, etc. on gain stability.

Simply put: Less feedback -> higher bandwidth and lower precision.

Kai

Hi Kai,

For this application, the 1st stage consists of a low-noise, high GBP opamp, which provides highest possible gain. As for the rest of the signal chain, we can care less about the noise figure and should choose opamp with highest GBP. Is this strategy correct?

The next step up in bipolar (lower noise) input op amps is the OPA858. Running inverting, might be able to tune this into a >1.5GHz with some gain. 

But really, 1.5GHz with that kind of total gain is pretty problematic. 

For RF amps, you give up the DC coupling but there are many choices if that is ok. 

Here I tried the OPA858 in inverting matched input impedance mode, not much gain and if I comp it flat with those external C's only 1.2GHz BW. Not sure how you are going to get there if you need DC coupling, The peaked curve is without the caps, 

Hi Yubing,

yes, the noise of the first stage usually dominates the overall noise of composite amplifier. You might want to have a look at this thread where I discussed this point in more detail:

https://e2e.ti.com/support/amplifiers/f/amplifiers-forum/993200/opa356-amplification-for-ultrasonic-ndt-signal

Kai

Hi Kai,

Thanks, I definately will. Appreciate your reply.

Thanks Micheal. I would recommend using the feedback capacitor, since otherwise the phase margin would be lower than 45deg.

Well Yubing, a feedback cap by itself will shape the noise gain to 1 which will oscillate for a decomp VFA. The cap on the inverting node to ground shapes the noise gain to a cap divider at higher F - so you need both in this case.