4 ghz opamp

--sp

Sure, CMOS input buffers followed by lots of really tiny bipolars.

The input capacitance is nice.

Yes, a very small die part, with high current density.

Thanks to clever bootstrapping, the input capacitance is truly awesome, only 0.45pF and 0.1pF differential!

I think the 0.45 is what you see at the inverting input in a TIA config, which is still damned good.

It only costs $6 or something like that.

The effective input capacitance is small, but the usual reason for caring about that is the rising current noise, which this part does nothing to fix.

Cheers

Phil Hobbs

The phase shift with that kind of op-amp is always very large in any of the data sheets I've looked at (only a few). Is there some reason for that? Is it reduced by negative feedback to more reasonable levels?

Not intended for unity gain.

Cheers

Phil Hobbs

I guess it has the same problem of a bootstrapped jfet, namely the voltage noise of the bootstrap driver, this one being cmos. But if the semiconductor geometry is really small, and the pre-bootstrap capacitances are small, there is some real benefit.

But 7 fA/rthz is mediocre. Some fet amps are 30, some are below 1. Not bad for a 4 GHz amp!

The open-loop gain -3 dB point (45 degree lag point) is about 3 MHz for this part. Most opamps are -3dB at 10 Hz or so.

All opamps have close to 90 degrees of lag in their operating range. Yes, negative feedback reduces closed-loop gain and phase shift.

You said it. But you would never use something that noisy as a bootstrap in a built-up circuit, and there's no way to turn it off.

Since it makes the noise worse, they obviously did it to keep from having to use 100-ohm feedback resistors to avoid instability. It sort of looks bad if your part has to have a phase lead cap to avoid oscillating.

Cheers

Phil Hobbs