Schottky enhanced PIN diode limiter operation

[mtwieg]

Recently I was reading up on power limiters and pin diodes, and found a couple informative **broken link removed** on schottky-enhanced PIN diode limiters. I understand the concept well enough, but if the schottky has a lower threshold for conduction, then why wouldn't you just use antiparallel schottky diodes instead of a PIN at all? Does using a PIN/schottky combo get you higher power handling or something?

 

[Darktrax]

Thank you for the link on informative papers.

I am also trying to find ways of protecting the inputs of amplifiers, though in my case, they are low-noise LNA inputs. Though I don't know for sure, I think there are a number of issues about using Schottky or other regular diodes as opposed to PIN diode attenuators. We may be able to limit the signal to safe levels to protect the gate of a FET, but we pay the price of having it there messing with the wanted signals.

The diodes are inherently non-linear devices, and are widely used as large signal detectors, but they introduce DC offsets and noise into the small signals when they are not acting as limiters. They have a notorious temperature dependence.

You can divide diode effect into three regions. Above 20dBm, they can be considered linear, with current proportional to signal, and they switch on and off as fast as charge storage allows.

There is a "transition region" 20dBm down to -20dBm, when there is no well-defined line between conduction and non-conduction for small signals, and it becomes inefficient as a rectifier.

For really small signals, like below -20dBm, diodes once again can give predictable and useful amplitude measurements, because the DC component is proportional to the square of the signal input voltage. This might be a handy square-law thing that can be scaled direct to dB, but only if you are in the measurement business. If it was there as a limiter, it is going to mess up your life!

The RF model of the diode is hard to reconcile in matching and bias circuits, and it brings with it 1/f shot noise as well as temperature dependent dissipative noise. It make DC components which need to be isolated from FET bias circuits. If a band of frequencies is present, you have all kinds of intermodulation products also.
There may be clever circuits that can use these OK, but I posted before, and came up with none. At least PIN diodes can be used at microwave frequencies to make a matched attenuator, but that makes noise in a front-end that adds directly to the noise figure.

Maybe use a higher power FET to do the job of the Schottky, in some circuit where overload turns it on.
Sorry this was all so negative, but if anybody has a nice limiter circuit - then do tell.