ICL7660 to generate -3v for negative gate bias is a good choice? (adds Phase noice?)

In few amateur projects i found 7660 IC used with few kOhm trimmer pot adjustable resistor (for voltage divider). Then we can adjust gate biasing from 0 to -3.3v using trimmer pot to tune transistor for best performance (multiplier, oscillator, amp). I do not know how to understand impact of 7660 voltage ripple on phase noise. Will it be bigger than without 7660 with only self-biasing?

Using 7660:
- more phase noise? (not obvious)
- less power consumption? (not obvious)
- easier biasing for certain load line
as i understand, when we have negative power source, some things become easier.

Using self-bias:
- less phase noise
- more power consumption? (not obvious, more power consumption if i increase voltage. less efficiency if voltage is the same, because power dissipated in Rs)
- more difficult to tune gate negative voltage bias
- cannot use bigger voltage to compensate drop around Rs, because load line can become dangerous (not obvious)

Hope to hear about that not obvious things.

My experience of 7660 being used to provide gate voltage has been in small signal HEMT microwave amplifiers, often receiving signals from low power orbital satellites. The reason it was used was nothing to do with noise, it was simply that self biasing relied upon a voltage being dropped across the source resistor and at 14GHz it wasn't easy to by-pass it it to keep the source impedance low enough. By directly grounding the source pins this problem was removed. Applying a negative voltage to a high impedance gate circuit was far easier to achieve, easier to adjust and a single 7660 could supply several amplifier stages. An alternative we used sometimes was a 74LS04 hex inverter wired as a simple RC oscillator, with it's output capacitively coupled to a voltage doubling negative rectifier. It did the same as a 7660 but worked out cheaper to produce.

Brian.

in general any switcher regulator will work. Since a gate typically draws no dc current, u might use a switched capacitor type regulator (no pesky magnetic field to shield). A simple R-C filter with a whopping big C should filter out the noise enough for most applications.

Of course, if you are looking for very low phase noise, linear supplies are the way to go.

Did see one application once where they rectified the RF signal using a schottky diode, and generated enough negative bias that way to control the Fet gate. Was....interesting. No phase noise with that...but maybe some turn-on issues.