A new mm-wave noise generator technology sought

[jiripolivka]

Up to now, in the mm-wave region, above say 50 GHz, there are several methods to generate noise, useful for calibration or creating noise fields:

1. Gas-discharge tubes matched to a waveguide: usually generate ENR (excess noise ratio) of 10-15 dB but need a high-voltage and high-power DC supply. Available from RF to ~400 GHz.

2. Avalanche-diode noise sources, in principle P/N junctions with reverse breakdown. Generate ENR of up to 30 dB, but less than 10 dB at >80 GHz. Poor performance above 80 GHz, not available above 100 GHz. Simple, low-power input, low-cost.

3. Using heated or cooled loads in a transmission line. Generate low ENR, like 400K against 300 K ambient, or, 77K (liquid nitrogen) against 300 K ambient. Very accurate and easy to use but need either liquid nitrogen (or hydrogen, or helium), or boiling water to function. Usable from DC to THz.

4. COLFET devices: FET or HEMT amplifiers "radiate" lower-than-ambient noise from their inputs while outputs are terminated. Usable at any frequency band where FET or HEMT amplifiers are available. ENR limited; e.g. at 23 GHz, generated noise temperature is ~70K, at 0 GHz, ~150 K. Useful to replace the cooled loads but barely available above 80-90 GHz.

5. In the THz frequency region, optical-fiber amplifiers were reported to generate a reasonable noise output, like 26 dB ENR over 300-330 GHz.
So far not available for 80- >100 GHz.

I would welcome anyone pointing out another technology suitable for >80 GHz, with ENR > 10 dB. Thank you all.

- - - Updated - - -

Please correct point 4 : at ~50 GHz, ~150 K. Thank you.

 

[biff44]

I have always thought that optical sources had a very flat and high microwave AM or PM noise content. For instance, a DFB laser (locked to one frequency mode), if you looked at its phase noise say 50 to 200 GHz from the carrier, it would be relatively flat. yes/no? maybe you can set up some sort of homodyne receiver with a wideband photodetector that would give you a flat noise output in the mmw region?

or maybe the same homodyne system, but driven by an LED with superbroadband optical noise output? The coherence length with an LED would be so short that you would get a strong mmw noise output in a homodyne system.

 

[jiripolivka]

I have always thought that optical sources had a very flat and high microwave AM or PM noise content. For instance, a DFB laser (locked to one frequency mode), if you looked at its phase noise say 50 to 200 GHz from the carrier, it would be relatively flat. yes/no? maybe you can set up some sort of homodyne receiver with a wideband photodetector that would give you a flat noise output in the mmw region?

or maybe the same homodyne system, but driven by an LED with superbroadband optical noise output? The coherence length with an LED would be so short that you would get a strong mmw noise output in a homodyne system.

Dear Biff44:

Thank you, I will think about that! My problem is that I have no access to optical laboratory to test similar ideas. I know that to generate THz signals, beating lasers is used on non-linear devices. So far this way seems too complex to me but maybe it can be realized soon.
My problem is to make noise sources for radiometer calibration. For such use avalanche noise diodes are the best, but their ENR drops fast above 90 GHz. Other methods are rather cumbersome.
I tested also various LEDs but their noise output can be used at RF, lower than 1 GHz.

Thank you very much!