John Larkin wrote:
On Wed, 11 Jan 2006 14:10:32 -0500, Phil Hobbs
wrote:


John Larkin wrote:

On Wed, 11 Jan 2006 10:08:19 -0800, "RST Engineering"
wrote:

A hot resistor.

How about a thermistor or a lamp filament that was 50 ohms at some
high temperature. You could heat it with DC, sense its
resistance/temp, and let it make noise, all in a single part.

Old vintage noise figure meters used gas tubes. And I think there was
a pencil tube that mounted in a waveguide and made shot noise.

And, of course, the old photomultiplier trick.

I still like the flashlight/photodiode trick. You can get a really good
calibration just from the dc, and can calibrate the frequency response
with a spark plug.

Cheers,

Phil Hobbs



What's the light-flash waveform look like from a spark plug? What do
you drive it with?

Don't you have gobs of femtosecond lasers around your place?

John


You use one of the circular-gap plugs, run it to a HV supply via a 10M
resistor, and just discharge the capacitance of the plug--you get a nice
irregular relaxation oscillation. It isn't the absolute most beautiful
pulse, but (a) it's easy to shield so you get rid of the pickup, (b)
it's surprisingly bright, and (c) the rising edge is way under 1 ns,
which should be fine for the VHF to low UHF range. I might stick one on
my sampling scope sometime and find out more about its actual
performance, but this is a pretty common trick.

There are femtosecond lasers around here--my fastest one is about 20 ps,
but it's continuously tunable from 420 nm to 10 microns, when it's working.

Cheers,

Phil Hobbs