Hi

One of the most basic reasons for putting out > +20 dbm is that you
had a spec of -195 dbc / Hz for the noise floor :)

Some of these specs are a bit mutually exclusive.

Bob

On 2/6/17 2:37 PM, Bob Camp wrote:

Sure.. And to be honest, I'm not sure that some of the folks coming up
with paper requirements for these speculative low power transmitters are
aware of that.  They take dBc values from 1 Watt transmitters and assume
you can meet that with your 1 mW transmitter.

Then again couldn't you cool your oscillator.. that gets the T part of
the kT down lower <grin>

Cool that puppy down to <1K and get 25dB noise improvement, eh?

HI

Sounds like a low cost solution :)

The other proposed solution is to source the signal out of a zero ohm source. It’s not clear
which one actually costs less.

Bob

Hi,

On 02/07/2017 12:36 AM, jimlux wrote:

Your 50 ohm termination resistor will be a great source of that noise.
For a narrow-band fixed signal you can terminate with whatever reactive
network you feel confident with instead. If you match impedance well
enough it will work fairly well. Some oscillators have far-out
impedances far from 50 Ohm anyway so impedance matching is so-so and
most of the noise comes from the termination resistor.

Besides, for the deep space stuff you have cheap access to 2.7 K or so
anyway, right? :)

Cheers,
Magnus

Hi

Ok so, I have a device that puts out 1 mw (0 dbm) of power. I want the phase noise to be -195 dbc. What
source and load resistance do I use?  :)

I can use an infinite load, that will get me a whopping 3 db of noise improvement. That only
leaves another 17 or so db still to be found. Of course I’m not going to deliver 1mw into
an infinite load so we now loop back through what does putting out 1 mw really mean?

If my source is purely reactive I have a phase angle between voltage and current. I now have energy
coming out and not power. Back to the same question about 1 mw.

Yes you can run in and out of rabbit holes for weeks on this one :)

Bob

hi Magnus, how about the effect of that cheap 2,7K on the active device
if it is bipolar?

Greetings

Alex

On 2/6/2017 4:35 PM, Magnus Danielson wrote:

On 2/6/17 6:24 PM, Alexander Pummer wrote:

I wish it were easy to get 2.7k in space.. that's the temperature you're
radiating to.. At 300K you can radiate a few hundred watts/square meter.
When you have finally got down to, say 50K, the radiative heat
transfer gets pretty small. That T^4 bites you pretty hard.  And on the
surface of Mars, or Europa, or (worst yet) Venus, that 2.7K radiative
sink is not so easy to get.  And there's also that 5500K source in the
sky putting a kW/square meter into you.

But more practically - you see all sorts of cool idea for mesh networks
and what not.  But they're all "plug the 802.11 node into the 5V wall
wart" kinds of things.  Just the VCO to make the 2.4GHz probably
consumes a significant amount of power.

I'd like those mesh nodes to be, say, 10s of mW total power, when
they're on.  Another problem, of course, is that a superhet receiver
needs an LO to receive.

So very low power oscillators are of some interest - and low phase noise
is because, well, this is time nuts.. An RC blocking oscillator isn't as
interesting.

By the way, some colleagues are building a box to get some Rb atoms down
to picoKelvins.. Check out Cold Atom Lab.  Push a button and a couple
seconds later, you have a Bose-Einstein Condensate.