OMG, you need two PSG's AND a PXI cage with lots of cards.
Another doomsday machine (it replaces the E5505A which I
have a lot of experience with, and THAT was a doomsday
machine for sure.  The blurb calls it the "gold standard".
Right, only when it comes to price).  It even says it is designed
for Power Users at the high-end of the market.
Power Users with deep pockets.

My guess is well into 6 figure$.

(Note:  I am a Keysight retiree).

BTW, I don't know what -177 dBm/Hz has to do with phase noise.
The relevant units are dBc/Hz as any time-nut knows.

Rick

On 8/15/2019 3:42 PM, Dr. David Kirkby wrote:

The thought of what it must cost is frightening. A colleague just acquired
a used 5503 for a bit under $50K.

On Friday, August 16, 2019, Richard (Rick) Karlquist richard@karlquist.com
wrote:

--
Homo sum humani a me nihil alienum puto.

It's a modular system like its 3048A and E5500-series forebears, so the only limits are your imagination and your bank account. :)  Very cool hardware, noting that it has some competition from all-in-one instruments like the FSWP that the earlier 'doomsday machines' didn't have to face.

I'm sure you could spend a quarter of a million dollars on either the FSWP or the N5511A if you checked all the boxes.

I haven't seen PN analyzers rated that way before, but it makes sense to a certain extent.  They are basically saying they can measure the noise on a 0 dBm signal down to -177 dBc/Hz, with signals above 0 dBm being measurable with a commensurately-lower floor.  Their rated power limit is +20 dBm, so -197 dBc/Hz would be the theoretical limit.

The brochure goes into that in some detail, but there's no mention of cross-spectral collapse, which I thought was interesting.  The white-noise floor in their example plots is enviably flat, with no divots, suspicious-looking valleys or other artifacts.  If they are able to avoid that problem in the general case, it would be interesting to hear more about the strategy being used.

-- john, KE5FX
Miles Design LLC / Jackson Labs Technologies, Inc.

Hi,

On 2019-08-16 08:30, John Miles wrote:

Indeed. You get some pretty good stuff when you do.

Well, assuming the internal noise becomes a limit, which does not
necessarily need to be, but it is the traditional limit.
Cross-correlation can measure below it, but you then need to avoid
spectral collapse as well as AM-to-PM conversion which is another issue.

Most hand-wave around that question. I've only see two reasonable
approaches, the onces we two provided, each with it's benefit and
drawback. A third approach has been proposed, but I have not seen a
realization of that.

Agree. It's a tricky one as one try to reach that level. The approach I
proposed is so far impractical even if it works.

Cheers,
Magnus

On 8/15/2019 11:30 PM, John Miles wrote:

It may be helpful to explain to the group how the Keysight
phase noise business model operates.  First of all, it is the same
business model as when they were Agilent, and before that,
HP, all the way back to the 3047.  I worked there through all
those eras.  Second, it is a unique business model vs the models
used for the underlying instruments.

Basically, phase noise as a market has never been worth enough
money to justify a conventional product line paradigm.  It
was always an orphan that no division manager wanted.  OTOH,
having nothing in that space detracted from HP as the complete
solution provider.  The solution was to have Spokane division
donate Bob Temple to champion a product composed of existing
instruments plus one purpose built "miscellaneous" box (eg,
the 11848A, etc) that contained everything else.  The divisions
got some incremental sales with no R&D effort.  There were
probably negotiations in the proverbial smoke filled room about
how to divide up the pie.  I remember on the 3048 they created
a "3048 system reference number" with a price of $ZERO that
was necessary to give the salesman the right commission and
perhaps collect an honorarium from each division to support
Bob.  The software had to run on a separate computer rather
than residing in any instrument.  Even the miscellaneous box
is totally dumb.  Bob, of course, has been retired for some
time, but can be hired as a freelance consultant.  I never
found out who stepped into Bob's very large shoes for this
latest product.  In any event, this business model predictably
results in high end prices.  Another part of the business model
is that the well-heeled customers will spend generously to
buy the time of Keysight customer engineers.  When I worked
there, I could simply walk down stairs and get free advice
from these guys.  Since the customers paid for these guys,
there was no issue about where to get the money for them.

For whatever reason, off shoring to Malaysia didn't seem to
result in lower factory costs than Santa Rosa.  Or maybe
it did and the price would be 0.5 million otherwise.

I wish my former colleagues at Keysight the best of success.

Rick

Its only necessary (as NIST have shown) to cool the splitters to reduce the correlated or anti-correlated noise between the outputs. Everything else can run at ambient temperature.

Bruce

That should have been:

Its only necessary (as NIST have shown) to cool the splitters to reduce the correlated or anti-correlated thermal noise between splitter outputs. Everything else can run at ambient temperature.

Bruce

Hi,

There is actually more approaches. Me and NIST showed you can setup a
cross-correlating interferometer, it actually works, but is so far hard
to maintain properties, but the principle works so we presented and
published it.

Another approach suggested by Enrico is to actually compensate with the
expected noise-level of the power-divider resistor, and you do not need
to match that extremely well to cover most of the noise.

Yet another being used when you do not need to go all the way down is to
have a common resistor and then use active buffers for power-splitting.
It's not perfect and there will be a common mode noise that remains, but
it will be robust to cancellation due to power-splitter noise of
anti-correlation, so for many purposes this is good enough.

Thus, there is three different methods to work around it and one
pragmatic to dodge it. So you can avoid cryogenic methods, but none of
these methods is perfect, we have not solved it completely.

Cheers,
Magnus

On 2019-08-21 11:16, Bruce Griffiths wrote: