Hi Dave, thanks for responding.
I looked into the Wenzel reference just now and found -170 dBc/Hz at 100
kHz for one of their 100-200 MHz products. That's pretty close to the
spec for a 25 MHz low phase noise circuit I bought on ebay for $100, and
just above the thermal noise floor for a clock with a few dBm or so of
power.
If you had a file of samples from the Wenzel or one of your other
sources you are willing to share, I would like to take a look.
My basic question is fairly simple. Are modern samplers low cost with a
good reference clock stable enough to measure phase noise that low with
real samples only and without the need for two channels or cross
correlations? I could answer that question I think with a file of
samples from a decently stable setup.
How are you clocking the TI ADC in your tests?
Best,
Karl
On 8/20/2024 2:49 PM, 'Hawkins, David W (US 334B)' via
casper@lists.berkeley.edu wrote:
Hi Karl,
Thanks in advance to anyone whose interest is piqued enough to respond.
My interest in piqued ...
I'm in the process of measuring phase noise using a Rhode&Swartz on some EVMs
and custom designs:
1. Texas Instruments DAC39RF10 DAC (I have rev1 and rev2 of their space parts)
2. Texas Instruments ADC12DJ3200 ADC (QMLV part)
3. LMX2615-SP synthesizer
4. LMK04832-SP clock distribution
5. Wenzel low-phase-noise 108MHz and 2700MHz reference (5400MHz and 10800MHz
also options)
6. Keysight N5183B synthesizer
The DAC and ADC are JESD204C devices. The application is radars (space-based).
The ADC EVM plugs into the FMC site on a KCU105, so could fill the DDR on that.
I'm playing with these toys to determine what we don't link. So far the LMX
device is the weakest-link (largest source of phase noise). My original plan
had been to take 108MHz to make 5400MHz for the DAC, 2700MHz for the ADC, and
168.75MHz for the FPGA REFCLK/CORE clock. But the phase noise of the LMX may
change my mind to bring the 5400MHz directly from the Wenzel. The Wenzel phase
noise is nicer as they use doubler to get to higher frequency rather than a PLL.
So my "toys" might just have the low-phase-noise and higher-phase-noise that you
are looking for to provide your algorithms. We would have the Rhode&Swartz analyzer for
comparison.
Regards,
Dave
Dr David Hawkins
Technical Group Supervisor
Radar Science & Engineering Section
Radar Digital Systems Group (334B)
Jet Propulsion Laboratory
4800 Oak Grove Dr
Pasadena, CA 91109
Office: 300-235R
Phone: 818-354-2252
Cell: 626-720-7079
https://radar.jpl.nasa.gov/
-----Original Message-----
From: casper@lists.berkeley.edu <casper@lists.berkeley.edu> On Behalf Of Karl
Warnick
Sent: Tuesday, August 20, 2024 1:38 PM
To: casper@lists.berkeley.edu
Subject: [EXTERNAL] [casper] Low cost phase noise analysis
Hi all,
I've spent some time this summer as part of a radar project digging into
calculating phase noise for highly stable tones. I have implemented what I
think is a decent algorithm. My next steps are to look for test data sets and
tips for the hardware.
Do you have a file of samples of a stable tone? If anyone has a test data set
consisting of samples of a pure tone that they would like to share as a test
data set, I'd like to apply my codes to that and check the phase noise. Both
the tone generator and the ADC sample clock should be phase stable to the order
of a Keysight signal generator, or ideally better. The data set length should
be a reasonable fraction of a second for ~1 Hz phase noise resolution. The
frequency of the tone and the sample rate are fairly arbitrary as I'm mainly
looking to benchmark the algorithm.
How cheaply can stable samples be acquired? I'm looking for low cost hardware
(a few $100s up to a few $k) that is stable enough to measure phase noise
comparable to a Keysight source or better. Phase noise can be measured with an
expensive phase noise analyzer, but I believe it should be possible to do this
with a low cost digitizer with a suitably stable sample clock. The sample clock
could (or perhaps must) be external. The sample rate should be around 80-100
Msps or higher and the platform should be able to store a burst of samples of
length on the order of 1 sec. We have done this using a ZCU 216 and it seems to
work, but that isn't really a low cost board. I've looked into Picoscope
products, which might be ideal, but their support people don't know anything
about the phase noise properties of their samplers.
Thanks in advance to anyone whose interest is piqued enough to respond.
Best,
Karl
--
Karl F. Warnick
Parkinson Engineering Research Professor Department of Electrical and Computer
Engineering Brigham Young University
450 Engineering Building
Provo, UT 84602
(801) 422-1732
--
Karl F. Warnick
Parkinson Engineering Research Professor
Department of Electrical and Computer Engineering
Brigham Young University
450 Engineering Building
Provo, UT 84602
(801) 422-1732
--
You received this message because you are subscribed to the Google Groups
"casper@lists.berkeley.edu" group.
To unsubscribe from this group and stop receiving emails from it, send an email
to casper+unsubscr...@lists.berkeley.edu.
To view this discussion on the web visit
https://groups.google.com/a/lists.berkeley.edu/d/msgid/casper/151d449f-5425-4afe-9102-6f3f692fb2a2%40ee.byu.edu.
