WRT makes a lot of TDR and VNA measurements for our Channel Modeling platforms
out to 50GHz, routinely.
Much of below has already been touched on, both Tom and Gert making great
comments. This works 99.9% of the time for really good VNA measurements
suited for signal integrity (not RF Microwave):
Before I get beaten up over âthis wasnât exactly the questionâ¦â, a good
VNA measurement process is really important before digesting what is up with my
DUT and set up.
1. Use a late model VNA with low RX phase noise. If you think your 8510A is
still cool, great, just donât use it for making measurements.
2. Calibrate with a focus on DUT characteristics: low insertion loss, use a
short low loss adapter and SOLR calibration. WRT has done a score of
tutorials on the subject at DesignCon, they are on our website. We also wrote
a DesignCon tutorial on measurements with Teraspeed Labs that is still really
relevant.
3. Validate your calibration with a stepped impedance NIST traceable airline
(Beatty standard), we sell them and can configure a measurement validation set,
ping me. (https://wildrivertech.com/products-2/ ;
<https://wildrivertech.com/products-2/>) We just wrote a paper on the Beatty
Standard for DesignCon (and the PENN State SI Symposium), it is on our website
also.
4. Validate your calibration with the Anritsu terminators at each port, same
as 3. We find those terminators to work with VNA and the really fast TDR,
again ping me. I recall model 28Mk50 is the model.
5. Use a metrology grade adapter THRU for forward transmission check. This
is where passivity should be checked as well. This test is not sufficient,
however. Use the Simbeor tool, it also has excellent rational compact model
fit (used for D.C. and extending frequency range).
6. Collect all the S-parameters and run them through Simbeor S-parameter
quality metrics. Add the D.C. point if your handing the models out to
customers for time domain analysis. We have had success with the Simbeor
tool, and not have had success with other tools.
7. We will NOT use keepers and minimize adapters in signal path. Rosenberg
will provide a phase matched cable with the right sexes. They are expensive,
and we consider cables to be consumables. You keep a clean lab space with good
connector organization. If your company can pay the CEO $15M/year, they can
afford to feed the $200K VNA good cables - justify it, and use the other cables
as backups. I good metric is new cables for every major project requiring a
lot of measurements. Build the cost into the program.
8. Minimize cable movement during calibration, it increases measurement
uncertainty and wreaks havoc on SOLT and SOLR calibration.
9. IF BW 1kHz, 4000 points 10MHz to 40GHz, harmonically sampled so all
frequency spacing is spaced evenly. RCM models in Simbeor gets around the
harmonic sampling, but that sampling is required to VNA time domain transform
in case you want to use that. You want enough sample points for multiple
round trips of your test fixture length.
10. If your test fixture launches are junk, stop, design another fixture after
getting some help. Anything more than 5-10% impedance blip with a fast TDR is
a bit too much. We can help with launch designs.
There is also a lot of misconception regarding ECAL and mechanical cals. A
good ECAL provides better NIST traceable calibration than mechanical, although
we still see a need to have both.
- Al Neves
Products for the Signal Integrity Practitioner
Alfred P. Neves
Chief Technologist
Office: 503-679-2429
www.wildrivertech.com <http://www.wildrivertech.com/>
2015 Best In Design&Test Finalist
On Apr 26, 2016, at 10:15 AM, Ryan Lott <ryan.lott.ee@xxxxxxxxx> wrote:
I also agree with what Josiah says.
The effect he is referring to is called IF Delay and it can eat your lunch
if you aren't aware of it. When VNAs are operating in a continuous sweep
mode and measuring long duts, if the IFBW is too large, then this can
happen. A way around this is to use a lower IFBW or to force the VNA into
stepped mode and set the correct dwell time to account for multiple round
trip times. There is a sweet spot that you can achieve by adjusting the
IFBW, dwell time of the stepped mode, and the power level for passive
measurements; providing optimum speed and S/N of measurements.
Ryan Lott
Spectra 7 Microsystems Ltd.
Ryan.lott@xxxxxxxxxxxx
------------------------------------------------------------------
To unsubscribe from si-list:
si-list-request@xxxxxxxxxxxxx with 'unsubscribe' in the Subject field
or to administer your membership from a web page, go to:
//www.freelists.org/webpage/si-list
For help:
si-list-request@xxxxxxxxxxxxx with 'help' in the Subject field
List forum is accessible at:
http://tech.groups.yahoo.com/group/si-list
List archives are viewable at:
//www.freelists.org/archives/si-list
Old (prior to June 6, 2001) list archives are viewable at:
http://www.qsl.net/wb6tpu
