DTF and Time Domain Measurements with Copper Mountain Tech. VNAs

DTF and Time Domain Measurements with Copper Mountain Tech. VNAs


Welcome to another video in the introductory
series from Copper Mountain Technologies in Indianapolis Indiana. Today we’ll introduce you to performing time-domain
measurements with our vector network analyzers, which is useful for measuring the length and
loss of a cable such as a feeder, for localizing impairments in a cable such as a mismatch
or damaged cable or connector, for tuning multi-pole filters, and many other purposes.
Time domain measurements are a standard feature of all our VNAs, but methods of setting up
and using measurements are a little different in different instrument types. So we’ll start
with an example of Distance to Fault in the R54 and R140 vector Reflectometer, then show
you time domain in the TR5048 analyzer, which is a full two port 1-path device, with 50
Ohm ports, which measures up to 4.8 GHz. First though, what is time domain? In VNAs,
time domain is just another display format, like Phase, Log magnitude, or Smith Chart.
Rest assured that the measurements actually being performed by the instrument are still
our trusty S-parameters over frequency, like S11, with the number of points, frequency
range and power which are all set up in the normal way. Of course, all the other display types always
show the measured parameter over frequency, whereas the time domain mode shows a result versus
time. So how does that happen? Well the result of the normal measurement is transformed through
algorithms to convert from frequency to time. The end result is in the time domain. You could say it’s similar to a TDR result, though again the VNA doesn’t physically transmit a pulse but a series of CW tones. So how do we set up and use time domain measurements?
The important settings are number of points, frequency range, and cable parameters. Let’s
talk about those in turn. And just a heads up, I’m going to use the
rule of thumb of a foot per nanosecond here, for convenience. First, the number of points that are measured
between the start and stop frequency. Obviously more points will take longer to measure, but
the benefit of more points per GHz is to increase the longest reflection time we can measure.
For example if we have 51 points over a 1 GHz span, about 50 points per GHz, the largest
time we can measure is about 25 feet at the speed of light. But if we increase that to
500 points, we can measure reflections out to about 250 feet. So more points per GHz,
more distance. A slower sweep time is the main tradeoff. Second, frequency range. There are two main
aspects here, the limitations of the device we’re testing, and then the tradeoff we can
make given that limitation. When we say DUT limitation, all we mean is the frequencies
where we expect either a small reflection if it’s a load-like DUT, or a strong reflection
if the DUT is something like a cable with a short or open its far end. In either case,
we should set the start and stop frequency according to the useful bandwidth of the system
we’re testing. And notice that we can set the distance start and stop to anything we
like, less than this maximum, so don’t forget to zoom in or out to the distance range you
want to see. Now within the DUT’s frequency range limit,
the greater a frequency span we can use, the better resolution we’ll have in the time domain
result. A 2 GHz span gives us a time resolution of about 1 foot. Doubling the span improves
resolution 2x, so a 4 GHz span gives a resolution of half a foot, and so on. At its maximum
frequency span of 85 MHz to 14 GHz, the R140 can give you a time domain resolution of about
140 picseconds, which works out to about 2 inches in free space, 4.3 centimeters to be
exact. Now of course, if you keep a certain points per GHz as above, increasing the frequency
range increases the number of points, which slows down the measurements, so there’s a
tradeoff here as well. The final group of settings are those used
to convert from reflection times to reflection distances in various units. I’ve been using
a foot per nanosecond so far for convenience, which is close to free space velocity of propagation,
but if you’re using a particular cable type you can find out its velocity factor and use
that for a more precise distance estimate. In fact, the reflectometer has a table of
such cable types and their Vps built in, which you can use as-is or customize with your own
cables and parameters. You see that table also contains a cable loss
coefficient. This is especially useful if you’re testing a cable with an open or short
at the far end. If that cable has 2 dB loss for example, a 0 dB reflection will arrive
back to the instrument with a magnitude of -4 dB, because it looses 2 dB each way. The
loss coefficient is used to “boost up” reflections coming from further down the cable, to better
approximate the true magnitude of the reflection. OK, so as we’ve gone through the important
considerations you’ve been seeing how to set up time domain measurements in the reflectometer
software. Let’s hook up a real DUT and look at its reflection. I’ve got a horn antenna
here, which is pretty effective above about 6 GHz, you can see its return loss is better
than about 10 dB. Now let me choose the number of points to use. I want to look for reflections from the
environment here in my office, 10 feet is plenty of distance, so I’ll choose 20 points
per GHz, that’s 140 points in total. So I’ll set my frequency span up for its full bandwidth, say 7 GHz to the instrument limitation of 14 GHz. That’ll give me a time domain resolution of about 4 inches. OK, so hey cool, you can see the reflection
coming from my hand as I move it closer to and farther away from the antenna. And here,
let me measure the distance from my desk to the ceiling, it looks like about 7 feet which
seems about right. Pretty cool. Finally then, I mentioned time domain is a
feature of all our instruments so let’s make a measurement with the TR5048 analyzer. One
difference here is the TR software has direct Vp entry instead of the cable table, so just
look up the Vp for your cable type and it in here. I’m testing a cable here in my office that
looks to be about 10 feet long, and I’ve got it terminated with a short. I’ll measure from 20 kHz to
4.8 GHz Since the span is a bit less than earlier, I’ll have about 6 inch resolution.
So I’ll put my marker on the peak, you can see that’s about 35 nanoseconds
total, 17 nanoseconds each way. I happen to know Vp for this cable type is 0.7, so the
marker is really telling us… 24.2 feet round trip. Just a little trick here, I can halve
Vp to get the one-way delay of 12.1 feet, which is good because I also happen to know it’s a 12
foot cable. So, that’s an introduction to how you can
set up and use time domain or distance-to-fault measurement modes in Copper Mountain Technologies
VNAs. I’d encourage you to download our VNA simulator at www.coppermountaintech.com and play around with the feature yourself, to get familiar with its capabilities. You
can download any of our software for free from the website. Thanks for watching!

1 Comment

  1. can it be converted as Spectrum Analyzer? hope this VNA has API or ar least existing apps cann save Spectral / Spectrum Amplitude into computer harddisk

Leave a Reply

Your email address will not be published.


*