Call Search

New to Ham Radio?
My Profile

Friends Remembered
Survey Question

DX Cluster Spots

Ham Exams
Ham Links
List Archives
News Articles
Product Reviews
QSL Managers

Site Info
eHam Help (FAQ)
Support the site
The eHam Team
Advertising Info
Vision Statement

donate to eham
   Home   Help Search  
Pages: [1]   Go Down
Author Topic: How do I measure the phase difference?  (Read 1600 times)

Posts: 67

« on: October 09, 2009, 06:36:39 PM »

How do I measure the phase difference between two signals at 1296 MHz?

Thanks in Advance.



Posts: 680

« Reply #1 on: October 09, 2009, 07:10:03 PM »

What kind of signals, at what level?  two CW tones at high levels (milliwatts?) or two noisy signals coming from two antennas?

If the signals are fairly clean and high level, a mixer will work as a phase detector. Feed one in the LO port, feed the other in the RF port, measure the voltage at the IF port.  There's some ambiguity (90 lead and 90 lag both give zero volts out)..

if you have a 1296 source to use as an LO, you can mix both signals down to something a lot lower, and put it on an oscilloscope and either look at the sine waves or make a lissajous pattern.

If the signals are low level, then it's tougher...

Posts: 17481

« Reply #2 on: October 09, 2009, 08:18:36 PM »

Are they radiated signals, or signals out of a box?  What
sort of phase difference do you expect to see?

One solution would be to build an open wire line at least
one wavelength long and feed the signals (carefully matched
to the impedance, of course) into each end of it.  Then
take the feedline to your receiver and slide it along the
line until you get a null.  You can calculate the phase
difference by the position of the null.  (In practice you'd
want to swap the sources and repeat the measurement from the
other end to correct for any construction inaccuracies.)
A length of aluminum U channel could be used for the outer
conductor with the inner conductor diameter chosen for the
correct characteristic impedance.

Of course that assumes that the signals are of the
same amplitude to start with.

With very low level signals you can run each through
an FM receiver chip using a common local oscillator,
then feed them both into the same phase detector (after
limiting, so the amplitudes aren't important.)  You'll
get a DC signal out that is proportional to the phase
shift between them.  (The SA605 is better than many of
the Motorola chips in the use because it doesn't have
a built-in capacitor between the two inputs to the
phase detector.)

I can probably think of a few other methods, but perhaps
if you can describe the application they will be more
likely to suit your needs.

Posts: 933

« Reply #3 on: October 11, 2009, 03:20:16 AM »

I would love to know why this question is in the Elmers section. A measurement like this at 1.2 GHz takes some serious kit and at the moment I have no idea how to even attempt it.

Let me give it some thought and if my mind is still blank about this I will ask at work tomorrow.


Posts: 1551


« Reply #4 on: October 11, 2009, 06:17:41 AM »

I would also be interested in just what he's trying to measure.  Without details, it's tough to answer.  And it's going to be tough to accurately measure without some serious test equipment.  As a microwave/RF engineer I've done it many times, but as a ham, an easy way doesn't come to mind.  Details?


Don, K2DC

Posts: 17481

« Reply #5 on: October 11, 2009, 07:53:56 AM »

The measurement also depends on the desired accuracy.

If you are trying to adjust them to be 90 degrees apart
for a phasing receiver, I'd run both signals into a DBM
and adjust for zero DC output.  For a bench measurement
(rather than real-time) I'd do the same thing, but add
a small coax jumper of known length to either side, for
a total of 3 measurements.  From those you should be able
to calculate the actual phase of the original signals,
with an accuracy that depends on cable lengths, signal
levels, etc.

This presumes, of course, that the signals are generated
rather than being received off the air.

Or just run the two signals into your dual-trace 10GHz
oscilloscope and measure the time shift between them. Wink

Posts: 4380

« Reply #6 on: October 11, 2009, 10:34:22 AM »

Some of the measurement techniques described assume that the phase relationship between the two signals will remain at a constant difference.  Such a circumstance is unlikely unless the two signals are derived from a single source.  Signals in the real world vary over time as a result of many variables.
My question is what are you going to do with this phase information after you measure it?  We are talking about fractions of a nanosecond.  I can't think of any amateur use for such measurements.

I wonder about folks who ask questions without giving enough info for anyone to give a useful answer.

Posts: 67

« Reply #7 on: October 12, 2009, 08:40:11 AM »


Let me tell you what I am doing. I need a 1296 MHz solid state 200 Watt power amplifier. A custom made one cost $2600. I have $800 into eight modules that I need to put in parallel. This requires 14 1296 power splitters that must be in phased. I need to test this before I put full power into this PA or it will blow up. Fifty-fifty odds on a mixer is scary to me.

Posts: 21764

« Reply #8 on: October 12, 2009, 11:31:14 AM »

Why would it blow up?

If the signals are not in phase, you'll have reduced gain and power.  Wilkinson dividers may be built with "n" ports, so you wouldn't need 14 (7 splitters, 7 combiners) for 8 stages; this could easily be reduced to four, or possibly even two.


Posts: 680

« Reply #9 on: October 12, 2009, 07:57:42 PM »

Hmm. And you think the phase shift through these amplifiers and combiners will be reasonably stable?

You can do a 8 way divider and an 8 way combiner (heck, you can BUY them from MiniCircuits, although they probably can't handle all the power on the output side, and you'll wind up spending a fair amount in SMA connectors and cabling)

You might look at quadrature hybrids (either lumped or Lange coupler style), or at a distributed amplifier scheme.

But.. to return to your original question.. you've got plenty of power.. An inexpensive mini-circuits double balanced mixer and some pads and a volt meter will do just fine.  Get yourself some transmission line sections to introduce known phase shifts to resolve the ambiguities.

Posts: 17481

« Reply #10 on: October 12, 2009, 09:01:57 PM »

Fifty-fifty odds of blowing up the circuit?  It doesn't
seem that way to me.

Let's look at it this way.  You need to divide the input
signal 8 ways, run each through an amplifier, then combine
them back together in phase.  Presuming all the amps are
identical, that shouldn't be a problem:  just use the same
feed system with identical cable lengths for each amp and
they should be in phase.  Combine the outputs with in-phase
combiners, again all identical and with equal cable lengths.

One advantage of Wilkenson Dividers is that they will
absorb any phase difference in the resistor rather than
passing it to the other port.  A little bit of phase shift
here or there won't be a problem - it will just reduce
your output slightly.  And you can detect any sort of
major problem by monitoring the temperature of the
resistors:  they shouldn't dissipate any power when
the sources are properly matched.

My suggestion would be to build a board with a 2-way
splitter driving a pair of 4-way splitters.  The 4-way
splitters can be built with a bunch of quarter wave
transmission line segments (semi-rigid cable may make
it easier to get all the lines the same length.)  The
2-way can use 75 ohm line.  Then you just need to cable
up the outputs to the individual amps, keeping cables
the same length.  There shouldn't be any uncertainty
in the process so far:  some phase shift due to
construction differences perhaps, but not a phase

If the amps are commercial units they should all be
identical and have the same phase shift.  We'll revisit
this issue later.

On the output side, you can do the same thing.  While
you don't really need Wilkenson dividers on the input,
they will give added security on the output.  Again
I'd suggest two 4-way combiners, then a single 2-way
combiner for the final output at higher power.

Now, perhaps you are afraid that your construction
techniques aren't good enough, or that there is a
transformer wound backwards.  It might also be that
one of the power modules has low gain or a bad transistor.
So the next step is how to test the units as you assemble

Start with the 2-way splitter/combiners and all the
power modules.  Connect the input to one module and
measure the output power at around 10% of the rated
output.  (If you plan to get 12.5 watts out of each,
then 1 watt output probably is a good level.)  Actually,
your first test is to see that all modules give about
the same output when driven by the same input.  This
just requires a power meter.

Once you know they are good, connect two modules in
parallel using the splitter/combiners and measure the
output power of the pair - it should be less than that
for a single module due to the losses in the splitter
and combiner.  You're running at low power here, so
you shouldn't damage anything.  If the units are out
of phase the power will be low and the resistor in the
combiner will get warm.  Use one unit on the right side
and test all the other 7 along with it:  the output
power should remain relatively constant.  Then pick
a different reference unit and repeat the test.  When
you're done you'll know if any of the modules have an
abnormal phase shift (or amplitude).  While you are at
it, monitor the current in each of the modules to see
how it responds to output power - they should all be
fairly close.  Then you can monitor the current in each
of the amps as you go along to see if they are sharing
the drive:  if one is putting out more power than the
others in the group, it should show a higher current.

You can test the 4-way splitter/combiners in the same
way - put them on the input and terminate two ports,
checking all pairs of ports as if it was a 2-way divider
(though it will have 3dB more loss because half the
power is being terminated.)  Any problems in construction
should show up as low output power and/or warming in
the resistor in the output combiner, and this testing
can be done at low power to protect the modules.  By
the time you get one watt out of all 8 modules combined,
each will be running less than half a watt output
(including allowance for losses) which shouldn't have
any serious side effects even into a mis-matched load.

Looking at the problem overall, your need isn't so much
to measure the phase but to make sure that all combinations
have the SAME phase relative to each other, regardless
of what the actual phase is.  One way to test that is
to use the 2-way power splitter and run one branch into
an amplifier module, then through an attenuator to a
double balanced mixer used as a phase detector.  (The
MiniCircuits web site should tell you what polarity
the output voltage will be for their mixers when used
as a phase detector.)  Run a cable from the other port
of the divider directly to the other port on the phase
detector.  If you change the length of this cable (or
any other in the setup) it should change the voltage
measured on the phase detector.  I'd look for something
that is around 1/4 of the highest voltage that you
measure with different cables.  Then swap out all the
different amplifier modules and make sure you get
about the same phase detector voltage.   You don't have
to know what the phase shift is through all the cables
and equipment, just that all the modules give you the
same result.  Then you know they are matched.

Now, what if they aren't consistent?  That is possible
if they require hand-tuning to frequency.  In that case
start by tuning up each one individually for maximum
power output (or whatever other criterion you want.)
Let's say that they all give different results on the
phase check test.  Take the one closest to the center
of the distribution and connect it to the phase detector
though a power meter.  Re-adjust the cable lengths to
get a reasonable phase detector reading.  Then swap
in each of the other amps in turn and adjust it for the
same output power and phase detector output.  (If you
use the lowest gain amp as the reference, you can add
parts to the input of the other amps to reduce the gain
and shift the phase as needed.)  When you're done you
know that all the amplifier modules are in phase.

Posts: 631


« Reply #11 on: October 13, 2009, 01:16:38 AM »

Seems to me that a network analyzer with an adjustable section in each amplifer would be the easiest way to tune this proposal up.....


Carl - W9PMZ
Pages: [1]   Go Up
Jump to:  

Powered by MySQL Powered by PHP Powered by SMF 1.1.11 | SMF © 2006-2009, Simple Machines LLC Valid XHTML 1.0! Valid CSS!