University of Toronto, Physics, Quantum Optics

Alan Stummer, Research Lab Technologist

RF Phase Shifter

Send your feedback! Feedback Request
I am curious who uses what.  Are these webpages a waste of time, or are they any help to others?  Are the circuits, software and utilities appearing in other labs?  Please send your comments or suggestions or what you have used (or not) or schematics of your version or pictures or anything!   Email me, or be creative and send a postcard!  I want to hear from the vacuum!

NOTICE: This webpage and associated files is provided for reference only.  This is not a kit site!  It is a collection of my work here at the University of Toronto in the Physics department. If you are considering using any schematics, designs, or anything else from here then be warned that you had better know something of what you are about to do.  No design is guaranteed in any way, including workable schematic, board layout, HDL code, embedded software, user software, component selection, documentation, webpages, or anything.

All that said, if it says here it works then for me it worked. To make the project work may have involved undocumented additions, changes, deletions, tweaks, tunings, alterations, modifications, adjustments, waving of a wand while wearing a pointy black hat, appeals to electron deities and just plain doing whatever it takes to make the project work.


For  Marcelo Martinelli in Aephram's lab, started Nov 2003.  The RF Phase Shifter is a method of shifting the relative phase of two 70MHz signals.  They are used to drive two AOMs (Acousto-Optical modulators) which in turn modulate the two halves of a split laser light.  These modulated beams are recombined in an interference pattern.  The phase shift, which occurs in <20nS, relocates the nodes of the interference pattern.

Modus Operandi

Refering to the schematic ( schematic in native Eagle or as as PDF), the 10MHz reference signal is sent in, either a sine or square wave. The lowpass filter FIL3 removes all but the fundamental frequency, leaving a sine wave.  This signal goes to three places: the reference comparator and the reference amplifier and the phase shifted amplifier.

Reference and Phase Shifted Signals  Part of the 10MHz sine is attenuated [with zero phase shift] via R10 and R9 then amplified by the VGA (Variable Gain Amplifier) IC2. At the same time, the 10MHz sine is attenuated via R15 and C13 then amplified by the VGA IC3.  The impedance of C13 at 10MHz is the same as the resistance of R9, resulting in more or less the same amplitude presented to the VGAs.  However, the R15 & C13 combination gives a phase shift of 83°.  The result is now two equal strength signals at the two VGA inputs, apart by 83° at 10MHz.  these are sometimes called quadrature signals.

VGAs Each VGA has two gain controls: a dB-linear (A.K.A. logarithmic) control called Vdbs and a true linear control Vmag.  For each VGA, the output signal is the input signal amplified by these two gains.  The logarithmic gains are manually set by the trimpots R6 and R11.  These calibrate out inherent errors such as differences in input signal amplitudes and VGA gains.
The two linear gains are complimentary, such that as one is rising, the other is falling.  The input phase control voltage, Vphase, is reflected around +2.5V by the opamps IC1, such that the sum of the two outputs is always +5V.  Specifically, IC1A has a gain of +1 and IC1B has a gain of -1 normalized to +2.5V.  The result is that the output of  IC2 is proportional to Vphase and the output of IC3 is inversely proportional to Vphase.

Variable Phase Shifted Comparator   The two complimentary outputs from two VGAs are averaged.  This is where the phase shifting actually takes place.  In one extreme where Vphase = +5V, IC2 (the reference 10MHz signal) is at full strength and IC3 (the phase shifted 10MHz signal) is shutdown.  Looking at the true outputs (VGA.13, meaning pin 13), the average signal on comparator IC4B.3 is one half of the IC2 reference signal due to the divider R13 & R16.  More importantly, the phase of this signal is the reference phase of 0°.  The complimentary outputs (VGAs.12) divide down via R17 & R18 to a signal at a phase of 180°.  The comparator output indicates the relative polarity of its inputs, therefore the output toggles on the zero crossings.
In the other extreme case where Vphase = 0V, the VGA IC2 (reference 10MHz) is shutdown but the VGA IC3 (phase shifted 10MHz) is at full strength.  Now the comparator IC4B has inputs at 83° and 263°, so its output is toggling at a phase of 83° relative to the reference.

Now take a middle example where Vphase = +2.5V.  Both VGAs have the same gain and therefore the same output amplitudes, however not in phase.  The average of the 0° signal and the 83° signal is one at 41.5° and 70% (is it 1/sqrt 2?) of the amplitude. Because the comparator IC4B detects only relative polarity and not amplitude (within reason), this 30% drop in amplitude is unimportant.  Its output toggles at 41.5° relative to the reference.

This phase can now be linearly adjusted anywhere between the two extremes of  0° and 83° by setting Vphase between +5V and 0V.  The slew rate (the rate at which the output can change, or dV/dt) of IC1 is >100V/uS so it can change from one extreme to the other in <5nS.  This means that the 10MHz phase can be changed at up to 83°/5nS or 17°/nS.

Variable Phase Shifted 70MHz The comparator IC4B turns the 10MHz phase shifted sine wave into a (approx) 4V p-p (peak to peak) square wave with 5nS edges and therefore strong harmonic content.  A true square wave has only odd harmonics with each harmonic strength equal to the reciprocal of the harmonic number.  Therefore, the 7 th harmonic amplitude is 1/7 th or 14% of the normalized fundamantal value.  The comparator IC4B's output is sent through the 70MHz bandpass filter FIL2, which removes all but the 7th harmonic at 70MHz.  Unfortunately, the filter is far from perfect.  It is not a true bandpass filter such as can be obtained in optical filters but rather a combination of imperfect bandpass, high pass and low pass filters.  This means that not only does it disspates power (lowers the Q) but it can better be defined as differentiating the square wave and then removing the lower harmonics.  The effect is that the 70MHz envelope decays slightly after each 10MHz edge but is synchronously boosted on the next edge.  This quirky 70MHz is output from the board as the phase shifted signal.  Note that amplifying this 70MHz signal into saturation will remove the envelope.

Reference Comparator The comparator IC4A monitors the incoming 10MHz but after the 10MHz filter, at a constant 0° phase shift.  In exactly the same way as the comparator IC4B converts its inputs to square waves, so does comparator IC4A, producing 70MHz at 0°.

Summary A phase shift range of 83° at 10MHz corresponds to a time shift of 23nS.  Because the 70MHz signal is a product of the 10MHz signal, it too is time shifted by 23nS.  However, at 70MHz that corresponds to 581° or 1.6 cycles.  Only pi or 2pi of shifting is needed so the excess capability is not used.


The board must be recalibrated whenever a new 10MHz source is used.  Put in the 10MHz signal to be used, disconnect Vphase or preferably set it to +2.5V (it will default to about +2.5V or equal VGA gains when disconnected).  Monitor test point "VGA-Shift" and adjust the trimpot R11 for approximately 0.25Vrms then monitor test point "VGA-Ref" and adjust trimpot R6 for approximately 0.25Vrms too.  The important parameter is the relative amplitudes - both can be set lower or higher.  Too low and the comparators IC4 will pick up noise and cause jitter, too high and the VGA outputs saturate.

The effect of this calibration is to straighten the transfer curve of Vphase to phase shift.  If the calibration is off, possibly if a different 10MHz source was used without recalibration, the transfer curve will become distinctly "S" shaped.  If this is not a problem then calibration is not important.

Return to homepage
Sorry, no more chance for asking direct questions, queries, broken links, problems, flak, slings, arrows, kudos, criticism, comments, brickbats, corrections or suggestions. Made with Nvu