Phase Shifters

phase shifter patent3568105

Click here to learn the tradeoff between TDUs and phase shifters (new for May 2019!)

The figure above shows a microstrip phase shifter, circa 1969, invented by Robert Felsenheld.  While others were at Woodstock skinny-dipping and watching Canned Heat's Alan Wilson croak out "Going up the Country", Felsenheld was working at ITT in Nutley, New Jersey and writing up this patent.  What's left of ITT Nutley is now part of Harris, if you like to keep track of old nameplates.  The phase shifter is a four-bit digital unit, using switched series lines.  It probably provides more of a true time delay than a constant phase shift over frequency, but you would have to analyze what the loading effects of path ACB are on the main path when diodes CR1 and CR8 are switched in. One funny thing about the image is that the draftsman did not understand that capacitors C6, C7, C8 and C9 must be in shunt with the diode DC bias lines, not in series. These capacitors need to be grounded to form RF grounds that choke off the signal from escaping out the control lines.  At the time there were no surface mount parts, so axial-leaded capacitors (perhaps 10 or 100 pF) would have one lead soldered to the trace, and the other going through a drilled hole in the board and soldered to the backside (with leads cut as short as possible!) It would also have been possible to make an RF ground with a quarter-wave open circuit stub to avoid having to drill a hole. Thanks to William for asking how the capacitors might be connected.  Felsenheld was a graduate of Lafayette College, and went on to form a manufacturer's rep company;  he passed away in 2015. Alan Wilson (Canned Heat) died in 1970 at the age of 27 after overdosing on barbiturates.  Engineers generally live a lot longer than musicians, which allows the IEEE to offer lower life insurance premiums to members.  Hey, that almost sounds like they are paying us for an advertisement!

Below is an outline of all of our material on phase shifters. This collection of material is easily the best free phase shifter resource on the planet, and it keeps getting better.

Check out the Phase Shifter Search Tool (on everything RF) if you are looking to buy a phase shifter.

On this page we have:

Phase shifter background

Digital versus analog phase shifters

Active versus passive phase shifters 

MEMS phase shifters (proceed with caution...)

Other Microwaves101 phase shifter pages include:

An important characteristic of phase shifters (We recommend that you read this first!)

Types of phase shifters

Switched line (delay line) phase shifters

Switched filter phase shifters

High-pass/low-pass phase shifters

Spreadsheet for high pass/low pass phase shifter calculations

Switched filter phase shifters (the so-called Campbell-Brown bit)

Loaded-line phase shifters

Ferroelectric phase shifters

Reflection phase shifters

180 degree hybrid phase shifters (such as rat-races)

Quadrature hybrid phase shifters

Varactor phase shifters

180 reflection bits using PIN diodes

Reflection phase shifters using circulators (coming soon)

Schiffman phase shifters

MEMS phase shifters

Related to phase shifters

True time delay

Time delay units (TDUs)

Vector modulators

RMS error calculations

RMS amplitude and phase errors (covers attenuators and phase shifters)

Phase shifter RMS phase error

Phase shifter RMS amplitude error

Phase shifter design examples

Multi-state phase shifter simulations in Agilent's ADS

Multi-state phase shifter simulations part II

Phase shifter data example 1

MMIC phase shifter examples

Example 1: AMTL six bit S-band phase shifter

Example 2: Marconi six bit C-band phase shifter

Example 3: Hittite HMC543 X-band phase shifter

Power handling example: Qorvo TGP1439

Applications of phase shifters

Frequency translators

Phased arrays

SSPAs (see phase error section)

Residual phase noise measurement (coming soon)

Some phase shifter background

Phase shifters are used to change the transmission phase angle (phase of S21) of a two-port network and they have four important characteristics. The first is insertion loss (or gain). Ideally, phase shifters provide low insertion loss in all phase states. While the loss of a phase shifter is often overcome using an amplifier stage, lower insertion loss phase shifters require less amplification and lower power to overcome the losses. The second important characteristic is phase shifters have equal amplitude for all phase states. Many systems using phase shifters must not experience amplitude changes in signal level as phase states are changed. The third important characteristic is that most phase shifters are reciprocal networks. This means they work effectively on signals passing through them in either direction. These three characteristics are used to describe the electrical performance of phase shifters. The fourth characteristic is whether they provide flat phase versus frequency, or true time delay.  Read more about this  important characteristic of phase shifters.

Other important characteristics of any microwave device are its usable bandwidth, and its power handling; we didn't call them out as they are not specific to phase shifters. Read about one case study on phase shifter power handling here.

Phase shifters can be controlled electrically, magnetically or mechanically. Most of the phase shifters described on this web site are electronically-controlled, passive reciprocal networks

While the applications of microwave phase shifters are numerous, perhaps the most important application is within a phased array antenna system (a.k.a. electrically steerable array, or ESA). In these systems, the phases of a large number of antenna elements are controlled to force the electromagnetic wave to add up at a particular angle to the array. For this very purpose, phase shifters are often embedded in TR modules. The total phase variation of a phase shifter need only be 360 degrees to control an ESA of moderate bandwidth. Networks that stretch phase more than 360 degrees are often called time delay bits or true time delays (part of a TDU), and are constructed similar to the switched line phase shifters that are described below.

Until recently, digital phase shifter was typically a GaAs MMIC comprised of passive reciprocal networks.  Today, vector modulator circuits designed in silicon are giving GaAs some competition. Vector modulators are may or may not be passive reciprocal networks, but in silicon implementation, their is no cost associated with adding amplifiers (which would make the vector modulator active and non-reciprocal).

Analog versus digital phase shifters

When we say "digital" phase shifter, we are still talking about analog electronics.  Don't get confused thinking that you can send a stream of ones and zeros down data bus and control their phase and accomplish something. Digital in the case of phase shifters means two-state devices, where the states have different insertion phases at microwave frequencies.

Phase shifters can be controlled using analog signals or digital bits. Analog phase shifters provide a continuously variable phase, most often controlled by a voltage. Electrically controlled analog phase shifters can be realized with varactor diodes that change capacitance with voltage, or nonlinear dielectrics such as barium strontium titanate, or ferro-electric materials such as yttrium iron garnet. A mechanically-controlled analog phase shifter is really just a mechanically lengthened transmission line, often called a trombone line. Analog phase shifters have been used in radar systems and more recently to down-tilt (steer) antennas used in cellular base stations. 

Most phase shifters are of the digitally controlled  variety because they are more immune to noise on their voltage control lines. Digital phase shifters provide a discrete set of phase states that are controlled by two-state "phase bits." The highest order bit is 180 degrees, the next highest is 90 degrees, then 45 degrees, etc., as 360 degrees is divided into smaller and smaller binary steps. A three bit phase shifter would have a 45 degree least significant bit (LSB), while a six bit phase shifter would have a 5.6 degree least significant bit. Technically the latter case would have a 5.625 degree LSB, but in the microwave world it is best to ignore precision that you cannot obtain. If you can't comprehend this point, you might want to consider a different career such as accounting.

Someone asked us about the relative advantages of analog and digital phase shifters.  We provide a list below, the two technologies we are considering are discrete varactor diode (analog) and MMIC (digitally controlled). We don't have a reference for this information, it is based on personal observation.  Feel free to use Microwaves101 as THE reference for phase shifter trade studies.

Analog phase shifter advantages

  • Lower loss
  • Lower cost of parts (but sensitive to assembly variations)

Digital phase shifter advantages

  • Immunity to noise on control lines
  • More uniform performance, unit-to-unit
  • Ability to achieve flat phase over wide bandwidth
  • Less susceptible to phase pulling when embedded in networks that are not perfectly impedance-matched
  • Easier to assemble
  • Potentially higher power handling and linearity

The convention followed for phase shifters is that the shortest phase length is the reference or "off" state, and the longest path or phase length is the "on" state. Thus a 90 degree phase shifter actually provides minus ninety degrees of phase shift in its "on" state.

Active versus passive phase shifters - historical perspective

One of the first phase shifter MMIC designs used dual-gate FETs in a way that it was non-reciprocal. You can learn of it here:

Vorhaus, J. L. et al, "Monolithic Dual-Gate GaAs FET Digital Phase Shifter", IEEE transactions on Microwave Theory and Techniques, Vol. MTT-30, No. 7, July 1982.

With an active phase shifter, it might be possible to counteract the loss of the phase shift elements and eliminate an amplifier stage. Seems worth exploring, right?

In the 1980s the idea of an active phase shifter quickly fell beside the wayside, because a passive reciprocal phase shifter is more versatile and requires fewer SPDT switches to route the transmit and receive signals through the phase shifter.

It is funny how over the past ten years, more and more TR module architectures are using a "common leg" circuit that includes phase control. Because of this, the one advantage of a passive phase shifter no longer matters. It seems just a matter of time before someone reinvents the active phase shifter in a TR module. Maybe that has happened already, just send us a reference and we'll post it here.

An important note on active phase shifters... you must consider both the gain and the noise figure of your phase shifter when you are analyzing the performance of the next higher assembly. Vorhaus' paper did not report the noise figure of his 1982 active phase shifter.

MEMS phase shifters

Since the beginning of RF MEMS, there has been about a billion dollars spent to try to develop a phase shifter with low loss such that a PESA is enabled and the inventors make a killing. In the beginning, it seemed possible that a three-bit Ka-band phase shifter could be made with 1 dB loss. Then the dark ages of MEMS began, when reliability was uncovered to be a major problem, and hence the MEMS Tree of Woe was born and all involved took the acronym off of their business cards. More recently, promises have been reduced to maybe 2.5 dB loss for three bits at Ka-band, 2 dB at X-band. This level of loss spells "game over" for MEMS phase shifters.

We'll keep an eye on the technology, and one of these days make a page of content about it.

Other

If you know of any phase shifter topologies not covered on one of our phase shifter pages that should be described here, drop us a line and we will add your knowledge to this chapter! Want to donate a photo of a phase shifter you designed? Send it our way and we might get your 15 minutes of fame if we decide to put it on this page

Check out what Wikipedia offers on phase shifters, it's slim pickings... they do link back to us, but don't consider us as a "reference". Someone ought to teach those eggheads some manners...