Understanding RF/Microwave Solid State Switches and Their Applications

Understanding RF/Microwave Solid State Switches and Their Applications

Understanding RF/Microwave

Solid State Switches and their Applications

Application Note

Table of Contents

1.0 Introduction

........................................................................................... 3

2.0 Types of Solid State Switches

....................................................... 4

2.1 PIN diode switches

................................................................... 4

switches

........................................................................... 10

2.3 Hybrid

...................................................................... 13

3.0 Solid State Switch Specifi cations

............................................ 14

3.1 Operating frequency range

................................................... 14

3.2 Isolation

.................................................................................... 15

3.3 Insertion

........................................................................... 16

3.4 Return loss and VSWR

.......................................................... 17

3.5 Switching

..................................................................... 18

3.6 Settling

............................................................................. 21

3.7 Video

.......................................................................... 25

3.8 Off-port

............................................................... 27

3.9 Phase

........................................................................ 27

3.10 Harmonics and intermodulation distortion

3.11 1 dB compression point

....................... 28

......................................................... 30

4.0 Applications

....................................................................................... 31

4.1

4.2

Application 1 Mobile handset: power amplifi er testing

... 31

Application 2 Signal routing: multiple-instrument,

multiple-DUT

............................................................... 33

4.3 Application 3 Filter bank: SAW fi lter testing

.................... 35

4.4

4.5

Application 4 Satellite: testing channel amplifi ers with ALC systems

................................................................... 37

Application 5 Base station and satellite: antenna testing

... 38

5.0 Conclusion

.......................................................................................... 39

6.0 Reference

............................................................................................ 39

2

1.0 Introduction

RF and microwave switches are used extensively in microwave test systems for signal routing between instruments and devices under test (DUT). Incorporating a switch into a switch matrix system enables you to route signals from multiple instruments to single or multiple DUTs. This allows multiple tests to be performed with the same setup, eliminating the need for frequent connects and disconnects. The entire testing process can be automated, increasing the throughput in high-volume production environments.

Introduced in the 1960s, the term “solid state” describes electronic devices which contain neither vacuum tubes nor mechanical devices such as relays. In solid state components, electrons fl ow through unheated solid semiconductor materials, Germanium or Silicon being the most well known, instead of fl owing through a heated vacuum as in vacuum tubes. As vacuum tubes were rapidly superseded by transistors, the term “solid state” generally refers to devices that

“contain no moving parts”.

RF and microwave switches can be categorized into two equally mainstream, and essential groups:

ii) Solid state switches

Currently, the micro-electromechanical-systems (MEMS) switch technology is emerging. However, historically the switch market has been dominated by the mature technologies of electromechanical and solid state switches. MEMS switch technology has to overcome problems such as reliability and quality packaging. These switches will need to provide convincing, if not far superior performance, over the two established families of switches, before achieving higher market adoption.

Solid state switches are more reliable and exhibit a longer lifetime than their electromechanical counterparts due to their superior resistance to shock, vibration and mechanical wear. They also offer a faster switching time. However, the higher innate ON resistance of the solid state switches gives them higher insertion loss than electromechanical switches. Therefore solid state switches are preferred in systems where fast switching and long lifetime are essential.

Browsing through the internet would give you some basic information on solid state switches, with minimal defi nition such as: solid state electrical switches operate without moving parts. This application note will deliver a more comprehensive understanding of solid state switches including: the types of solid state switches, the technologies used to design them, the important parameters to consider when selecting a solid state switch and how to measure them. A comparison of the different types of solid state switches is made as selection parameters are discussed. Finally, a number of application examples are given to illustrate the signifi cance of the parameters and the role of solid state switches in today’s demanding test and measurement systems.

3

2.0 Types of Solid State Switches

There are three main types of solid state switches:

• PIN diodes switches

• Field-effect transistors (FET) switches

• Hybrid (FET and PIN diode) switches

2.1 Pin diode switches

Pin diode

A PIN diode is a semiconductor device that operates as a variable resistor at RF and microwave frequencies. Its resistance value varies from less than 1Ω (ON) to more than 10kΩ (OFF) depending on the amount of current fl owing through it. As a current-controlled device, the resistance is determined only by the forward biased DC current. When the control current is switched on and off, the

PIN diode can be used for switching. An important feature of the PIN diode in switching applications is its ability to control large RF signals while using much smaller levels of DC excitation.

Forward biased

The PIN diode structure consists of an intrinsic (I) layer with very high resistivity material sandwiched between regions of highly doped positively (P) charged and negatively (N) charged material as shown in Figure 1a. When the PIN diode is forward biased, positive charges from the P region and negative charges from the N region are injected into the I layer. These charges do not recombine immediately. Instead, a fi nite quantity of charge always remains stored and results in a lowering of the I-region resistance. The resistance of the PIN diode under forward bias, R s

, is inversely proportional to the total forward bias current,

I f

., making the PIN diode perfect for achieving excellent isolation at high frequencies. Figure 2 shows a typical R s

vs. I f

graph of a PIN diode. Figure 1(b) shows the equivalent circuit with the PIN diode forward biased. L is the parasitic inductance which depends on the geometrical properties of the package.

4

Reverse biased

With reverse or zero biasing, the I layer is depleted of charges and the PIN diode exhibits very high resistance (R

P

) as shown in Figure 1(c). C

T

is the total capacitance of the PIN diode which is the sum of the diode junction capacitance

(C

J

) and the parasitic capacitance (C

P

) of the package. This capacitance limits the switch’s performance at high frequencies in the form of insertion loss and isolation degradation. A low capacitance diode helps to improve performance at higher frequencies.

P

Metal pin

Glass

Intrinsic layer

L

R s

N

(a) Cross section of a basic diode (b) Forward bias

Figure 1. PIN diode switches

C

T

Rp

(c) Reverse bias

10,000

1000

100

10

1

0.01

0.1

1

I

F

– forward bias current (mA)

10

Figure 2. Typical R s

vs. If graph of a PIN diode

100

PIN diode characteristics, such as: high switching speed, low package parasitic reactance and small physical size compared to a signal wavelength, make them ideal for use in broadband switch design. The performance of PIN diode primarily depends on chip geometry and the nature of the semiconductor material used in the fi nished diode, particularly in the I region.

5

Lower frequency limitation of PIN diode switches

Forward biased

The key limitation of a PIN diode switch is its lower frequency limit. The forward biased PIN diode is a current controlled resistor when forward biased. The resistance of the I region under forward bias, R s

is inversely proportional to the stored charge Q = I

ƒ

τ

where I

ƒ

is the forward current and

τ

is the recombination time or carrier lifetime.

R

S

=

(μ p

W

2

+ μ n

=

W

)Q (μ p

+ μ

2 n

)I

ƒ

τ ohms where

W = I-region width

μ

μ p

= hole mobility n

= electron mobility

This equation is valid for frequencies higher than the transit time frequency of the I region, defi ned as f transit frequencies less than f transit

= 1300/W 2 (f in MHz and W in microns). At

, the PIN diode acts like a PN-junction diode and rectifi es the RF signal, making the PIN diode unsuitable for use at these frequencies. f transit

typically ranges from a few kHz to 1 MHz.

Reverse biased

The reverse bias equivalent circuit consists of the PIN diode capacitance (C

T shunt loss element, (Rp), and the parasitic inductance (Ls). C

T tance consisting of the junction capacitance C

J

and package capacitance C

P

), a

is the total capaci-

C

J = e

A

W which is valid for frequencies above the dielectric relaxation frequency of the I region.

ƒ d =

2

1

π

ρ e e

= dielectric constant of I-region material

A = diode junction area

W = I-region thickness

ρ = resistivity of I region

At frequencies much lower than

ƒ d

, the capacitance characteristic of the PIN diode resembles a varactor diode. Changes and variations in the capacitance affect the usefulness of PIN diode as a switch at lower frequencies.

This is why PIN diode switches have low frequency limitations.

6

PIN diode switch design

PIN diodes are often used to design a switch that controls the path of RF signals. Depending on the performance requirements, the switch can consist of all series diodes, all shunt diodes or a combination of series and shunt diodes.

Bias 1

RF port 1

C

L

D

1

RF port 3

D

2

L

C

Bias 2

RF port 2

L

Common junction

Figure 3a. Series pin SPDT switch[1]

Bias 1

RF port 3

RF port 1

C

L

C

Z o

λ /4

D

1

Bias 2

Z o

λ /4

C

L

C

RF port 2

D

2

Figure 3b. Shunt pin SPDT switch

Bias 1

RF port 1

C

L

D

3

D

1

L

RF port 3

D

4

Bias 2

L

C

RF port 2

D

2

Figure 3c. Series-shunt pin SPDT switch

Series PIN diodes (Figure 3a) are capable of functioning within a wide bandwidth, which is limited by the biasing inductors and DC blocking capacitors.

Also, when PIN diodes are reverse biased, parasitic capacitance gives rise to poor isolation at microwave frequencies, with a 6 dB per octave roll-off as a function of frequency.

Shunt PIN diodes (Figure 3b) feature high isolation relatively independent of frequency. To turn a switch on, PIN diodes are reversed, and this means a dominant reverse biased capacitance exists. Circuit designers commonly use a circuit transmission line to create series lumped inductance to achieve a low pass fi lter effect which enables the switch to work up to the desired frequency.

7

However, shunt diodes switches have limited bandwidth arising from the use of quarter wavelength transmission lines between the common junction and each shunt diode. At the midband frequency f o

, where the transmission lines are λ/4 in length, the switch operates as follows: when diode D1 is forward biased and diode D2 is reverse biased, the RF signal fl ows from port 3 to port 2, and the RF port 1 will be isolated. The λ/4 line will transform the short circuit at D1 into an open circuit at the common junction, eliminating any reactive loading at that point. However, as the frequency is changed from f o

, the transmission lines will change in electrical length, creating a mismatch at the common junction.

To illustrate this, Figure 4 shows a model of two shunt diodes. The PIN diodes are turned on (assume R

ON

= 1.2 ohms) to shunt the RF signal to ground. In

Figure 4a, there is no spacing between the diodes (for explanatory purpose only, it is of course impossible to have zero physical length in between diodes). The isolation is around 32 dB as Figure 5 shows (red trace).

In Figure 4b, a physical spacing of 116 mils (quarter wavelength of 10 GHz at alumina substrate with K effective

= 6.466) will result in a maximum isolation at 10

GHz. This is shown in Figure 5 as the blue trace. The λ/4 spacing of 116 mils has transformed the short circuit of the diodes into an open circuit.

λ

=

c

effective

.4.10 x 10 9

= 116 mils where c = 3 x 10

8

m/s is the speed of light.

Multiple shunt PIN diodes separated at different midband frequencies can be used to achieve high isolation across a broadband frequency range. This is illustrated in Figure 4c where two additional shunt diodes are spaced at quarter wavelength of 20 GHz (58 mils).

Figure 4a. PIN diodes with no spacing between diodes

Figure 4b. Quarter wavelength spacing of shunt PIN diodes

Figure 4c. Quarter wavelength spacing of shunt PIN diodes with two additional diodes

8

As can be seen from the pink trace in Figure 5, the design achieves good isolation up to 20 GHz. Isolation at 20 GHz has improved from around 30 dB

(blue trace) to 100 dB (pink trace). However, with more shunt diodes, there will be some degradation in insertion loss and as well as higher current and power consumption.

-20

-80

-100

-40

-60

Maximum isolation at 10 GHz

-120

-140

0 5 10 15

Improved isolation at 20 GHz

35 40 20 25 freq, GHz

30 45

Figure 5. Isolation performance of different diode spacing and quantity

Circuit designers often resort to using a combination of both series and shunt diodes to achieve optimal insertion loss and isolation performance in a PIN diode switch (Figure 3c). However, this switch is more complicated to design and consumes higher biasing current and power compared to series or shunt

PIN diode switches.

Notice that in a PIN diode switch design, the biasing path is connected to the

RF path of the switch design. This is unavoidable when designing a PIN diode switch with series PIN diode. DC blocking capacitors are also needed at the RF ports. This will degrade the insertion loss performance of the PIN diode switch at low frequencies due to the high pass fi lter effect of the capacitor, and at high frequencies due to transmission loss through the capacitor.

RF chokes (inductors) are used along the biasing paths to avoid RF signal leakage while allowing the DC signal to fl ow through to bias the diodes. The

RF choke must have adequately high impedance at low frequencies so that the

RF signal will not leak through the biasing path leading to higher insertion loss.

At the same time, the RF choke should have a high self resonant frequency to enable broadband switch design.

For power consumption, PIN diode switches with current-controlled PIN diodes always consume higher current than FET switches.

9

2.2 FET switches

Field-effect transistors (FET)

Field-effect transistors (FETs) are a semiconductor device which depends on an electric fi eld to control the shape and hence the conductivity of a channel in the semiconductor material. FETs control the current between source and drain connections by a voltage applied between the gate and source. FET switches are very stable and repeatable due to good control of the drain-to-source resistance

(R

DS

).

Figure 6a shows the low channel resistance between drain and source (R

R

DS

=

ON

) that occurs when the FET is biased ON by the switching voltage. Low channel resistance enables these switches to operate at low frequencies (down to DC).

The application of a reverse-biasing voltage between gate and source causes the depletion region at that junction to expand, thereby “pinching off” the channel between source and drain through which the controlled current travels.

In the off state, as shown in Figure 6b, the conduction channel is depleted

(pinched off), which causes the FET to exhibit very high resistance (R

OFF

). This mechanism ensures that FET switches provides excellent isolation at low frequencies.[1]

Source

R (on)

V > 0 volts

Gate

Drain n

+

GaAs n-type GaAs

Depletion layer

“Channel”

Semi-insulating GaAs substrate

Figure 6a. Low channel resistance

Source n

+ n-type GaAs

R (off)

Semi-insulating GaAs substrate

V < – V

p

Gate

Depletion layer

No “Channel”

Drain

Figure 6b. High channel resistance

10

D

G

C

DS

= 0.05 pF

S

Figure 6c. Schematic of a FET

The isolation of FET switches degrades at higher frequencies due to the effect of drain-to-source capacitance (C

DS

). Figure 6(c) shows a Gallium Arsenide (GaAs) metal-semiconductor fi eld-effect transistor (MESFET) schematic, with the equation below showing the drain-to-source impedance as equal to 320 Ω at 10

GHz. This is equivalent to an isolation of only 10.5 dB between the drain and the source, which is clearly not suffi cient to satisfy isolation performance.

Assume C

DS

= 0.05 pF ; f = 10 GHz

│X c

│ =

j

1

ω

│ = │

1

│ ≈ 320Ω

C j2

π

ƒ

C

11

FET switches

Figure 7 shows a simplifi ed schematic diagram of a SPDT switch using FETs. To switch the RF from the common port to RF port 2, FET1 and FET4 are reversed biased so that the channels between source and drain are pinched off; while

FET2 and FET3 are forward biased so that low channel resistance exists between drain and source. FET1 and FET3 act as a series device to switch the RF on and off, while FET3 and FET4 are used to shunt away RF that leaks to the off port for better isolation.

Notice that the biasing path is not connected to the RF path of the switch design. This gives FET switch a simpler DC biasing path, eliminating the need for an expensive high-performance RF choke. These chokes are used to reduce the insertion loss that results from the biasing path being connected to the RF path in PIN diode switches. Nevertheless, the ON resistance of a FET is typically higher than a PIN diode, resulting in inferior insertion loss performance for FET switches compared to PIN diode switches.

As voltage-controlled devices, FET switches consume much less current than

PIN diode switches.

Figure 7. Simplifi ed SPDT switch using FETs as switching devices

12

2.3 Hybrid switches

It is clear that both PIN diodes and FETs provide distinctive advantages. However, neither exhibits superior bandwidth and isolation requirements at the same time. So, hybrid switches using FET and PIN diode technology were created to provide wide bandwidth and high performance switching.

Hybrid switches use series FETs to extend the frequency response down to DC with excellent isolation and shunt PIN diodes at λ/4 spacing to provide good isolation at high frequencies. Utilization of series FETs instead of PIN diodes also provides better repeatability performance due to the well controlled R

DS

ON.

Hybrid switches contain shunt PIN diodes and thus draw a certain amount of current. Nevertheless, the RF path and biasing path should not converge; that might lead to RF leakage in the DC biasing path as with PIN diodes.

The previous sections discussed the different types of solid state switches.

Table 1 summarizes the general performance of the different types of solid state switches; a surface comparison with electromechanical mechanical switches is shown as well.

Table 1. General performance for different types of solid state switches

Frequency range

Solid state switches

Pin diode

from MHz

FET

from DC

Hybrid

from kHz

Electromechanical switches

from DC

Insertion loss

Medium High High

(Roll off at low (Roll off at high (Roll off at high frequencies) frequencies) frequencies)

Agilent switch examples

P9400/2/4

85331 /2

U9397, U9400

Low

Isolation

Good at high frequencies

Good at low Good at high Good across broad

Repeatability

Excellent Excellent Good

Switching speed

Fast Average

Power handling

Low Low Low High

Operating life

High High

Power

Moderate

Current interrupt

consumption

reduces

Sensitive to

RF power overstress,

RF power RF power

Vibration

L series, 8710x/20x

13

3.0 Switch Specifi cations

3.1 Operating frequency range

As consumer technologies push the use of the frequency spectrums to their limits, broadband switches undeniably have an upper hand over narrowband designs. Test systems with broadband accessories increase test system fl exibility by extending frequency coverage. This reduces costs and saves resources.

Consider this case, you are currently running a test system that tests GSM and

UMTS products at 2.2 GHz now want to test WLAN products at 5 GHz. In this case, you would need to buy equipment to extend the frequency of the system to

5 GHz or more. Equipment such as network analyzers which commonly run on a broadband basis, are less likely to be the limitation. On the other hand, accessories such as switches might be a constraint. Agilent technologies offers a series of broadband accessories. By early adoption of broadband accessories, you can save the time and cost required to extend the capability and frequency range of your test systems.

Agilent Technologies has standard broadband solid state switches operating from

300 kHz to 50 GHz. The switches are carefully designed to meet broadband requirements while accomplishing excellent RF performance. The 85331B/32B

PIN diode switches operate from 45 MHz to 50 GHz with excellent RF performance while U9397, U9400 hybrid FET switches offer astounding RF performance from 300 kHz to 18 GHz. The PIN diode families of P9400, P9402 and P9404, offer outstanding performance from 100 MHz to 18 GHz.

Table 2. Frequency range comparison of solid state switches

Lower frequency limit

*

Upper frequency limit

*

PIN diode switches

100 MHz

> 50 GHz

*Typical. Based on performance requirements.

FET switches

DC

> 20 GHz

Hybrid switches

300 kHz

> 20 GHz

14

3.2 Isolation

Isolation is defi ned as the ratio of the power level when the switch’s path is

“off” to the power level when the switch is “on”. In other words, it is the suppression of a signal in excess of the insertion loss at the “off” port.

High isolation in switches is crucial in most measurement applications. Good isolation prevents stray signals from leaking into the desired signal path.

High isolation is especially critical in measurement systems where signals are consistently being routed to and from a variety of sources and receivers through various switch test ports. If these stray signals are allowed to get through, measurement integrity is severely compromised.

PIN diode switches typically offer better isolation performance at high frequencies than FET switches due to the FETs’ drain-source capacitance when the FET is turned off. However, this disadvantage can be resolved with the hybrid design by using shunt PIN diodes for isolation.

PIN diode switch isolation is poorer than FET switches at lower frequencies, around tens of MHz, due the low frequency limitations of PIN diodes. From the point of view of a circuit designer, this can be improved by proper selection of the pin diodes, examining the I-region characteristics and optimization of the

I-region thickness of the PIN diode design.

Table 3. Isolation performance comparison of solid state switches

PIN diode switches

Isolation at low frequencies

Average

(100 MHz range)

Isolation at high frequencies

Excellent

(18 GHz range)

FET switches

Excellent

Average

Hybrid switches

Excellent

Excellent

-20

-40

-60

-80

-100

-120

-140

-160

0

Agilent solid state switches isolation performance

0 5 10 15 20 25 30 35 40 45 50

Frequency (GHz)

Figure 8. Agilent solid state switches - typical isolation performance

U9397A 8 GHz hybrid FET SPDT

P9400A 8 GHz PIN diode transfer switch

P9402A 8 GHz PIN diode SPDT

P9404A 8 GHz PIN diode SP4T

85331B 50 GHz PIN diode SPDT

U9400A 8 GHz hybrid FET transfer

U9397C 18 GHz hybrid FET SPDT

P9400C 18 GHz PIN diode transfer switch

P9402C 18 GHz PIN diode SPDT

P9404C 18 GHz PIN diode SP4T

U9400C 18 GHz hybrid FET transfer

15

0

0

-1

-2

-3

-4

-5

-6

3.3 Insertion loss

Insertion loss, expressed in decibels (dB), of a switch is determined by measuring the power loss of a signal that is sent in through the common port and out from the port that is in the “ON” state.

Insertion loss plays an important role in many applications. In receiver applications, the effective sensitivity and dynamic range of the system is lowered by insertion loss. In system applications where the additional power needed to compensate for the loss is not available (amplifi ers in particular), insertion loss will be a critical specifi cation of a switch.

Insertion loss in solid state switches is generally attributed to three factors.

1. Resistance losses due to the fi nite resistance of series connected components, particularly PIN diodes and fi nite “Q” capacitors.

2. VSWR losses due to mismatch at the terminals of the switch or within the switch. With proper matching compensation techniques, mismatches

can reduced.

3. Conductor or transmission line loss within the switch itself due to the presence of microstrip, coaxial line or wave-guide inter-connection lines.

Switches also get more “lossy” as the number of arms or throws of the switch increases. Other factors that contribute to extra insertion loss include off-arm terminations and video fi lters.

As previously discussed, PIN diode switches typically have higher insertion loss at low frequencies due to the sharing of DC biasing and RF paths, while FET switches have higher insertion loss performance at higher frequencies due to the higher ON resistance of FETs.

Table 4. Insertion loss performance comparison of solid state switches

Insertion loss

PIN diode switches

Excellent

FET switches

Good

Hybrid switches

Average

2

Agilent solid state switch insertion loss performance

4 6 8 10 12 14 16 18 20

U9397A 8GHz hybrid FET SPDT

P9400A 8 GHz PIN diode transfer switch

P9402A 8 GHz PIN diode SPDT

P9404A 8 GHz PIN diode SP4T

U9400A 8 GHz hybrid FET transfer

U9397C 18 GHz hybrid FET SPDT

P9400C 18 GHz PIN diode transfer switch

P9402C 18 GHz PIN diode SPDT

P9404C 18 GHz PIN diode SP4T

U9400C 18 GHz hybrid FET transfer

Frequency (GHz)

Figure 9. Agilent solid state switches – typical insertion loss performance

16

3.4 Return loss and VSWR

Return loss, expressed in dB, is a measure of voltage standing wave ratio

(VSWR). Return loss is caused by impedance mismatch between circuits. At microwave frequencies, the material properties as well as the dimensions of a network element play an important role in determining the impedance match or mismatch caused by the distributed effect.

VSWR is an indicator of refl ected waves bouncing back and forth within the transmission line; this increases RF losses. Mismatched impedances increase

SWR and reduce power transfer. If VSWR is high, higher power in the transmission line also leaks back into the source, which might potentially causes it to heat up or oscillate. Solid state radios which have a lower tolerance for high voltages may automatically reduce output power to prevent damage. Looking from the standpoint of an entire test system, ripples due to the refl ection between system accessories and components can be minimized with low VSWR accessories.

In optical transmission, circuits which have high VSWR network elements are much more prone to inter symbol interference (ISI), which causes a low quality of service (QoS).

In solid state switch design, there will be a fi nite ON resistance whether the designers choose to use FET or PIN diode as the series on-off mechanism.

This causes an impedance mismatch which results in poor return loss. FETs, as mentioned, typically have a higher ON resistance than PIN diodes. Agilent switch designs incorporate proper matching circuits to improve the VSWR or return loss of the switch without sacrifi cing the other specifi cations of a switch.

All Agilent’s P940x, U9387, and U9400 solid state switches guarantee very low

VSWR and excellent return loss performance of at least 15 dB (VSWR = 1.43) up to 8 GHz and 10 dB (VSWR = 1.92) up to 18 GHz. This ensures optimum power transfer through the switch and the entire network.

17

3.5 Switching speed

Fast switching speed is important in ATE applications where product testing throughput is vital. It is especially important in applications that require the stacking of multiple switches in series. Another new technology usage is in the automotive industry, namely for adaptive cruise control (ACC) and collision avoidance systems (CAS), where high-frequency transmitting and receiving rates need to be thoroughly analyzed.

Defi nition

Switching speed is defi ned as the time needed to change the state of a switch arm (port) from “ON” to “OFF” or from “OFF” to “ON”. Switching speed is often characterized in two ways: rise/fall time and on/off time. Figure 10 shows a timing diagram of a switch and the defi nitions used to describe the switching time.

Rise time is the time it takes for the detected RF output* to raise from 10% to

90% of the fi nal value, when a switch arm is changed from an “off” state to an

“on” state.

Fall time is the time it takes for the detected RF output* to drop from 90% to 10% of the initial value, when a switch arm is changed from an “on” state to an “off” state.

Rise and fall times do not include the switch driver delay time.

ON time is the time period from 50% of the transition of the control signal to

90% of the detected RF output* when the switch arm is changed from an “off” state (isolation) to an “on” state (insertion loss).

OFF time is the time period from 50% of the transition of the control signal to

10% of the detected RF output* when the switch arm is changed from an “on” state (insertion loss) to an “off” state (isolation).

The ON and OFF times include the switch driver propagation delay.

*

Measured using a square law detector

18

50% of control signal transition

Driver propagation delay

50% of control signal transition

Driver propagation delay

Control signal

RF OFF

10% RF 10% RF

RF OFF

RF square law detected voltage*

* If a negative detector is used, the

detected RF voltage will portray

negative values when RF power is ON.

Rise time

ON time

90% RF

RF ON

90% RF

Fall

time

OFF time

Figure 10. Timing diagram and defi nition of switching time of a switch

Switching time can be measured by using a square law RF detector and an oscilloscope.

Oscilloscope

Signal generator

CW, Frequency = 6 GHz

Power = 10 dBm

Negative detector

*Drawing is not to scale.

15V

U9397 A/C

FET SPDT

GND

CTRL

+15V Pulse wave, period = 50 ms,

Duty cycle= 50%,

Rise time = 10 ns

Function generator

Figure 11. A typical setup for measuring the switching time of a switch

19

RF Off

50 % TTL control

Detected RF voltage

90% RF

Switching time (ON)

RF On

Figure 12. Typical switching time performance of Agilent P9400C PIN diode and U9400C hybrid FET transfer switches

Figure 12 shows the ON time measurement of a typical Agilent P9400C

18 GHz transfer switch. For P9400C, the measurement result shows an ON time

(including driver delay) of only 70 ns, while the rise time is < 30 ns. The measured OFF time (not shown in the graph) for this switch is much lower than the

ON time, exhibiting a typical value of 30 ns.

Table 5. Typical switching time performance of solid state switches

PIN diode switches FET switches

Rise time/fall time

Excellent (tens of nanoseconds)

Average

Switching time

Excellent (tens to

(On/Off)

hundreds nanoseconds)

Average

Hybrid switches

Average

Average

20

3.6 Settling time

As discussed in the previous section, switching time specifi es an end value of only 90% of the settled/fi nal value the RF signal. As the need for accuracy and precision has increased, settling time performance of a solid state switch, which measures to a level closer to the fi nal value, has become an important specifi cation. The widely used margin-to-fi nal-value of settling time is 0.01 dB (99.77% of the fi nal value) and 0.05 dB (98.86% of the fi nal value).

This specifi cation is commonly used for GaAs FET switches because they have a gate lag effect caused by electrons becoming trapped on the surface of the

GaAs. Agilent GaAs FET switches have a patented design that dramatically reduces the gate lag effect on and reduces the settling time to less than

350 microseconds.

There are two common ways to measure the settling time of a FET switch.

a) Oscilloscope measurement

Signal generator

Oscilloscope

CW, Frequency = 6 GHz

Power = 10 dBm

Negative detector

*Drawing is not to scale.

15V

U9397 A/C

FET SPDT

GND

CTRL

+15V

Pulse wave, period = 50 ms,

Duty cycle= 50%,

Rise time = 10 ns

Function generator

Figure 13. Measurement setup for measuring settling time using an oscilloscope

21

―0.01 dB + 0.115% Vp

10% Vp

Signal

}

90% Vp

Switching wave

Rise time/speed t = 0

Settling time

Figure 14. Switching waveform and settling time diagram

To translate 0.01 dB to detected voltage in volts (V).

Assume the detected V fi nal

10 * log (V settled

V settled

V settled

/V fi nal

= 200 mV

) = –0.01

= 10 (–0.01/10) x 200

= 199.54 mV

Note that a value of 10 is used instead of 20 when a square law detector is used because the instrument’s output DC voltage is proportional to the RF power delivered to the 50 ohm input.

As a result, you are looking for the point where the detected voltage is 0.46 mV down (99.77% from initial value) from the fi nal voltage. This method of measuring the settling time is less accurate than a network analyzer because the noise of the oscilloscope impairs measurement accuracy.

22

b) Network analyzer measurement

Measurement with an oscilloscope always has some uncertainties due to the resolution of the oscilloscope as well as the linearity and response time of the square law detector. A faster and more accurate measurement can be made using a network analyzer with an external trigger output. Figure 15 shows the setup for measuring an Agilent U9397C’s settling time using an Agilent PNA-L network analyzer.

Since the settling time measurement is measured to 0.01 dB of the fi nal value, the network analyzer is used to normalize it; this is shown in dB. The two output ports of the network analyzer are connected to switch input and output ports.

The trigger signal from the function or pulse generator is used to synchronize the switch control input with the PNA-L’s external trigger. Below are the steps used to measure the settling time.

1) Set the network analyzer to continuous wave (CW) and 100 KHz IF BW for a fast sweeping time. When measuring the settling time of the switch, turn on the switch (close) for the path that the network analyzer is measuring. Then wait for a few seconds until it’s fully settled and save the

S21 data (insertion loss) to the memory.

2) Set the trace to data/mem. Now, a 0 dB trace is seen on the screen.

3) Set the network analyzer to external trigger (or use rising/falling edge depending on which switch port is being measured).

4) When the trigger is set correctly, the transient response will show on the screen. Some averaging may need to be applied to get consistent measurement

results.

Network analyzer

External trigger

*Drawing is not to scale.

+15V

2

COM 1

U9397 A/C

FET SPDT

15V

GND

CTRL

50 ohm

Function/pulse generator

Figure 15. Measurement setup for measuring settling time using an Agilent PNA-L network analyzer

23

Figure 16. Typical settling time performance of the U9397A measured using a network analyzer

(0.05 dB settling time < 100

µ

s)

Figure 17. Typical settling time performance of the U9400A measured using a network analyzer

(0.01 dB settling time < 100

µ

s)

The U9397A (SPDT FET switch) settling time measurement shows that a 0.05 dB settling time is 91 µs while the 0.01 dB settling time is 212 µs. The U9400C achieves a settling time of 0.01 dB of 100 µs. The superb settling-time performance of Agilent’s solid state FET switches is the result of a patented design Agilent Technologies has incorporated into its switches.

Table 6. Typical settling time performance of Agilent solid state switches

Settling time

PIN diode switches

Excellent (< 50 μs)

FET switches

Good (< 350 µs)

Hybrid switches

Good (< 350 µs)

24

3.7 Video leakage

The word “video” was adopted from television, where the video signal (the picture) is carried on a VHF or UHF signal. As the name “video” suggests, video leakage or video feed-through signal spectrum is in the MHz to GHz range. [2]

Video leakage refers to the spurious signals present at the RF ports of the switch when it is switched without an RF signal present. These signals arise from the waveforms generated by the switch driver and, in particular, from the leading edge voltage spike required for high speed switching of PIN diodes. When measured using a 50Ω system, the magnitude of the video leakage can be as high as several volts. The frequency content is concentrated in the band below

200 MHz, although measurable levels can be observed as high as 1 GHz.

Most switches contain video leakage; the magnitude can be as low as a few mV to as high as 3 V in a 50Ω system. FET or hybrid switches generally offer lower video leakage, a few mV; while PIN diode switches have higher video leakage due to innate design requirements*. In PIN diode switches, the RF and DC biasing share the same path. When control voltage is applied to turn the switch on and off, a current surge will be generated. The DC block capacitor used on the

RF path causes the current to surge along the RF path when the switch is turned on and off. Some PIN diode switches have video leakage as high as

±10 V. Note that the current surge level depends on how fast the control voltage changes as well as the voltage level and the capacitor’s value.

Video leakage is simply a voltage dividing condition that occurs at 10 MHz or below. If the load’s DC input impedance is much higher than 50 ohm, the load will suffer higher video leakage than a load with lower input resistance.

Agilent’s solid state switches are carefully designed to ensure extremely low video leakage. For instance, P940x PIN diode switches offer video leakage of

500 mVpp which is extremely low for a PIN diode switch, while U9397 and

U9400 FET switches have less than 10 mVpp video leakage.

The amplitude of the video leakage depends on the design of the switch and the switch driver. Video leakage can damage sensitive devices, such as satellite transponders, which use low-power level switching (–100 dBm ON/OFF) and instruments, depending on the amplitude of the video leakage.

Oscilloscope

2 COM 1

Agilent

U9397C

FET SPDT

Function generator

Figure 18. Video leakage measurement setup

25

* Refer Agilent Application note “Video

Leakage Effects on Devices in Component

Test” for a more detailed explanation.

Figure 18 shows the video leakage measurement setup: an oscilloscope, power supply and function or pulse generator. The switch needs to be biased and all the ports are terminated by a 50-Ohm load except one, which is connected to the oscilloscope’s 50-Ohm input channel. The function/pulse generator is connected to the control line of the switch to toggle the switch on and off, the video leakage will be captured on the oscilloscope’s screen.

Table 7. Typical video leakage performance of solid state switches

Video leakage

PIN diode switches FET switches Hybrid switches

Figure 19. Typical video leakage performance of the U9400C hybrid FET switch (a negligible 2 mV)

Figure 20. Typical video leakage performance of the P9400C PIN diode switch

(0.5 V – very low for a PIN diode switch)

26

3.8 Off-port termination

A switch can be refl ective or absorptive. With refl ective switches, the RF signal at the “off” port is refl ected back to the source due to the poor match. In general, these switches have a simpler design, a slightly lower cost, and can handle higher power than absorptive switches. Absorptive switches provide a matched termination at the inactive ports. Because they absorb the RF signal, they are limited by the power-handling capability of the terminations. These switches are slightly more complex in their design.

3.9 Phase tracking

2

With tight control on process and lower level material tolerance, Agilent switches feature a phase difference of less than 10 degrees at microwave frequency range up to 18 GHz. Figure 21 shows the phase difference (in degrees) of ports of Agilent’s P9404C 18 GHz PIN SP4T switch with one of the ports acting as the reference port. It is obvious that the phase differences of the ports are tightly controlled.

Phase tracking of U9397C SP4T hybrid FET switch

3

0 2 4 6 8 10 12 14 16 18

1

0

-1

-2

Reference port 1

Phase tracking2

Phase tracking3

Phase tracking4

-3

-4

Frequency (GHz)

Figure 21. Typical phase difference (in degrees) performance of Agilent’s P9404C 18 GHz PIN SP4T switch

Test systems often require switches that are “phase tracked”. Phase tracking is the ability of a system with multiple assemblies or a component with multiple paths to closely reproduce their phase relative to each other. A phase tracking requirement is best achieved by fi rst equalizing the time delay between arms of a multi-throw switch. This requires a tightly controlled physical length of the arms from the input port to the output port. Secondly, the difference in phase from one unit to another within a product line should be minimized. Since the switch is made up internally of many elements, i.e. diodes, capacitors, and chokes with their accompanying mounting parasitic reactance and losses, it is necessary to control the uniformity of parts and assembly techniques to achieve the best phase tracking possible.

27

3.10 Harmonic and intermodulation distortion

Harmonic distortion

is a single-tone (single-frequency) distortion product caused by device nonlinearity. When a non-linear device (all semiconductors including solid state switches inherently exhibit a degree of non-linearity) is stimulated by a signal at a single frequency f1, spurious output signals can be generated at the harmonic frequencies 2f

1

, 3f

1

, 4f

1

,...Nf

1

. (Nth harmonic is the Nth order product). Harmonics are usually measured in dBc, dB below the fundamental output signal (see Figure 22).

Harmonic distortion

Fundamental

Amplitude x dBc ydBc f

1

2f

1

3f

1

Figure 22. Harmonics of a fundamental frequency

Frequency

Intermodulation distortion

arises when the nonlinearity of a device or system with multiple input frequencies causes undesired outputs at other frequencies, causing the signals in one channel to interfere with adjacent channels. Reducing intermodulation distortion has become more important as the spectrum becomes more crammed and channels are more tightly spaced.

These spurious products are mathematically related to the fundamental input signals. It is common practice to limit the analysis to two tones (two fundamental frequencies, f

1

and f

2

, which are normally separated by a small offset frequency of around 1 MHz) due to the complexity analyzing more than two input frequencies at a time. The output frequencies of the two-tone intermodulation products are:

P f

1

± Q f

2

, where P, Q = 0, 1, 2, 3,.....

The order of the distortion product is the sum of P + Q. So, the third order interf modulation products of the two signals, f

1

1

and 2f

2

– f

1

(see Figure 23).

and f

2

, would be 2f

1

+ f

2

, 2f

1

– f

2

, 2f

2

+

28

Amplitude

Passband f

1 f

2 f

1

+f

2

2f

1

2f

2

2f

1

-f

2

2f

2

-f

1 f

2

-f

1

Frequency

Figure 23. Second and third order intermodulation distortion

Output power

Calculated/interpolated third order intercept point

3f

1

2f

1

+f

2

2f

2

+f

1

3f

2

Fundamental signal

Compression effects

Third order product

Input power

Figure 24. Realistic representations of third order intercept point

Broadband systems may be affected by all the non-linear distortion products.

However, unlike harmonic and second order distortion products, third order intermodulation distortion products (IP3) are always too close to the fundamental signals to be easily fi ltered (see Figure 23). When strong signals from more than one transmitter are present at the input of the receiver, as is commonly the case in cellular telephone systems, unwanted distortion products will be generated affecting the receiver measurement.

The nonlinear transfer function of a device or system can be expressed as a series expansion:

V out

= a

1

Acos (ωt) + a

2

A

2 cos (ωt) + a

3

A

3 cos (ωt)…

Higher order intermodulation products will increase in power much faster than the fundamental as shown in Figure 24. The intermodulation signal will increase in power three times faster than the carrier signal. The increase is in dBm units, which are a logarithmic function. The same relationship holds true for other harmonics or intermodulation products. For instance the second harmonic and the second order intermodulation products’ power will increase two times (in dBm) faster than the fundamental, while the third harmonics will increase three times as fast.

Agilent solid state switches have mainstream distortion performance. For instance, the U9397A/C SPDT switches have a typical third order intermodulation of 55 dBm (input power) and second harmonic intercept (SHI) and third harmonic intercept (THI) of 70 dBm and 55 dBm respectively.

29

3.11 1 dB compression point

Another nonlinearity of a system or device is measured by the compression point. In the linear region, a 1 dB increase in input power to a device will correspond to 1 dB increase in the output power. The nonlinearity effect becomes apparent when the output power starts to increase less than the input power.

When the power difference increment is 1 dB, the device has a 1 dB compression point. If not explicitly stated, the 1 dB compression point refers to the output power (P out

) at that point.

Agilent U9397A has high 1 dB compression point typically at 29 dBm, while the

U9397C is typically at 25 dBm.

Output power

Output

1 dB compression point

Fundamental signal

1dB

Input

1 dB compression point

Input power

Figure 25. 1 dB compression point

30

4.0 Applications

Application 1

Mobile handset: power amplifi er testing

Critical parameters: settling time, video leakage

Figure 26 shows a simplifi ed test setup of a dual-band mobile handset power amplifi er. A signal generator with digital modulation capability supplies the test signal to the power amplifi er and a vector signal analyzer (VSA) is used to measure the output signal from the power amplifi er. Two switches are used to switch between the PCS and GSM bands and two attenuators are placed at the output of the power amplifi er to protect the switches. The triggering signal

(frame trigger) from the signal generator is used to synchronize the VSA and trigger the switches to test the correct band of the power amplifi er at the right time.[3]

Switch selection is very important in this application for two reasons: First, the switch must have a settling time that is fast enough to allow the VSA to capture any timeframe of the signal. Figure 6 shows a timing diagram for a GSM/EDGE signal, as you can see one slot equals 577 μs. So, when the signal generator sends a frame trigger signal out, the switches must switch and settle within

577 μs so the VSA can start to capture data within the time frame of the slot 1 signal to ensure accurate measurements.

Agilent U9397A/C incorporates a patented design which reduces the settling time to < 350 μs (measured to 0.01 dB of the fi nal value). Most FET switches available in the market today have a typical settling time of > 50 ms.

The second reason careful switch selection is needed is video leakage. Typical

PIN diode switches have video leakage of ≥ 1 Vpp due to the nature of PIN diode switch design. This can potentially damage power amplifi ers because their maximum input power is typically < 13 dBm. The other alternative would be electromechanical switches that have low or no video leakage but the switching speed (typically in ms) is too slow for this application.

As discussed in Section 3,

“Solid State Switch Specifi cations”

, PIN diode switches have higher video leakage than FET or hybrid switches. Nevertheless, careful design of Agilent P9400x PIN diode switches gives them very low video leakage, typically less than 0.6 Vpp, which is excellent for PIN diode switches.

If lower video leakage is needed, both Agilent FET hybrid switches (U9397x and

U9400x) offer extremely low video leakage of < 10 mVpp.

31

Signal generator

Triggering

DCS band

GSM band

Power amp

Attenuator

Attenuator

Signal analyzer

Figure 26. Simplifi ed test setup for testing GSM/EDGE handset power amplifi ers

4.615 ms

Frame trigger

Slot

576.92 μs

Figure 27. Timing diagram of a GSM/EDGE signal

32

Application 2

Signal routing: multiple-instrument, multiple-DUT testing [4]

Critical parameter: isolation

This section shows how isolation affects measurement accuracy. Figure 28 shows the Agilent P9400C 18 GHz transfer switch confi gured to test two DUTs simultaneously, each through different sets of equipment targeting distinctive measurement parameters.

Signal generator

DUT 1

Agilent P9400A/C transfer switch 1

3 4

DUT 2

1 2

Network analyzer

Spectrum analyzer

Agilent P9400A/C transfer switch 2

4 2

–30 dBm

3 1

Figure 28. Switching and testing two DUTs between two different test sets simultaneously

Table 8. Switching paths of different test setups

Control input

High

Low

State

DUT 1 connected to

Network analyzer

Spectrum analyzer and signal generator

DUT 2 connected to

Network analyzer

Spectrum analyzer and signal generator

Tests

S-parameter

Spurious signal, harmonics

DUT 1 is tested for S-parameters using network analyzer, while DUT 2 is tested on harmonics and spurious signals using signal generator and spectrum analyzer.

33

Agilent U9400C transfer switch’s isolation performance is shown in this application example. Assume that the transmitted power coming out of the DUT

1 is –30 dBm. At the same time, DUT 2 is tested on spurious signal, which might be measured to a value as low as –110 dBm or more. This requires switch #2 to have an isolation of more than –110 dBm – (–30 dBm) = –80 dB, which is clearly fulfi lled by the isolation performance of U9400C of > 90 dB isolation at 18 GHz.

This is only one of the countless measurement uncertainties that require careful attention. Switches with low or moderate isolation performance can impair system accuracy and time is wasted calculating for the uncertainties and dealing with the complexities that result from the precise timing requirements of ultra fast switching and testing systems.

Isolation is a parameter that is more diffi cult than insertion loss to calibrate out of a test system, therefore isolation is often seen as the critical parameter of a switch.

Table 9. Agilent solid state switch isolation performance

U9397A/C

Hybrid

P9402A/C

PIN diode SPDT

P9404A/C

P9400A/C

PIN diode SP4T

PIN diode transfer

U9400A/C

Hybrid

at 8 GHz (dB)

100

80

80

80

100

Isolation specifi cation at 18 GHz (dB)

80

80

90

80

90

34

Application 3

Filter bank: SAW fi lter testing [5]

Critical parameters: switching speed, settling time, insertion loss, compatible logic

Network analyzer

P9404A/C P9404A/C

Filter bank

Figure 29. Typical test setup for fi lter bank testing

SAW fi lters are a high-volume product which could be produced in quantities of two million units per month, so fast switching speed/settling time is needed.

With fast switches, test system designers can avoid unnecessary delays between measurements.

Figure 29 shows a typical test setup for fi lter bank testing. Two SP4T absorptive

PIN switches are needed for S-parameter measurements with an Agilent ENA network analyzer. Mobile handset and semiconductor manufacturers use PIN diode switches because fast switching speeds are needed for the high volume testing. P940xA/C PIN diode switches are particularly suitable for this application because they provide very fast switching and rise time, which prevent premature measurements; low insertion loss, which optimizes the dynamic range; and TTL control, which enables easy switch control using +5 V or 0 V.

35

* For normal distribution, the probability of

exceeding ±4σ is 0.00006, half of it will be on

the right hand side (failing)

* For normal distribution, the probability of

exceeding ±2σ is 0.05, half of it will be on the

right hand side (failing)

Settling time

A typical FET hybrid switch has a settling time of 10 ms compared to Agilent

U9397 and U9400 FET hybrid switches which offer a settling time of < 350 µs.

If you sweep the network analyzer for an insertion loss measurement and the sweeping time across the screen is 1 ms, it is not diffi cult to imagine the effect of a switch with moderate or even substandard settling time on the measurement’s accuracy.

For example, consider a production line where SAW fi lters have a typical mean performance of 1 dB insertion loss. The test specifi cation was set at 1.2 dB.

The assumption is that the SAW fi lters’ insertion loss performance will fall into a normal distribution with a standard deviation of 0.05 dB. At 4 standard deviations away where the test specifi cation is set, 0.003%* or 30 pieces of

SAW fi lters out of a million will fail the specifi cation in a normal production run.

Now, consider a case where a switch with a long 0.01 dB settling time, 10 ms for instance, is used in the test system. The switch might not settle in time to allow the you to make an accurate measurement on the SAW fi lters. Assume that at the mid span of the network analyzer’s 1 ms-sweep, the switch has just reached 0.1 dB of its fi nal value and you take the measurement at this point of time. This means that an extra loss of 0.1 dB has been unjustly “imposed” on the insertion loss performance of the SAW fi lter, measuring it as 1.1 dB (mean) instead of 1 dB (mean). The measurement has now shifted to only 2σ away from the test specifi cation, so 2.5%* or 25,000 SAW fi lters out of a million will fail the specifi cation (of which only 30 pieces are true failures!). What is worse is that this might not be perceived by the manufacturer, leading to unnecessary yield loss.

Insertion loss

For this insertion loss example, the stopband of the SAW fi lter is about –70 dB to –80 dB and the network analyzer dynamic range is about 110 dB. In SAW fi lter manufacturing testing, a comprehensive switch matrix which contains multiple switches in series is commonly used, so the insertion loss measurement might have to go through four or more switches for a measurement.

If each switch has about 4 dB loss and the cable loss is 4 dB, then the total loss of the measurement path will be 20 dB. So, the total insertion loss including the

80 dB stopband-attenuation of the fi lter itself will be about 100 dB. This is only a 10 dB margin from the dynamic range of the instrument. If each of the switch contributes an extra loss of 1 or 2 dB, you might end up measuring the noise fl oor of the instrument instead of the actual SAW fi lter stopband performance.

36

RF IN

A1

Application 4

Satellite: testing channel amplifi ers with ALC systems

Critical parameter: video leakage

AT1

A2

AT2

A3

RF OUT

Detector

VR1

Step gain control

VR2

DC amp

Figure 30. Channel amplifi er with an automatic level control (ALC) system for satellite applications

Figure 30 shows a typical channel amplifi er with an automatic level control (ALC) system for satellite applications. The ALC controls the input of the fi nal power amplifi er, for instance a traveling wave tube amplifi er (TWTA), that is dependant on the input power level of the channel amplifi er. If a PIN diode switch is used to test this device, the video leakage signal could perturb the detector causing the gain of the device under test to alter signifi cantly. This can overstress or damage the TWTA. In this situation, the device could not be installed in a satellite for fear of premature failure.

Both Agilent FET hybrid switches (U9397 and U9400 series) offer extremely low video leakage of < 10 mV, which protects sensitive devices from video leakage damage.

37

Application 5

Base station and satellite: antenna testing

Critical parameters: switching speed, settling time, isolation, impedance matching

Source antenna

Antenna under test

V

H

Switch control unit

SP2T

PIN switch

SP4T

PIN switch

Switch control unit

From transmit source

To receiver

Figure 31. A typical multiple-channel, multiple-frequency system confi guration[6]

The high cost associated with the manufacture and installation of systems containing antennas, such as base stations and satellites, followed by the subsequent inaccessibility to such systems, makes it imperative that they be designed to provide highly reliable performance throughout their life cycle. As key components of these systems, antennas are extensively tested in many modes of operation to ensure they meet necessary specifi cations prior to launch.

Input/output beam characteristics are tested as a function of different variables.

For instance, amplitude or gain might be measured against frequency, azimuth, elevation, distance, or time. In addition, phased-array antennas have a plurality of sensing and/or receiving elements (feeds), beam forming networks, input/output ports and/or channels which may require many tests. In fact, antenna testing is so extensive that millions of gain measurements alone may be required![7]

Figure 31 shows a typical confi guration with Agilent PIN diode switches connected to the source antenna and the antenna under test. Agilent solid state switches provide fast-switching and settling time which makes them ideally suited for testing antennas with multiple channels or ports, as well as for applications requiring both co- and cross-polarization measurements. For instance, one PIN diode switch can switch transmit polarization, while a second switch switches between the separate test ports of the antenna. With this technique, the co- and cross-polarization response of each test port can be measured in one rotation of the antenna. The ability to rapidly switch transmit and receive polarization also enables full polarimetric radar cross-section (RCS) measurements to be made quickly and easily.[8]

In cases where switches with a slow settling time are used, test developers have to idle the test program for a few milliseconds or more before each switching cycle in order to allow the switches to fully settle. Agilent FET switches, with a settling time of 350 µs at 0.01 dB (99.77% of fi nal value); and PIN diode switches, with settling time of typically less than 50 μs; provide fast, reliable switching in antenna test systems.

38

5.0 Conclusion

6.0 Reference

This application note has delivered a detailed overview on the types and technologies of solid state switches. Important parameters for selecting a solid state switch have been discussed throughout the application note, and comparisons have been made of the different types of solid state switches, namely the

PIN diode switch, the FET switch and the hybrid switch.

PIN diode switches have better insertion loss performance at higher frequencies compared to hybrid or FET switches. In contrast, FET and hybrid switches have better insertion loss performance at lower frequencies, around kHz to MHz.

PIN diode switches have excellent rise time and settling time performance, in the nanosecond range compared to FET or hybrid switches. However, Agilent’s patented design improves the settling time of its FET switches to < 350 μs.

Hybrid or FET switches, have minimal or negligible video leakage of < 10 mV p

, this is much lower than PIN diode switches at around 1 V. Agilent’s FET switches has also have excellent isolation performance, up to 100 dBm.

Finally, a number of application examples that illustrate the critical parameters of solid state switches were given, and their signifi cance in today’s performancedemanding test and measurement systems were examined. The application examples explained the pros and cons of using different types of switches helping you to understand how switch performance affects the measurement integrity of test systems.

1. Agilent Technologies, “Agilent Solid State Switches Application Note:

Selecting the right switch technology for your application”, Literature number: 5989-5189EN, 2006

2. Agilent Technologies, “Agilent Video Leakage Effects on Devices in

Component Test Application Note” Literature number: 5989-6086EN, 2007

3. Agilent Technologies, “Agilent U9397A/C FET Solid State Switches (SPDT)

Technical Overview”, Literature number: 5989-6088EN, 2007

4. Agilent Technologies, “Agilent P9400A/C Solid State PIN Diode Transfer

Switches Technical Overview” Literature number: 5989-7215EN, 2007

5. Agilent Technologies, “Agilent P940xA/C Solid State PIN Diode Switches”

Literature number: 5989-6695EN, 2007

6. Agilent Technologies, “Agilent 85331B/85332B Solid State Switches 85331B

SP2T 45 MHz to 50 GHz 85332B SP4T 45 MHz to 50 GHz Technical

Overview”, Literature Number: 5989-4960EN, 2006

7. Theodore S. Fishkin, Lawndale; Mark Skidmore, Long Beach; Peter

Dimitrijevie, Redondo Beach, all of Calif., “Antenna Test and Measurement

System” United States Patent, Nov 28. 1989, Patent Number 4,884,078

8. Agilent Technologies, “Agilent Antenna Test Selection Guide”, Literature number: 5968-6759E, 2005

39

www.agilent.com

www.agilent.com/find/mta

Agilent Email Updates www.agilent.com/fi nd/emailupdates

Get the latest information on the products and applications you select.

www.lxistandard.org

LXI is the LAN-based successor to

GPIB, providing faster, more efficient connectivity. Agilent is a founding member of the LXI consortium.

Agilent Channel Partners

www.agilent.com/find/channelpartners

Get the best of both worlds: Agilent’s measurement expertise and product breadth, combined with channel partner convenience.

Remove all doubt

Our repair and calibration services will get your equipment back to you, performing like new, when promised. You will get full value out of your Agilent equipment through-out its lifetime.

Your equipment will be serviced by Agilent-trained technicians using the latest factory calibration procedures, automated repair diagnostics and genuine parts.

You will always have the utmost confi dence in your measurements.

Agilent offers a wide range of additional expert test and measurement services for your equipment, including initial start-up assistance, onsite education and training, as well as design, system integration, and project management.

For more information on repair and calibration services, go to:

For more information on Agilent Technologies’ products, applications or services, please contact your local Agilent office. The complete list is available at:

www.agilent.com/find/contactus

Americas

Latin America

United States

Asia Pacific

Australia

China

Hong Kong

India

Japan

Korea

Malaysia

Singapore

Taiwan

Thailand

305 269 7500

(800) 829-4444

1 800 629 485

800 810 0189

800 938 693

1 800 112 929

0120 (421) 345

080 769 0800

1 800 888 848

1 800 375 8100

0800 047 866

1 800 226 008

Europe & Middle East

Austria

Belgium

Denmark

Finland

France

43 (0) 1 360 277 1571

32 (0) 2 404 93 40

45 70 13 15 15

358 (0) 10 855 2100

0825 010 700*

www.agilent.com/fi nd/removealldoubt

Ireland

Israel

Italy

Netherlands

Spain

1890 924 204

972-3-9288-504/544

39 02 92 60 8484

31 (0) 20 547 2111

34 (91) 631 3300

Sweden

Switzerland

0200-88 22 55

0800 80 53 53

United Kingdom 44 (0) 118 9276201

Other European Countries: www.agilent.com/find/contactus

Revised: October 1, 2009

Product specifi cations and descriptions in this document subject to change without notice.

© Agilent Technologies, Inc. 2008, 2010

Printed in USA, May 21, 2010

5989-7618EN

Was this manual useful for you? yes no
Thank you for your participation!

* Your assessment is very important for improving the work of artificial intelligence, which forms the content of this project

Download PDF

advertisement