Application Notes - How To Specify Pin Diode Switches
When purchasing PIN diode switches, it is important that they are completely specified to assure system performance. It is also important that the specifications be achievable. This page is designed to help a systems designer specify realizable PIN diode switches.
There are six key parameters essential to specify PIN diode switches. These are:
- Type, i.e., SPST, SPDT, SP3T, DPDT, etc.
- Operating frequency band
- Insertion Loss
- Switching speed
- Power handling
There are five secondary parameters that may require specification. These are:
- Logic compatible driver type and speed
- Phase tracking arm to arm and/or unit to unit
- Off arm terminations
- Intercept point or compression point
- Video transients
II. SWITCH TYPE
Most PIN diode switches are of the single pole multiple throw type. They range from single throw up through 8 to 64 throws. The most popular type is the SPST or pulse modular type. In general, the greater the number of throws, the less popular the switch, and, hence, the less readily available it is. PMI has standard switch designs up through 5 throws in the three popular bands of interest: HF, UHF/VHF, and Microwave. We also have designs for 8 and 10 throws at HF and Microwave.
The most popular multi-pole throws is the DPDT type, commonly known as the Transfer Switch. These units are available in UHF/VHF and Microwave bands. High order multipole switches are generally referred to as switch matrices, which is a whole subject matter by itself.
III. OPERATING FREQUENCY BANDS
PMI classifies PIN switches into five operating frequency bands. They are:
- Video, which covers from 100kHz to 2MHz.
- HF, which covers 2MHz to 32MHz.
- UHF/VHF, covering 10MHz to 2000MHz.
- Microwave, covering 10MHz to 20 GHz.
- Millimeter wave switches, 20GHz to 40GHz.
The above bands have loosely defined boundaries which overlap. They are more indicative of the five different technologies available to the switch manufacturer as well as distinct application area of switch requirements.
There are some special application bands and technologies such as the high speed, low transient IF switching technology which PMI offers.
IV. THE PIN DIODE
A simplified equivalent circuit of the PIN diode is shown in figure 1. The forward biased diode is a current controlled resistor. The resistance vs. current behavior of a typical PIN diode is shown in figure 2. The reversed biased diode is a voltage-controlled capacitor. The capacitance vs. voltage of a typical PIN diode is shown in figure 3.
V. INSERTION LOSS
Simple, most basic switches have the lowest loss for any given operating band. For a given technology or operating band, insertion loss increases with increasing frequency proportional to the square root of frequency in a well-designed PIN switch. Insertion loss originates in three basic areas.
- Conductor or transmission line loss within the switch itself due to the presence of micro strip, coaxial line, or wave-guide inter-connection lines.
- Resistance losses due to finite resistance of series connected components such as PIN diodes and /or finite "Q" capacitors.
- VSWR losses due to mismatch of components within the switch or at the terminals of the switch. VSWR losses at the terminals of the switch can be tuned out externally to improve losses; those within the switch must be minimized in design. These actually are the cause of ripples in the insertion loss vs. frequency characteristic.
Assuming a switch is well designed, i.e., lowest loss transmission media, lowest resistance diodes and other series components are employed and all internal VSWRs are minimized, the loss of the switch is then dependent on the complexity of the design. In general, multi-throw units are more lossy as the number of throws increases. The addition of off-arm terminations and video filters increases the loss of the switch for a given technology. Also, increases on/off isolation will contribute slightly to the loss. The insertion loss is lowest in the least complex switch configurations. For low loss switches, keep the specification simple.
PIN diodes are connected to the transmission line in series or in shunt. Isolation is achieved by reverse biasing series connected diodes for forwards biasing shunt connected diodes. The shunt mounted diode provides the most effective means for achieving broadband, relatively frequency independent isolation. It is ideally frequency independent, but, practically, small parasitic reactance generally affects broadband performance. Isolation is also achieved by reverse biasing series mounted diodes. Isolation for the series mounted diode decreases with increasing frequency.
Series-shunt diode configurations are frequently employed in multi-throw broadband switches to achieve relatively high isolation in a simple structure. An example of the performance of a series-shunt connection is shown in figure 4. Note how the isolation decreases with increasing frequency. Multiple diodes connected in series or in shunt are frequently employed in PIN switches to achieve relatively high isolation over a broad band of frequencies. The isolation vs. frequency characteristic of a shunt connected array of forwards biased diodes is shown in figure 5. An example of a shunt mounted switch is shown in figure 6, which achieves 85 dB isolations over the 2-18 GHz band by judiciously spacing four shunt connected diodes. An example of a switch employing an array of reverse biased series connected diodes is shown in figure 7, which achieves 70dB minimum isolation over the 10-2000 MHz band.
For narrowband applications, the possibilities are endless for combining and tuning diodes for excellent trade-offs between insertion loss and isolation. Many designers have employed series and shunt inductors to resonate the capacitance of reverse biased PIN diodes to achieve excellent isolation-insertion loss performance over limited frequency bands. (See reference 1.)
VII. SWITCHING SPEED
Switching speed of a PIN diode switch is generally defined as the time for the RF to traverse 10% to 90% levels. Other definitions, such as the time from 1 dB to 60 dB levels, are occasionally employed for high isolation requirements. The switching speed is generally controlled by two factors, the time required to remove the stored charge from the diode junction and the theoretical maximum speed at which the charge can be removed from the junction. The time required to remove the stored charge from the junction is limited by the transit time of the PIN diode.
The transit time given by:
Where W1=the device I-region thickness (cm)
Vs=maximum saturated velocity=10*7 cm/sec
The I-region thickness is related to the breakdown voltage Vb by:
Additionally, the stored charge in the forwards biased diode junction is related to the minority carrier lifetime of the junction by:
Where Qs=stored charge (coulombs)
I=forward current (amperes)
T=minority carrier lifetime (seconds)
As a minimum for operation as a PIN switch, the diode lifetime is shown vs. the lowest operating frequency in figure 8. Further, the transit time as a function of breakdown voltage is shown in figure 9 (see reference 2). For minority carrier lifetimes shorter than 10 nS, state-of-the-art PIN drivers can switch in approximately the transition time of the device. Longer lifetimes require higher currents and large, slower switching transistors causing switching times to be longer than the transition time.
Low intermodulation and harmonic distortion PIN switches require diodes with longer than minimum minority carrier lifetimes and hence switch more slowly.
High power PIN switches require higher Vb diodes which results in slower transition times and slower switching times.
VIII. POWER HANDLING
The power handling capability of PIN diode switches is controlled by three parameters. First is the upper operating temperature of the device. Second is the breakdown voltage and third the charge storage capability of the device. For silicon PIN diodes, best reliability is achieved by keeping junction operating temperatures below 200 degrees centigrade. Since series mounting diodes are more dissipative and have poorer heat sinking capabilities than shunt mounted configurations, switch designers tend to avoid series configurations in high power applications. Since series configurations are essential to wideband multi-throw switches, these units tend to be the lowest power handling configurations. Hence, high power broadband switches are difficult to realize. One usually ends up trading power for bandwidth.
It is necessary that the breakdown voltage be at least twice the peak RF voltage that the diode will see and that the forward charge stored in the junction be greater than the charge moved on one-half cycle of the RF current waveform. The former requirement will assure that the diode does not exceed its voltage breakdown and the latter that the forward biased junction will not be depleted in operation. The elements are essential to linear non-destructive operation of the diode under high power operation.
IX. LOGIC COMPATIBLE DRIVERS
The three most popular logic families are Transistor-Translator-Logic (TTL), Emitter Coupled Logic (ECL) and Metal Oxide Semiconductor (MOS/CMOS).
Of the three, TTL logic is by far the most popular, ECL and CMOS are a distant second. Four of the most popular forms of TTL driver circuits are shown in figure 10. We will confine this discussion to TTL compatible drivers. For best performance, switch drivers must be electrically as well as mechanically integrated in the switch unit. It is possible to achieve clean, transient free switching by designing electrically compatible drivers.
"Unit load" drivers are highly desirable because they are compatible with the widest range of TTL product line I.C.s. a "unit load" is defined as 40 microamperes maximum source current and 1.6 milliamperes maximum sink current. Drivers are available in multiples of "unit load." True TTL compatibility also requires a logic "low" to be 0–.8 volts and a logic "high" to be 2.0–5.0 volts at the input (0.8–2.0 volts is an undefined region.)
All TTL compatible drivers have delay. Generally the driver delay is defined as the time from 50% TTL level to where the RF signal changes by 10%, i.e., 1–10% for turn-on or 100–-90% for turn-off. It is caused by energy storage in the driver and/or RF circuitry. The delay is a result of the time required to remove the stored energy before the switch state can be changed. The stored energy can be stored charge in the base region of a switching transistor or stored in various capacitors and inductors in the driver circuit or the bias decoupling circuit. Often this delay is different for turn-on or turn-off. This phenomenon can lead to pulse shrinkage or pulse expansion when the PIN switch is operated in a pulse mode. Since driver delay is consistent from unit to unit in a well designed PIN switch, a system designer can often pre-trigger the switch and essentially "program-out" the driver delay. When it is not possible to anticipate the delay, it is necessary to specify delay equalization. An example of a PIN switch with equalized delay is shown in figure 11. This unit has on/off delay equalization to 5 ns, maximum.
Another phenomenon of driver delay is minimum pulse width. Since delay involves charging and discharging of components within the driver circuit, it is necessary to "charge" or "discharge" the driver before any RF changes in signal level are observed. This results in minimum pulse width for any switch with integral logic drivers. The minimum pulse width is approximately equal to the delay in the driver.
X. PHASE TRACKING
Often systems require switches that are "phase tracked." A phase tracking requirement is best achieved by first equalizing the time delay between arms of a multi-throw switch (if a multi-throw is indicated) and equalizing the time delay from unit to unit within a production run or product line, if required.
Since the PIN 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 best phase tracking.
For unit-to-unit phase tracking on a lot-to-lot basis, it is necessary to build a phase standard unit that is maintained at the switch vendor’s facility which has an impact on the price of the initial lot of switches.
Typical state-of-the-art phase tracking is as follows:
XI. OFF ARM TERMINATIONS
Often PIN switches are employed to commutate or switch VSWR sensitive components such as antenna elements in an array, oscillators or amplifiers. Normally, switches have an infinite VSWR in the OFF position. Figure 12 shows a switch with off arm terminations having an extra switching section that switch the terminal in question into a matched load when that arm is turned off. This, in effect, controls and stabilizes the VSWR in both the ON and OFF condition of the switch. You must specify off arm terminations when it is necessary to control OFF VSWR.
Be aware that when the specified arm is commutated or switched there is a period of time when the VSWR is unspecified. This is particularly important in high power switches where momentary high reflected power levels can be troublesome.
The addition to off arm terminations adds complexity to the switch.
XII. INTERCEPT POINT OR COMPRESSION POINT
Compression in a PIN switch is a less well defined parameter than in, say, an amplifier. So, we will limit our remarks in this section to intercept point. The concept of intercept point is well documented in the literature and we will not go into it here. Rather, we will examine the elements that control intercept point of PIN diode switches and their tradeoff on overall switch performance.
Intermodulation is a result of nonlinear mechanisms within the PIN diode primarily and occasionally caused by other elements such as nonlinear capacitors, resistors, and/or ferrite cores in the bias decoupling chokes. We will confine this discussion to the PIN diode only.
The primary intermod generator in a PIN switch is the forward biased series PIN diode. Intermod is generated in the diode when the stored charge becomes close to being swept out (or depleted) from the “I”layer region. Hence, low intermod switches employ diodes with longer than minimum minority carrier lifetimes and are biased at relatively high forward currents to store a log of charge in the junction. The degree of linearity is controlled by the percentage of charge depleted from the junction by the RF cycle. Highly linear switches have small percentage of charge depletion. See reference 3 for a more complete discussion of Intermodulation Distortion Mechanisms.
A secondary intermod generator is the non-linear capacity vs. voltage characteristic of the reversed biased PIN diode. This phenomenon is relatively easily controlled by selecting diodes with flat capacitance vs. voltage characteristics and biasing the device into that region of the curve.
XIII. VIDEO TRANSIENTS
Refer to figure 13, the equivalent circuit of a typical PIN switch. When the diodes are switched between biasing conditions, a change of voltage or current occurs at the bias decoupling element adjacent to the output terminals. The element acts to differentiate the waveform (current for the shunt inductor and voltage for the series capacitor) and cause a pulse, spike, or video transient at the output terminal. This transient occurs in all PIN switches but is controlled by various means.
The most effective means of controlling video transients are:
- Slowing the switching waveform
- Filtering the video spectrum
- Balancing or canceling two equal video transients
The first is very effective when switching speed is not important. Slowing the switching waveform will slow switching speed. The second is effective when the switch operating band is above the frequency band where the video spectrum is concentrated. The addition of high pass filters at the input and output terminals of PIN switches at frequencies above 500 MHz has proven very effective in reducing transients. Typically, the highest speed switches (1ns) have at least 90% of the video spectrum below 1 GHz. Filtering has its accompanying side effects. It will often introduce unwanted "ringing" in the switching waveform. Balancing has been employed very effectively as a means of reducing video transients without affecting switching speed or introducing "ringing." Unfortunately, present state-of-the-art technology has limited balancing technique to UHF/VHF band. An example of the balancing technique is the series of IF switches shown in figure 14.
XIV. CONCLUSIONSix essential and five supplementary parameters have been presented to aid in the specification of PIN diode switches. Tradeoffs between the various parameters have also been explored. It is hoped that this will help bridge the gap between switch users and switch designers.