Manufacturers of satellites, satellite launch vehicles, and other high reliability equipment face many challenges when designing electromechanical relays (EMR) into their systems. Preventing false relay operation from shock and vibration requires additional safety considerations. In addition, "hash filters" are sometimes necessary to debounce the contacts, which add space and weight. However, solid state relays (SSRs) are immune to the shock and vibration levels normally encountered in the application and do not need contact filters. As a result, the use of SSRs in place of the mechanical type leads to a more reliable end product. However, the transition to SSRs requires understanding a number of new parameters to ensure proper operation.
Changes With SSRs
Various electronic components in an SSR take the place of the electromechanical elements in a mechanical relay. The SSR uses a MOSFET in place of mechanical contacts to eliminate contact bounce. However, if there is any significant power dissipation involved, proper heat sinking will be necessary to ensure safe operation (see sidebar, MOSFETS Make it or Break it).
The traditional relay uses an electromagnet in conjunction with other mechanical components to affect contact actuation. In contrast, the SSR employs a photovoltaic isolator (PVI). As a minimum, a PVI consists of an LED and a photovoltaic array. When sufficient current flows through the LED, the light output falls on the array, generating a voltage that charges the FET gate. This turns on the SSR. To ensure that the MOSFET turns off when required, there is a gate discharge circuit present. In addition, an input buffer is also included allowing direct interface to logic devices.
The physical separation between the LED(s) and the array in a SPDT (Form C) dc SSR provides at least 1000V of isolation between the input and the output of the SSR. The enhancement mode MOSFET requires a gate voltage to turn it on and the depletion mode requires a gate voltage to turn it off.
In relay terms, this particular configuration is a Single Pole, Double Throw (SPDT) or a "Form C" relay. Since the MOSFET has a body diode, this arrangement does not block current flow in both directions, limiting its use to dc circuits. For ac usage, an additional MOSFET in series, source-to source, blocks current in both directions.
With its inherent isolation, the SSR can drive either the high side or the low side, or even between sides, as in a solar array battery charger. In this application, an additional isolation diode prevents battery discharge into the solar array through the FET body diode when the charger is off. Another way to accomplish this with less power dissipation is to use an ac SSR (or to use both channels of a dual dc SSR, and connect them in series, source to source) in place of the dc SSR/diode combination.
SSR Switching Times
EMRs do not switch very fast, perhaps in the neighborhood of 5 to 20 msec. The switching time of a normally constructed SSR is in the same ballpark, due primarily to the poor current transfer ratio of the PVI. The PVI output current has to charge the gate capacitance of the output FET and this takes time. For most applications, this medium speed response is acceptable. For faster switching, multiple "hot" (or so-called fast) PVIs can be used. These produce more current to charge the FET gate, resulting in a faster turn-on.
In other applications, however, system designers may be concerned with dV/dt and/or dI/dt. Keeping these values low helps keep electromagnetic interference (EMI) at acceptable levels. For instance, many satellites have multiple heater wires that keep the electronics and/or sensors at proper temperatures. A fast dV/dt on these wires could broadcast noise into internal electronics dictating the need for slower rise times.
To accommodate these EMI requirements, SSRs are constructed with additional internal circuitry that keeps dV/dt and dI/dt at "controlled" (i.e., slower) levels, helping to reduce system RFI and EMI. For example, a fast 100V, 1.5A octal SSR has a maximum 2 msec rise time and 9.5 msec fall time. The controlled version's specifications for these same parameters are 3 msec and 15 msec, about 1.5 times longer. The fast units have a maximum input supply current of 25 mA for a 1A output compared to a maximum of 15 mA for the controlled units.
A consequence of this controlled (i.e. slow) switching is that the MOSFET transitions its active region slowly. If turn-on transients are expected, for example due to a capacitive load, an analysis must be performed to ensure that the design does not violate the MOSFET safe operating area (SOA) limits (refer to sidebar). Fortunately, heater wires are mainly resistive, and therefore do not have a turn-on transient issue.
Surge Current Limiting
Solid-state devices are susceptible to failure with excessive current. One method of limiting the current surge uses a small resistor in series with the load. A second SSR can bypass the current limiting resistor if proper circuit operation does not permit leaving it in the circuit permanently. The second SSR turns on slightly delayed from the main load SSR. Some SSRs have a multi-purpose input circuit for automatically bypassing the current limiting resistor. The current limiting resistor for the second (bypass) SSR sets the input LED current in the 10 to 20 mA region.
Other applications may require a faster turn-on because of the presence of turn-on transients. Under these conditions, the following guidelines apply.
Use an external resistor to parallel the internal Rlimit in the SSR so that the nominal LED current is 2 to 3 times nominal, but no more than the maximum allowable per the device data sheet.
Use an external Rs and Cs to increase the peak current at turn-on to 100 mA (or as much as the specification allows). Rs sets the peak current, while Cs sets the pulse width, usually 1 msec.
Calculating the values of the components for turn on speed-up is straightforward. Using the RDHA702FT10A2FK as an example, the datasheet states that the internal LED diode drop is a nominal 2.6V at a 10 mA drive current. The Rlimit resistor sets this current, and is therefore equal to (5-2.6) V/10 mA, which is 240Ů. The data sheet also states that the maximum allowed LED direct current is 40 mA. In order to add an additional drive current of 30 mA, Rp calculates to be 2.4 V/30 mA, which is 80Ů.
The maximum peak LED current allowed per the datasheet is 100 mA for 1 msec. In order to add an additional peak current of 60 mA (so that the peak of 60 mA plus the dc level of 40 mA equals the maximum
Surge current limiting with automatic bypassing of a 10 V current limiting resistor and turn on speed-up circuitry.
of 100 mA), Rs calculates to be 2.4V/60 mA, which is 40Ů. Cs must set this 60 mA peak to have a time constant of no more than 1 msec. Since Cs sees basically just Rs for this time constant, Cs calculates to be 1 msec/40Ů, which is 25 µF.
At these higher currents, the LED voltage drop will be higher than the 2.6V used in the above calculations. This means the direct and alternating currents will not quite hit their mark. The designer has a choice here. Either recalculate the values based on actual voltage drop, or leave the values as-is since it will lead to a slightly more conservative design.
Designing for Survivability in a Radiation Environment
Any electronic components designated for operation in a radiation environment must be designed for graceful degradation. The idea is to build in enough performance to last through the expected mission lifetime.
For SSRs this starts with radiation hardened MOSFETs. For example, the MOSFETs in IR's SSRs can withstand 100 Krad (Si) Total Ionizing Dose (TID) and a Single Event Effect (SEE) of 37 MeV/ (mg/cm2).
The PVIs have similar ratings. However, PVIs exposed to radiation suffer some degradation. Some ways to minimize any negative effects on SSR performance include:
Run an LED current of at least 10 mA, and preferably more. Higher LED current increases the degradation capability.
Choose a PVI with two LEDs. This yields more room for degradation.
If possible, choose a slower PVI because they are simpler and therefore more immune to radiation than their faster counterparts.
With the proper design considerations, solid-state relays eliminate the inherent problems with electromechanical relays without creating new ones. For high reliability applications, the result is the improved reliability that electronic components have brought to other consumer, automotive, and industrial electronic products.
A DC SSR, IR’s RDHA710SE10A2QK, used in a solar array battery charger requires a diode to prevent reverse current flow and battery discharge in the off mode.
Alan Tasker is Engineering lead, Space Products, IR HiRel Products Group. You can reach him at firstname.lastname@example.org.
MOSFETs Make It or Break ItEMR contacts typically have only a few mŮ of resistance and subsequently very low power dissipation. MOSFETs in the SSR can easily have 10 times or greater drain-to-source resistance, or on-resistance. The resulting I2R losses (i.e. heat) must be removed by adequate heatsinking, especially at higher currents. Data and curves on the individual device datasheets aid the engineer in assessing the heatsinking needs for both steady state and transient heating to stay within the MOSFET's safe operating area (SOA) and below the maximum junction temperature. For steady state operation, the temperature rise, ÄT, is P x RčJA, where RčJA = RčJC + RčCS + RčSA. MOSFET packaging has a direct impact on the RčJC. For example, the typical RčJC value is 18 C/W for a 64-pin flat pack, 0.65 C/W for a 5-pin surface mount, and 1.7 C/W for an 8-pin surface mount. A power dissipation of 5W in the MOSFET in a 64-pin flat pack with RčJC =18 C/W produces a temperature rise of 90C. If the case can be held to 25C, the junction temperature of 115C is safely below the maximum operating temperature of 150C, but the added heat from any transients must be considered as well.
Truchard will be presented the award at the 2014 Golden Mousetrap Awards ceremony during the co-located events Pacific Design & Manufacturing, MD&M West, WestPack, PLASTEC West, Electronics West, ATX West, and AeroCon.
In a bid to boost the viability of lithium-based electric car batteries, a team at Lawrence Berkeley National Laboratory has developed a chemistry that could possibly double an EV’s driving range while cutting its battery cost in half.
For industrial control applications, or even a simple assembly line, that machine can go almost 24/7 without a break. But what happens when the task is a little more complex? That’s where the “smart” machine would come in. The smart machine is one that has some simple (or complex in some cases) processing capability to be able to adapt to changing conditions. Such machines are suited for a host of applications, including automotive, aerospace, defense, medical, computers and electronics, telecommunications, consumer goods, and so on. This discussion will examine what’s possible with smart machines, and what tradeoffs need to be made to implement such a solution.