Contactors and Power Supplies

Many high voltage DC applications require switching to disconnect a power supply or load from a shared bus, manage inrush current, or provide galvanic isolation. However, DC switching is not so straightforward as the hardware store light switch. The simple binary control input of a high voltage DC relay belies many nuanced considerations such as arc management, contact limitations, speed, and test integration. Choosing a practical relay or solid state switching solution requires detailed understanding and planning in all these areas to ensure dependable operation.

Switches within a feedback loop must be carefully managed so that they do not interfere with regulation. Simple static errors, like leaving a relay open, will break the feedback path. If the connection is closed later by an automated test program, the DUT will be connected to an unregulated source well above its setpoint. Similarly, dynamic conditions such as contact bounce will break the feedback path and must be allowed to settle before enabling the power supply. 

AC relay specifications do not imply similar DC performance. Contactors used with DC power supplies or loads must be specifically rated for DC operation. An electric arc is drawn between the contacts when they open under load. This arc is a major wear mechanism and reduces practical device lifetimes to much less than the purely mechanical lifetime listed on some datasheets. In ordinary AC contactor, the duration of the arc is limited by periodic zero crossings of the AC waveform. Using an AC contactor to break a DC load risks fire or other catastrophic damage. With no zero crossing to assist in the switching process, DC arcs can persist indefinitely, welding the contacts or destroying the relay.

If galvanic isolation is not required, solid state relays (SSR) reduce or eliminate many of the difficulties associated with conventional relays or contactors.  With no moving parts, SSRs are much faster than mechanical switches. Although there are no contact arcing concerns, commercial DC SSRs are still different from their AC counterparts. Some AC designs employ thyristors as the switching element. They latch on after a trigger pulse, but only shut off when the load current falls below the thyristor latching threshold – typically at a zero crossing in the AC cycle. DC SSRs use a continuous signal to enable FETs or IGBTs but might not be set up to block reverse current.

While it is possible to build a custom SSR, practical considerations extend well beyond simply choosing an appropriate IGBT or FET. Blocking reverse current requires two anti-series power devices doubling cost, losses, and heat sink volume. The metal backplate on many transistor packages is electrically connected and will require isolation if the heat sink is grounded or shared by other devices.

High power SSRs should be recognized as requiring substantial mechanical engineering effort in addition to the electrical development. Power dissipation from the IGBT saturation voltage, FET channel resistance (Rds_on), or diode forward voltage will usually necessitate heat sinking. The heat sinks are likely to dominate the overall weight and volume of the SSR solution and increase BOM and assembly costs.

Keysight RP7900 and MP4300 regenerative power supplies and loads include an internal SSR for every output. Timing and control of the SSR is seamlessly integrated into the standard SCPI commands and DUT protection features such as overvoltage or overcurrent detection. It operates automatically without additional programming effort from the user. If mechanical relays are desired, Keysight N6700, E4300, RP7900, MP4300 lines offer a 8 pin Digital IO port. Pin state, direction and polarity can be assigned via simple SCPI commands. Digital outputs can control large relays with an inexpensive logic level FET driver and a diode to prevent inductive kickback. Digital inputs can provide feedback on relay armature positions by using auxiliary contacts