A programmable DC power supply is sized by three hard limits: maximum output voltage, maximum output current, and the resulting power envelope in watts, with a 20% derating margin above the worst-case load to absorb inrush and future expansion [S2][S3].
Beyond the electrical envelope, selection turns on four cross-cutting decisions: topology (linear versus switching), control mode (constant-voltage versus constant-current priority), remote-interface protocol (SCPI over USB, LAN, GPIB or RS-232), and protection suite (OVP, OCP, OTP) [S1][S4][S6].
Defining the Operating Envelope: Voltage, Current, and Power
Total system power consumption plus 20% overhead is the standard sizing rule, then verify the efficiency curve holds at the typical operating load rather than only at full scale [S2]. Bench-class supplies cluster at 96 W to 192 W (e.g. 20 V/5 A, 32 V/6 A, 60 V/3.2 A, 80 V/13.5 A), while mid-power programmable units step to 360 W to 1080 W in the same voltage class [S5]. Higher-voltage autoranging chassis reach 250 V at 4.5 A to 13.5 A and 800 V at 1.44 A to 4.32 A within a 360 W to 1080 W frame, useful where the DUT swings from low-voltage logic to bus rails in one test [S5]. A 1U/2U/3U half- or full-rack mechanical footprint is the practical differentiator when multiple channels must share a bench; the 2230-30-6 triple-output 180 W chassis (30 V/6 A on two channels plus 5 V/3 A) targets that exact bench profile [S5].
For a deeper treatment of the broader DC power supply category, including regulated versus unregulated variants and ripple behavior, the reference page lays the groundwork. The same 20% overhead rule applies whether the load is a production tester or a single bench setup; under-sizing is the single most common field failure because designers compute average draw and forget motor-start or capacitor-charge peaks [S2][S3].
Topology Choice: Linear, Switching, or Hybrid
Switching supplies win on efficiency, weight, and power density and are the right pick for battery charging, power backup, and any portable or compact installation [S1]. Linear supplies win on ripple and noise: the 2280S-series at 32 V/6 A and 60 V/3.2 A delivers low-noise linear output with 10 nA to 3.2 A or 6 A high-accuracy measurement resolution and 140 µs dynamic-load capture, the kind of spec that matters for sensitive analog and reference design [S5].
For most ATE racks and lab benches, a switching power supply autoranging chassis is the default because the same 360 W or 720 W frame can cover 30 V/36 A and 80 V/13.5 A without manual range switching [S5]. Linear is reserved for low-noise, high-precision work; switching for everything else [S1].
Control Modes and Programming Behavior
Constant-voltage (CV) and constant-current (CC) are the two fundamental operating modes, and programmable supplies add a third axis: timing, sequencing, and readback through SCPI commands over a host computer [S4][S6]. A "constant-current priority" mode, available on the 2260B-series, lets the supply behave as a true current source with fast CC-loop response, a configuration used for laser-diode constant-current sourcing and power-LED binning [S5].
Programmable rise and fall times, list-mode sequencing, and battery-simulation profiles are now baseline features on mid-range chassis: the 2281S-20-6 stores up to 14 battery models and graphs state-of-charge against open-circuit voltage and internal resistance, which turns the supply into a battery emulator for PMIC and power-management tests [S5]. For a wider view of how programmable sources compare to general power supply categories, including AC and bidirectional variants, the encyclopedia entry walks through the topology matrix.
Interface, Software, and Rack Integration
USB and LAN are the dominant modern interfaces, with GPIB and RS-232 still common on legacy ATE; SCPI is the de facto command language across vendors, which means a test sequence written for one programmable supply usually ports to another with only minor syntax edits [S6][S7]. A 0.1 mA measurement resolution, remote sense to compensate for lead drop, and KickStart or vendor equivalent software support are the entry-level marks of a serious bench supply [S5].
LAN-connected chassis slot cleanly into 19-inch racks at 2U half-rack (e.g. 2380-150-15, 2380-120-60) or 3U full-rack widths, and the same chassis can sit beside a programmable electronic load (Series 2380) to form a closed-loop source-and-sink test bed [S5]. The industrial UPS reference page covers the closely related problem of ride-through and hold-up time, which matters when the programmable supply itself sits on a backed-up test line.
Protection, Safety, and Environmental Limits
Over-voltage protection (OVP), over-current protection (OCP), and over-temperature protection (OTP) are non-negotiable; remote sense is required whenever the lead resistance can shift the programmed voltage by more than the load's tolerance [S5][S7]. Operating temperature, humidity, dust, and moisture ingress ratings determine whether a supply belongs on the bench, on the production floor, or in a sealed enclosure [S2].
Input range selection matters at installation: single-phase 115 VAC (90-138 VAC) or 230 VAC (180-264 VAC), three-phase 208 VAC or 400-480 VAC, and DC-input variants at 12 VDC, 24 VDC, or 28 VDC for vehicle and military platforms; mismatch causes PSU failure, so input tolerance is the first thing to verify in the datasheet [S2]. For a more detailed view of how the DC-DC converter family differs from a bench programmable supply, the encyclopedia contrasts the two duty cycles.
Selection Shortlist by Use Case
For a R&D bench doing general analog and digital bring-up, a 100-192 W linear or autoranging switching chassis (e.g. 2200-30-5 at 30 V/5 A, 2280S-32-6 at 32 V/6 A) is the entry tier [S5]. For automated test of power-LEDs, laser diodes, or PMIC battery emulation, step up to a 360 W to 720 W chassis with constant-current priority, battery simulation, and programmable rise/fall times (2260B-30-36, 2260B-80-13, 2281S-20-6) [S5]. For automotive circuit, power-supmit, and reliability tests, a triple-output 90 W to 180 W chassis with independent isolated channels (2230-30-3, 2230-30-6) covers most rails in one box [S5].
Do not pick a linear supply for a high-power ATE rack: efficiency and heat dissipation will kill the budget. Do not pick a switching supply for sub-milliamp reference design: ripple and noise will dominate the measurement floor. Match topology to the noise floor first, then size watts, then pick the interface, then verify protection [S1][S5].
Comparison: Main Programmable DC Supply Classes
Bench linear (45-192 W): low noise, high accuracy, single-channel, USB/LAN/GPIB; best for R&D analog and reference work [S5]. Bench autoranging switching (360-1080 W): higher efficiency, wider V/I range, multi-interface, constant-current priority option; best for ATE and power-device test [S5]. Triple-output isolated (45-180 W per chassis): three independent rails, channel-tracking, series/parallel modes; best for automotive and appliance circuit test [S5]. Battery-emulation programmable (120 W): 14 stored battery models, SOC graphing, PMIC test mode; best for power-management and battery-powered device test [S5].
Across these four classes, decision criteria cluster on noise floor (linear lowest), power density (switching highest), channel count (triple-output only), and application-specific emulation (battery model only on the dedicated emulator class) [S5].
For a broader view of how programmable power fits into automated test alongside spectrum and signal analysis tools, the How to Choose a Spectrum Analyzer: Spec-Anchored Selection Map article applies the same envelope-plus-interface methodology. To verify the bench can deliver clean power to RF test lines, the How to Size and Select a Spectrum Analyzer: Spec Map and Selection Logic reference complements this guide.
Verifiable next nodes: confirm the input VAC range on the installation site matches the supply's spec (115 VAC, 230 VAC, 400-480 VAC three-phase) [S2]; pull the latest SCPI command reference and verify LAN/LXI compliance before any ATE integration [S6][S7]; check OVP/OCP trip accuracy against the DUT's absolute-maximum ratings on the bench before first power-up [S5].