A 0.01% (100 ppm) decade resistance box carries ten times the relative accuracy of a 0.1% (1000 ppm) unit, but that single spec line hides meaningful differences in switch architecture, resistor technology, and price band [S1][S2].
Decade boxes are passive substitution devices that synthesise resistance from series-connected internal resistor decades, typically one to eight decades, selected by rotary or push-button switches [S2]. They are widely used for calibration of process instruments (especially Pt100/RTD simulators), verification of multimeter ohms ranges, and rapid prototyping in circuit design [S1][S2].
What the Accuracy Class Actually Means
Accuracy on a resistance decade is specified as a percentage of the dialed value, plus in some cases a fixed floor resistance in ohms [S1][S3]. The IET Labs PRS-330 programmable resistance box, for example, is rated at 70 ppm + 1 mΩ (lab grade) [S1], meaning a 100 Ω setting would be accurate to within ±(0.007 + 0.001) Ω = ±0.008 Ω, while a 1 MΩ setting would be accurate to ±(70 + 0.001) Ω ≈ ±70 Ω. By contrast, a 0.1% class box at 1 MΩ would tolerate ±1000 Ω, more than 14x the PRS-330 error budget at that point [S1].
On mid-range commercial units, the relative error scales with the dialed value: 0.1% on a 1000 Ω step equals ±1 Ω, whereas 0.01% equals ±0.1 Ω, a difference that matters when calibrating 6.5-digit or 8.5-digit DMMs (Keysight 3458A, Fluke 8588A) [S1][S3]. At the low end, the absolute error of a 0.01% specification on a 1 Ω step is only ±100 µΩ, which approaches the contact-resistance uncertainty of the switch stack itself [S8].
Resistor Technology and Stability Behind Each Class
The 0.01% tier is built on hermetically sealed wirewound or precision metal-film resistors with temperature coefficients typically under 5 ppm/°C, often aged and screened for long-term stability [S1]. IET Labs specifies 20 ppm/year stability on its VRS-100 series high-resistance standards, and the Time Electronics 1067 quotes better than 20 ppm/year [S1][S2].
The 0.1% tier usually uses standard metal-film resistors in open-frame or enclosed housings; the Time Electronics 1061 cites "stable metal-film resistors" without naming a stability figure [S2]. IET's RS-Series substitution box spans 0.1% to 1% accuracy across its model range, with 0.01 Ω to 100 MΩ coverage and 36 GenRad 1433 variants for the precision tier [S1]. At the 1% educational level, the GME RD-300 and similar 7-decade substitution boxes (1 Ω to 11 MΩ) use 1 W carbon-film or metal-film resistors and are explicitly designed for student labs and field substitution, not calibration [S5].
Range, Resolution, and Decade Count
Accuracy class alone is meaningless without a matching range and resolution. The Time Electronics 1067 precision decade covers 10 mΩ to 12 kΩ with 6-digit resolution at 0.01% accuracy, ideal for low-resistance and Pt100 simulation work [S2]. The IET GenRad 1433 spans 1 mΩ to 111.11111111 MΩ across 36 models with 3 to 11 dials, and the HARS series offers single through 11-decade configurations at 0.01% or 0.1% depending on model [S1].
Wide-range commercial units like the Transmille 2090 cover 0.1 Ω to 99.9999 kΩ with 0.1% basic accuracy and an optional 0.01% upgrade path [S4]. At the extreme high end, the IET HRRS-5 kV/10 kV series reaches 1 TΩ to 11 TΩ with 5000 V or 10,000 V maximum voltage for megohmmeter calibration [S1]. Resolution per decade varies: a 0.01% spec on a 0.1 Ω step delivers 10 µΩ steps, but switch contact resistance and thermal EMF often dominate the measurement before the resistor tolerance does [S1][S8].
Decision Matrix: Which Class Fits the Job
Selection boils down to four criteria: required accuracy, resistance range, switching method, and budget. [S1]
0.01% (100 ppm) class resistance decade boxes are used for verifying the accuracy of measuring devices and test equipment including multimeters, and for Pt100 simulation to calibrate RTD transmitters and resistance temperature devices [S1][S2]. Price band: typically $1500-$6000 USD for a programmable or wide-range unit, with manual rotary units somewhat lower [S1].
0.1% (1000 ppm) class suits: production-line functional test of analog inputs; field substitution for instrument loops; electronics-design prototyping; general troubleshooting. Price band: $300-$1500 USD depending on decade count and switch quality [S1][S4]. The IET HARS-X2 double-power version and Transmille 2090 both sit in this band with selectable accuracy tiers [S1][S4].
1% (10,000 ppm) class suits: educational labs, student experiments, and any use where the decade box replaces a handful of discrete resistors rather than functioning as a calibration reference. Price band: under $100 USD for the GME RD-300 or HEALLILY 0-9999 Ω substitution box [S5]. A 1% box on a 1 kΩ step has ±10 Ω error, which is acceptable for circuit behaviour studies but disqualifies it for instrument calibration [S5].
Failure Modes and Practical Constraints
Switch contact resistance is the dominant error source below 10 Ω, regardless of the resistor tolerance class [S1][S8]. A rotary switch with silver-alloy contacts (IET HARS) adds a few milliohms of contact resistance that drifts with wear and oxidation; for 0.01% accuracy below 1 Ω, four-terminal (Kelvin) connections are mandatory [S1]. The IET PRS-330 explicitly notes that its 1 mΩ floor term reflects switch and lead resistance rather than resistor tolerance [S1].
Temperature coefficient drives long-term drift. A 0.01% resistor with a 5 ppm/°C TC, operating across a 10 °C lab swing, contributes 50 ppm of error, half of the accuracy budget, before any other factor [S1]. High-power units like the IET HPRS series (225 W/decade, 250 W max) self-heat during use and can drift several hundred ppm unless duty-cycled [S1]. For RTD simulation, an active heat-soak or thermal enclosure is recommended when chasing 0.01% at sub-1 Ω settings [S1][S2].
Switch wear and oxidation cause long-term drift. The IET HARS series cites "long-life switches with silver-alloy contacts" for the 0.01% tier [S1]; cheaper 1% educational boxes use plated-brass or copper-alloy switches that oxidise within a few years of bench use [S5]. The Time Electronics 1067 quotes better than 20 ppm/year stability, indicating that the manufacturer has measured and warranted the drift over a full year, not just at the time of shipment [S2].
Standards, Calibration, and Interfacing
No single ISO or IEC standard defines decade-box accuracy classes the way IEC 61010 defines multimeter safety. Traceability is established by calibration against laboratory standards traceable to national metrology institutes (NIST, NPL, PTB), typically using a substitution method where the decade box under test is compared step-by-step against a reference standard resistor of known uncertainty [S1][S2]. The calibration interval is usually 12 months for 0.01% units and 24-36 months for 0.1% units, though high-humidity or high-vibration environments shorten this [S1].
For Pt100 simulation specifically, the Time Electronics 1049 and 1050 Pt100 simulators are based on ITS-90 EN 60751, the platinum resistance thermometer standard, and exceed class A tolerance with ±0.3 °C accuracy across –200 °C to +800 °C [S2]. A 0.01% decade box can achieve better than ±0.1 °C equivalent at 0 °C (100 Ω) when used with low-thermal-EMF connections, making it the preferred manual substitute for a dry-block calibrator in field work [S1][S2]. Programmable units like the IET PRS-330 add USB, Ethernet, and IEEE-488 (GPIB) interfaces for automated test, an important consideration when a RTD transmitter needs to be swept across its full range in a production ATE setup [S1].
Who 0.01% Is For, and Who It Is Not For
The 0.01% class is the right tool for calibration laboratories, national-metrology-institute traceable measurements, 8.5-digit DMM verification on the 1 Ω to 10 kΩ ranges, and Pt100/RTD simulation in metrology contexts. It is overkill for circuit-design prototyping, field troubleshooting of 4-20 mA loops, and any application where an insulation resistance tester is being checked at the 1% level rather than the 0.01% level. Engineers specifying decade boxes for a general-purpose decade resistance box application should match the accuracy class to the DMM or calibrator under test; specifying a 0.01% box to verify a 1% handheld meter wastes budget, while specifying a 0.1% box to verify an 8.5-digit reference produces out-of-tolerance readings. [S2]
The 0.1% class covers the bulk of industrial and educational substitution work and remains the workhorse tier for most procurement decisions. Specifiers should request the per-decade accuracy (not just the headline percentage), the long-term stability in ppm/year, the switch contact material, and the operating temperature range. For specialised metrology work, the comparison between 0.01% and 0.1% boxes parallels the broader trade-off in precision instrumentation selection: pay ten times the unit cost to cut measurement uncertainty by ten times, but only when the downstream process actually requires that uncertainty.
For related coverage, see One-Channel vs Four-Channel Controller and Probe: Real Cost Split.