ResistaMet GUI

Open-source graphical interface for electrical characterization using Keithley 2400-family sourcemeters, with ASTM F84-aligned four-point-probe analysis.

Version: 1.8.2 Author: Brenden Ferland

More tabs (click to expand)

Tab	Screenshot
Resistance Measurement
Voltage Source
Current Source
4-Point Probe
I-V Sweep

Generated reproducibly via python tools/generate_screenshots.py — runs headless under Qt's offscreen platform, no instrument required.

Try it now (no instrument required)

git clone https://github.com/PEEKPerformer/ResistaMet-GUI.git
cd ResistaMet-GUI
pip install -e .
resistamet-gui --simulate

This launches the full GUI against an in-package Keithley 2400-family simulator. Every measurement mode works end-to-end with no NI-VISA, no pyvisa-py, and no GPIB hardware. The simulator is the same one tests validate byte-equivalent against captured hardware traces — see docs/sim_fidelity.md.

Optional: --sim-resistance 1000 (1 kΩ DUT) or --sim-model 2410 (advertise a different Keithley model).

Overview

ResistaMet GUI is a PyQt5 desktop application for controlling Keithley sourcemeters and performing electrical measurements. It supports five measurement modes (including hardware-driven I-V sweeps), real-time data visualization, multi-spot four-point-probe analysis with delta mode and ASTM F84-aligned correction factors, and dual-format data export.

Statement of need

The Keithley 2400/2450 family is standard equipment in materials, sensor, and device-physics labs. A typical day on one of these instruments combines long-duration fixed-bias logging with event annotations, periodic four-point-probe surveys, and routine 2-/4-wire resistance measurements — usually on the same instrument and the same sample. No single open application covers that combined workflow.

Researchers today either commit to a scripting framework (pymeasure, QCoDeS) and write Procedure subclasses, buy per-seat vendor software (KickStart, Windows-only, no four-point-probe analysis), or assemble narrow single-mode tools. ResistaMet GUI is for researchers who own a Keithley sourcemeter, do not write instrument-control code, and need every mode of that bench workflow in one point-and-click application.

Features

Measurement Modes

Mode	Sources	Measures	Use Case
Resistance	Current (up to 3A)	Resistance	2-wire/4-wire resistance measurement
Voltage Source	Voltage (-200 to +200V)	Current	Bias stress, I-V characterization
Current Source	Current (-3 to +3A)	Voltage	Material characterization
Four-Point Probe	Current	Voltage	Sheet resistance, resistivity, conductivity
I-V Sweep	Voltage or Current	Current or Voltage	Diode/device curves, hysteresis, breakdown
Van der Pauw	Current	Voltage	Sheet resistance + resistivity on arbitrary-shape samples (ASTM F76)

Four-Point Probe

• Sheet resistance (Rs), resistivity, and conductivity calculated in real time
• True 4-wire sense path — outer probes carry current, inner probes measure voltage via the Keithley's Sense terminals (:SYST:RSEN ON). Bench-verified against a Signatone S-302 probe head; see tests/hardware/rsen_diagnostic.py
• ASTM F84-02 correction factors — geometry-aware F₂ from Table 3 (circles) and the Smits 1958 table (squares, rectangles L/W ∈ {2, 3, 4}), thickness correction F(w/S) from Appendix X1.1 (valid out to w/S = 2.0), optional temperature correction F_T from Table 5 for n-/p-type silicon. Activated automatically when the user supplies a finite diameter D, non-circle geometry, or temperature + dopant
• Multi-spot tracking — save measurements at multiple probe positions, compare uniformity
• Live histogram of Rs distribution
• Current reversal (delta mode) — alternates +I/−I to cancel thermoelectric EMF; CSV exports include per-polarity V₊, V₋, R_f, R_r columns (F84 §13.1 diagnostic)
• Probe safety envelope — configurable warn / hard-stop power thresholds (default 10 mW / 100 mW). A pre-flight check refuses to start if the configured worst-case I_source × V_compliance exceeds the hard stop; a runtime check aborts the run and disables output if measured V × I exceeds it. Sized for tungsten-carbide tips (Signatone SP4 family) and conservative for thin-film / conductive-polymer samples where local Joule heating can damage the sample before the probe
• Legacy K · α · (V/I) path retained for non-Si materials and custom calibration factors (e.g. NIST-traceable reference standards)
• Inter-spot uniformity statistics in export

Van der Pauw

• ASTM F76-08 Method A sheet-resistance and resistivity for arbitrary-shape, hole-free samples with four periphery contacts (numbered 1–4 counter-clockwise)
• Walks the user through F76's 4 cabling geometries one at a time; current reversal (+I then −I) is automated at each geometry so thermal-EMF offsets cancel cleanly
• Solves F76 Fig. 5's implicit f(Q) equation (Q-1)/(Q+1) = (f/ln 2)·arccosh{(1/2)·exp(ln 2 / f)} numerically (hand-rolled bisection; no scipy dep)
• F76 §11.1 homogeneity gate — automatically flags samples where ρ_A and ρ_B disagree by more than 10%
• Per-geometry resistance bar chart visualizes uniformity; bars are color-coded by deviation from the mean (green/orange/red)
• Schematic diagram of the sample + lead routing updates as the user advances through the protocol; filmstrip preview of all 4 geometries makes the cyclic CCW rotation pattern obvious
• 37 unit tests pin values directly to F76 Section 11; bench-verified on a Keithley 2420 against a 100 µm conductive foil (1.4% asymmetry, F76 gate passes 7× over)

I-V Sweep

• Hardware staircase sweep using the Keithley sweep engine (precise inter-step timing via the instrument's trigger model)
• Source voltage or current with configurable start, stop, step, and per-step delay
• Sweep directions: up, down, or up-down (forward + reverse for hysteresis curves)
• Live X-Y I-V plot (separate canvas, not time-series)
• Per-point NPLC and compliance limits

Instrument Optimizations

• Hardware averaging filter -- repeat or moving average, 1-100 readings, runs on the Keithley itself (:SENS:AVER)
• Auto zero control -- on (accurate), once (fast), or off (fastest, drifts); ~3x speed boost in fast mode
• Offset-compensated ohms -- resistance mode option that cancels thermoelectric EMF in low-R DUTs
• Cable null -- one-button measure-and-subtract of cable/lead resistance (software-side reference)
• Auto source delay -- lets the instrument pick the optimal post-source settling time (:SOUR:DEL:AUTO ON)
• Non-concurrent functions -- :SENS:FUNC:CONC OFF for cleaner readings on the 2400 series
• High-impedance output-off -- :OUTP:SMOD HIMP protects the DUT when the output is disabled

Engineering Notation Input

Type natural lab notation instead of raw decimals:

• 1mA instead of 0.001000 A
• 100uA or 100uA instead of 0.000100 A
• 10mV instead of 0.010 V

The live readout displays in engineering notation too: V: 2.830 mV I: 1.000 mA R: 2.830 Ohm

Data Export

• Dual format -- JSON (with full metadata) + CSV (Excel-friendly) written simultaneously
• Crash recovery -- periodic checkpoints saved as .json.tmp, recoverable after power loss
• 4PP summary export -- per-spot breakdown with inter-spot uniformity RSD
• Configurable auto-save interval

Instrument Safety

• Compliance monitoring via Keithley STAT word bit 3 + threshold fallback
• Non-blocking compliance warnings (status bar flash, no modal popup spam)
• "Test Connection" button on every tab for pre-flight verification
• Configurable stop-on-compliance
• System sleep prevention during long measurements

UI

• Live numeric readout (large font) on all tabs
• Real-time matplotlib plots with interactive toolbar
• Resizable panels via splitters
• Tooltips on every setting explaining what it does and typical values
• Scroll-wheel protection on all spinboxes
• Tab switching allowed during measurement (read-only)
• "Run until stopped" checkbox on timed modes
• Custom event markers with text labels (press M)
• Multi-user profiles with per-user settings

Installation

Requirements

• Python 3.9+
• PyQt5, PyVISA, NumPy, Matplotlib (installed automatically)
• A VISA backend if you want to talk to real hardware (see below)

Setup

git clone https://github.com/PEEKPerformer/ResistaMet-GUI.git
cd ResistaMet-GUI
pip install -e .
resistamet-gui              # real instrument (needs VISA backend)
resistamet-gui --simulate   # in-package simulator, no hardware

pip install -e . reads pyproject.toml and registers the resistamet-gui console command. You can also run python resistamet-gui.py from the repo root.

VISA backend (real-hardware only)

If you launch without --simulate you need a VISA backend so PyVISA can reach the instrument:

• NI-VISA (download) — needed for NI GPIB adapters; Windows and Linux only. NI dropped macOS support after NI-VISA 18.5 (2020); on Apple Silicon the legacy installer will not run at all.
• pyvisa-py (pip install pyvisa-py) — pure-Python backend; the recommended path on macOS. Reportedly works with Prologix USB-GPIB adapters and serial sourcemeters, but I haven't tested either configuration here — only NI-VISA + a Keithley 2420 on Windows. Reports welcome.

If neither is installed and you're not using --simulate, Test Connection will fail with ValueError: Could not locate a VISA implementation.

Linux system packages (PyQt5 runtime)

Headless Linux distributions (and many CI images) don't ship the X11/Qt platform shared libraries that PyQt5 dynamically loads. On Debian/Ubuntu:

sudo apt-get install -y \
    libegl1 libxkbcommon-x11-0 libdbus-1-3 \
    libxcb-icccm4 libxcb-image0 libxcb-keysyms1 \
    libxcb-randr0 libxcb-render-util0 libxcb-shape0 \
    libxcb-sync1 libxcb-xfixes0 libxcb-xinerama0 libxcb-xkb1

(This is the same list our CI uses — see .github/workflows/test.yml.)

Quick Start

1. Launch and create a user profile
2. Click Test Connection to verify instrument communication
3. Enter a sample name
4. Set source level and compliance (type 1mA, 5V, etc.)
5. Click Start

Four-Point Probe Workflow

1. Set source current, probe spacing, and thickness
2. Start measurement -- readings appear in table with live histogram
3. Click Save Spot to archive current position's stats
4. Move probe to next position, repeat
5. After all spots: histogram shows bar chart of Rs uniformity
6. Click Export Summary for full report

Delta Mode (Thermoelectric Cancellation)

1. In the 4PP tab, expand Advanced
2. Check Current Reversal (Delta Mode)
3. Set settling time (default 0.1s between polarity flips)
4. Each reading now alternates +I/-I, reporting V_delta = (V+ - V-) / 2

Project Structure

• resistamet_gui/ — application package. Core modules: calculations.py (pure F84 / Smits formulas), instrument.py (Keithley VISA wrapper), workers.py (QThread measurement loop), data_export.py (dual JSON+CSV with crash recovery), _simulator.py (in-package Keithley simulator), and ui/ (main window, dialogs, canvas, widgets).
• tests/ — pytest suite. Unit/integration tests for every module, test_gui_smoke.py for Qt widget lifecycle, test_e2e_simulator.py for end-to-end UI-to-CSV pipeline, test_workers.py::TestSCPIContract for per-mode SCPI command assertions, tests/fakes/ (the fake instrument), tests/fixtures/scpi_traces*/ (captured hardware traces), tests/hardware/ (real-instrument tests gated by RESISTAMET_HARDWARE_ADDR).
• docs/sim_fidelity.md — simulator validation methodology and intentional gaps.
• scripts/community_capture.py — self-contained tool for contributors to capture SCPI traces from other 2400-family models.

Testing

# Unit + integration suite (pytest.ini ignores the e2e file by default)
QT_QPA_PLATFORM=offscreen pytest tests/ -v

# End-to-end suite (drives every tab through the in-package simulator,
# asserts recorded values against Ohm's law on a known fake DUT)
QT_QPA_PLATFORM=offscreen pytest tests/test_e2e_simulator.py -v

# Unit tests only (no Qt dependency)
pytest tests/ -v --ignore=tests/test_gui_smoke.py --ignore=tests/test_e2e_simulator.py

The e2e suite runs in its own pytest invocation because it leaves process-wide state (a pyvisa.ResourceManager monkey-patch and a live QApplication) that interacts poorly with module-scoped fixtures from earlier test files. CI runs both invocations in sequence — see .github/workflows/test.yml.

Instrument Compatibility

Hardware-validated (29 SCPI fixtures + 6 documented quirks, three DUT decades):

• Keithley 2420 (3 A model, firmware C30) — primary capture source
• Keithley 2400 (1 A model, firmware C30) — cross-model validation, 29/29 pass

The production code identifies the connected model from *IDN? against a static specification table covering the 2400 / 2401 / 2410 / 2420 / 2425 / 2430 / 2440 / 2450 family and surfaces the matching source/measure envelope at connect time. The 2400-family SCPI surface is largely uniform, so the other models in that table should work in principle, but only the 2400 and 2420 are hardware-validated in-house. If your model isn't yet captured, see "Help validate cross-model fidelity" below — that's how the table gets filled in honestly.

Help validate cross-model fidelity

If you have a different Keithley 2400-family instrument (2400/2401/2410/ 2420/2425/2430/2440/2450) and a precision reference resistor in 4-wire Kelvin (100 Ω, 10 kΩ, or 1 MΩ recommended), please consider running:

pip install pyvisa pyvisa-py
python scripts/community_capture.py

The script auto-detects the GPIB instrument, runs a polarity sanity check, captures a small set of representative SCPI traces, and prints instructions for opening an issue with the output zip attached. Each accepted submission becomes one row of cross-model validation in tests/fixtures/scpi_traces_community/ and runs in CI on every push, helping ensure the simulator faithfully reproduces every supported instrument. See docs/sim_fidelity.md for the full validation methodology.

Version History

See CHANGELOG.md for the full version history.

Citation

If you use ResistaMet GUI in your research, please cite:

Ferland, B. (2026). ResistaMet GUI: An Open-Source Electrical Measurement Suite
for Keithley Sourcemeters [Software]. Zenodo. https://doi.org/10.5281/zenodo.19919751

The DOI above is the concept DOI and always resolves to the latest archived version. Zenodo also provides a per-version DOI for citing a specific release. A CITATION.cff is included in the repository for machine-readable citation metadata.

Contributing

Contributions are welcome. Please see:

• CONTRIBUTING.md — bug reports, feature requests, development setup, and the cross-model trace contribution path.
• CODE_OF_CONDUCT.md — community guidelines.
• Issue templates — bug report, feature request, and cross-model SCPI-trace submission.

The fastest way to help right now is to run scripts/community_capture.py on any Keithley 2400-family instrument other than the 2400 and 2420 already validated in-house — see the "Help validate cross-model fidelity" section above.

License

MIT License — see LICENSE.md.