Instrumation is a Python Hardware Abstraction Layer (HAL) for RF test instruments — oscilloscopes, spectrum analyzers, signal generators, network analyzers, multimeters, power supplies, electronic loads, and frequency counters — built on PyVISA/SCPI, with a Digital Twin simulation mode for offline development.
21 instrument drivers across 8 categories · 8 supported brands (Keysight, Rigol, Siglent, Tektronix, Rohde & Schwarz, Anritsu, Keithley, TDK-Lambda) · 496 tests · 6 real-hardware validation reports · MIT licensed
Actively maintained — see release notes and the development log for the full history.
RF test bench automation is painful. Every instrument brand has its own quirks, SCPI dialects vary, and testing your scripts requires physical hardware on your desk. Instrumation fixes all three:
- Digital Twin mode — develop and debug offline with simulated instruments that emit realistic Gaussian noise
- Smart auto-detection — scans connected devices and loads the right driver automatically, no manual config
- Built-in CSV logging — test results logged out of the box, no extra dependencies
Instrumation has been validated against real lab hardware. See our experiment reports:
- Tektronix AFG3022C + Keysight DSOX2002A — Loopback Validation — plug-and-play AUTO discovery
- Keysight PXA N9030A — Spectrum Analyzer Validation — 32-bit binary trace transfers at high speed and comprehensive SCPI command coverage (29 driver methods)
- Keysight MXG N5183B — Signal Generator Validation — pulse modulation and frequency sweeps
- Keysight PNA N5232A — Network Analyzer Validation — S-parameter measurements and Smith charts
- Rigol DS1054Z — Oscilloscope Validation — 4-channel calibrated waveform readout and edge trigger
- TDK-Lambda Z+ — Power Supply Validation — handshake and setpoint validation
Programming a signal generator the traditional way vs. with Instrumation:
| Aspect | PyVISA (raw SCPI) | Instrumation |
|---|---|---|
| Discovery | Manual — find the resource string, manage ResourceManager |
Auto — just pass "AUTO" and the HAL scans USB + LAN for you |
| Connection | rm.open_resource("TCPIP0::192.168.1.100::...") — hardcoded address |
connect_instrument("AUTO", "SG") — type-aware routing |
| Configuration | sg.write(":FREQ:CW 2.4e9") — raw SCPI strings, no validation |
sg.set_frequency(2.4e9) — typed method with bounds checking |
| Cleanup | Manual sg.close() — easy to forget |
Context manager — automatic with block cleanup |
| Offline Dev | Requires real hardware connected | Digital Twin — set INSTRUMATION_MODE=SIM and develop anywhere |
| Portability | Vendor-specific SCPI — rewrite for each brand | One API — works on Keysight, Rigol, Tektronix, Siglent, R&S, Anritsu |
# ─── PyVISA: 8 lines of boilerplate ───
import pyvisa
rm = pyvisa.ResourceManager()
# Manually find the right resource...
sg = rm.open_resource("TCPIP0::192.168.1.100::inst0::INSTR")
sg.write("*RST")
sg.write(":FREQ:CW 2.4e9")
sg.write(":POW:AMPL -10")
sg.write(":OUTP ON")
sg.close()
# ─── Instrumation: 5 lines, zero config ───
from instrumation import connect_instrument
with connect_instrument("AUTO", "SG") as sg:
sg.set_frequency(2.4e9)
sg.set_amplitude(-10)
sg.set_output(True)No resource manager. No SCPI strings. No hardcoded addresses. Just your test logic.
Instrumation includes a built-in DataBroadcaster for streaming live readings over UDP. This allows you to build real-time dashboards or loggers with zero external dependencies.
- Non-blocking — UDP delivery runs independently of your test flow, so streaming telemetry never stalls a measurement.
- Zero-config — Broadcast to any host/port as JSON packets.
- Zero-dep — Built-in with Python standard library.
See examples/broadcast_demo.py and examples/dashboard.py for usage.
pip install instrumationOr install from source:
git clone https://github.com/abduznik/instrumation.git
cd instrumation
pip install .Windows users: You may need NI-VISA or Keysight IO Libraries Suite for physical hardware access.
If your packaged app starts but fails the moment it connects to an instrument, read this.
Instrumation loads part of its functionality at runtime instead of at import time:
pyvisa_py— the pure-Python VISA backend used by PyVISA when no vendor VISA library is installed.instrumation.drivers.*— device drivers discovered dynamically viaload_plugins()and lazy imports.
PyInstaller only bundles modules it can see by static analysis, so a default single-file build will be missing these modules. The .exe starts fine but fails as soon as it tries to connect, with errors like:
ModuleNotFoundError: No module named 'pyvisa_py'
AttributeError: 'NoneType' object has no attribute 'rsplit'
Include them explicitly at build time:
pyinstaller --onefile --windowed --name MyApp ^
--collect-all pyvisa_py ^
--hidden-import pyvisa_py.protocols.hislip ^
--hidden-import pyvisa_py.protocols.vxi11 ^
--collect-all instrumation ^
app.pyNote for Windows:
instrumationnever passesNonetopyvisa.ResourceManager(it passes an empty string), so PyVISA automatically selects system VISA if installed and falls back to the bundledpyvisa_pyotherwise.
import instrumation
sa = instrumation.connect_instrument("USB0::0x2A8D::...")
peak_power = sa.get_peak_value()
print(f"Peak Power: {peak_power} dBm")# Linux/macOS
export INSTRUMATION_MODE=SIM
# Windows PowerShell
$env:INSTRUMATION_MODE="SIM"from instrumation.factory import get_instrument
# Safer usage with context manager
with get_instrument("DUMMY_ADDRESS", "DMM") as dmm:
print(dmm.get_id())
result = dmm.measure_voltage()
print(f"Voltage: {result}")Manage complex test stations with multiple instruments using a station.toml file.
[instruments.sa_main]
driver = "SA"
address = "USB0::0x2A8D::0x0101::MY12345678::0::INSTR"
[instruments.psu_dut]
driver = "PSU"
address = "TCPIP0::192.168.1.100::inst0::INSTR"Use it in your code:
from instrumation import Station
station = Station("station.toml")
station.connect()
# Access instruments via dot notation
res = station.instr.sa_main.get_peak_value()
print(f"Peak: {res}")
station.disconnect()Instrumation comes with a powerful CLI for quick interaction and diagnostics.
# Scan for connected hardware
instrumation scan
# Take a quick measurement
instrumation measure USB0::... DMM measure_voltage
# List instruments in your station.toml
instrumation station list
# Measure using a named instrument from your station
instrumation station measure sa_main get_peak_value| Command | Description |
|---|---|
scan() |
Lists all connected Serial and VISA devices |
connect_instrument(addr) |
Connects to a specific instrument with auto driver detection |
| Platform | Status |
|---|---|
| Windows | Supported |
| Linux | Supported |
| Termux (Android) | Supported |
| macOS | Supported |
# Install in editable mode (required for tests to pick up local changes)
pip install -e .
# This project uses pyproject.toml (PEP 517/518). No setup.py is required.
# Install test dependencies
pip install pytest ruff
# Run tests (simulation mode)
export INSTRUMATION_MODE=SIM # Linux/macOS
pytestA lightweight Python library for generating structured UUT (Unit Under Test) test reports from measurement data. Works standalone or alongside Instrumation.
Features:
- HTML, Excel, and PDF report output
- Automatic pass/fail evaluation with configurable conditions
- Structured sections and test tables
- Easy integration with Instrumation measurements
pip install instrumation-reportfrom instrumation_report import Report, Section, TestTable, Measurement
report = Report(header=ReportHeader(
title="RF Subsystem Validation",
engineer="Yan",
uut_serial="SN-2024-001",
))
section = Section(number="1", title="Voltage Tests")
table = TestTable(title="Power Supply Checks", sub_number="1.1")
table.add(Measurement("3.3V Rail", 3.31, "V", condition=(3.2, 3.4)))
section.add_table(table)
report.add_section(section)
report.generate_html("report.html")
report.generate_excel("report.xlsx")- Language: Python 3.9+
- Libraries: PyVISA, PySerial
- Architecture: Smart Factory Pattern, Polymorphism
- Standards: SCPI (Standard Commands for Programmable Instruments)
Instrumation is maintained in my spare time alongside a full-time RF technician job. If it's saved you hours of boilerplate or made your test bench easier to automate, consider supporting:
Commercial support or custom driver development? Reach out via GitHub Issues or Discussions.
PRs and driver contributions are welcome. Open an issue first to discuss larger changes.
See LICENSE for details.