Goobay 9579 Filament Lamp: Measure a Resistance That Changes with Power

The 6 V, 100 mA rating of this miniature lamp gives 60 Ω at the rated operating point. Use the same low-voltage fixture with the lamp and a 62 Ω resistor to see why that calculation gives only one point on the lamp's current–voltage curve. The lamp dissipates electrical power in a filament whose resistance changes as it warms.

Measure the voltage across the load and the voltage across a small series resistor. The series resistor is a shunt, which lets you infer current from its voltage drop. From four settled voltage pairs you can estimate the load's resistance, its electrical power and a slope between two points.

Wire the lamp and the current shunt

  1. Use the Goobay 9579 incandescent bulb. Wentronic lists a G11 glass envelope, a C-2R tungsten filament and a Miniature Edison Screw E10 base. The data sheet gives 6 V direct current (DC), 100 mA ±10% and 0.60 W. Fit it in an insulated E10 holder. Connect the holder center to LOAD+ and the shell to S. Use a 62 Ω, 1%, 2 W resistor as the comparison load between those same points.
  2. Use channel 1 of a SIGLENT SPD3303X bench supply in independent mode. Channel 1 positive connects to LOAD+. Its negative terminal is G0. Fit a 10 Ω, 1%, 0.5 W resistor from S to G0. Keep other supply outputs disconnected and protective earth intact. With the output disabled, set the current limit to 0.12 A. Start each run at low voltage and raise it in steps.
  3. Use two battery-operated Fluke 115 digital multimeters. Set both to their 6.000 V DC range, with red in the voltage input and black in COM. Meter A measures the load with red at LOAD+ and black at S. Meter B measures the shunt with red at S and black at G0. Attach the sense leads close to the resistor and holder terminals. Keep both meters floating. A shared grounded return at the black terminals would join S to G0 and bypass the shunt.
The isolated bench supply feeds an E10 lamp or a 62 ohm resistor and then a 10 ohm shunt. One floating voltmeter reads across the load from LOAD plus to S. The other reads from S to G0 across the shunt.
One load at a time between LOAD+ and S. The two battery meters have their black leads on different nodes. Editable SVG · Full-size PNG.

The highest supply setting is 5.50 V, with a 5.60 V acceptance limit for the sum of the two measured drops. The lamp voltage is lower because the shunt takes part of the supply voltage. At the nominal 0.12 A current setting, a nominal 10 Ω shunt would dissipate about 0.14 W. Including the SPD3303X specified current-setting error of 0.3% plus 10 mA and a 1% high shunt gives about 0.17 W as a steady-state bound. The 62 Ω comparison resistor would dissipate less than 0.52 W even if all 5.60 V were across its 1% low-tolerance value. Keep the bulb and resistor clear of the work surface and handle them after they cool.

Take four settled observations per load

  1. Check the wiring with the output off. Fit the lamp and set the supply to 0.20 V. Turn the output on. Check both voltage polarities and the supply's constant-voltage indicator. The supply should remain in voltage regulation after the reading settles. Stop if it remains in current limit. Check the shunt and meter wiring before raising the voltage.
  2. Wait at least 30 seconds at each setting. Compare a second pair of readings five seconds later. If either reading moves by more than its meter bound, wait and repeat. Once they settle, record the two readings in volts. Keep the orientation of the holder and airflow around it consistent.
  3. Raise the setting to 1.50 V, 3.50 V and 5.50 V in order. Repeat the settling check at each point. At 0.20 V the nominal available load power is at most about 1 mW, from VS² ÷ (4 × RSH). This small-excitation reading approximates the low-power operating point. The experiment does not give the resistance at exactly zero power.
  4. Disable the output, let the bulb cool, then replace the lamp with the 62 Ω control at the same terminals. Restart at 0.20 V and take the same four measurements. Record any supply current limiting or poor contacts before interpreting the data.

Use the blank worksheet and enter load_v from Meter A and shunt_v from Meter B. The fixture checks that the two voltages add to within 0.04 V of the 0.20 V setting or within 0.10 V of the higher settings. Those windows and the 1.24 V shunt ceiling screen this worksheet for setup errors. They do not represent the supply’s current-limit specification. A reading of the lamp across its holder includes the holder contact drops, which matter most at low voltage.

Separate operating resistance from slope

Compute I ≈ VSH ÷ RSH for every row. The estimate neglects the small currents through the meters' approximately 10 MΩ inputs. Then compute the operating resistance R = VLOAD ÷ I and the estimated electrical power P = VLOAD × I. This is the ratio at one settled point.

The finite slope between the 3.50 V and 5.50 V source settings is ΔVLOAD ÷ ΔI. It describes a different quantity. Each end of that segment can have a different filament temperature. The slope is a secant through two settled points, so it does not give the instantaneous small-signal resistance of a filament at a fixed temperature.

The ideal 62 Ω control in series with a nominal 10 Ω shunt draws 5.50 V ÷ 72 Ω ≈ 76.39 mA at the highest setting. That puts about 4.736 V across the control and 0.764 V across the shunt. At 0.20 V the corresponding nominal values are 2.78 mA, 0.172 V and 0.028 V. These are calculated reference values. Compare the control with your measurements before drawing conclusions from the lamp. The nominal 6 V and 100 mA imply 60 Ω at the rating, but the voltage across this fixture's lamp stays below 6 V.

Analyze the worksheet and its limits

Download the experiment ZIP with the full diagram, worksheet and code. Run python3 lamp_curve.py observations.csv on a computer. The default shunt is 10.0 Ω with a 1% bound. A verified measured value and its tolerance can be passed with --shunt-ohms and --shunt-error-percent. Keep the actual shunt calibration and temperature changes in your bench notes.

The Fluke 115 specifies ±(0.5% of the reading + 2 counts) for DC volts. On the 6 V range each count is 1 mV. The stated accuracy applies within one year of calibration, from 18 to 28 °C and at 0 to 90% relative humidity. For the operating resistance, the code combines each voltage bound with the worst end of the shunt bound. It reports an increase only when the high point's lower resistance bound exceeds the small-excitation point's upper bound. It reports an unresolved slope if the positive voltage or current change cannot be established from those bounds. These are conservative instrument and shunt intervals. Drift, contact changes and departures from voltage regulation need separate checks.

Complete Python worksheet analyzer
#!/usr/bin/env python3
"""Estimate an incandescent lamp's settled I–V curve from two DC voltages."""

import argparse
import csv
import json
import math
from pathlib import Path
import sys


FIELDS = ("device", "point", "load_v", "shunt_v")
DEVICES = ("lamp", "resistor")
RAILS = {"small": 0.20, "low": 1.50, "middle": 3.50, "high": 5.50}
KEYS = {(device, point) for device in DEVICES for point in RAILS}
EPSILON = 1e-12


def bounded(raw, lower, upper, name):
    try:
        value = float(raw)
    except (TypeError, ValueError):
        raise ValueError(f"{name}: enter a number") from None
    if not math.isfinite(value) or not lower <= value <= upper:
        raise ValueError(f"{name}: expected {lower:g} to {upper:g}")
    return value


def meter_bounds(volts):
    # Fluke 115 on the 6 V DC range: 0.5% of reading plus 2 × 1 mV.
    error = 0.005 * volts + 0.002
    return volts - error, volts + error


def read_point(key, values, shunt_ohms, resistor_error):
    _, point = key
    load_v = bounded(values[0], 0.001, 5.60, "load_v")
    shunt_v = bounded(values[1], 0.001, 1.24, "shunt_v")
    source_v = load_v + shunt_v
    allowable = 0.04 if point == "small" else 0.10
    if abs(source_v - RAILS[point]) > allowable + EPSILON:
        raise ValueError("load and shunt voltages do not add to the selected rail")
    load_low, load_high = meter_bounds(load_v)
    shunt_low, shunt_high = meter_bounds(shunt_v)
    if load_low <= 0 or shunt_low <= 0:
        raise ValueError("a voltage is too close to zero for this meter range")
    current = shunt_v / shunt_ohms
    current_low = shunt_low / (shunt_ohms * (1 + resistor_error))
    current_high = shunt_high / (shunt_ohms * (1 - resistor_error))
    return {
        "estimated_source_v": source_v,
        "load_v": load_v,
        "load_v_bounds": [load_low, load_high],
        "shunt_v": shunt_v,
        "estimated_current_a": current,
        "estimated_current_bounds_a": [current_low, current_high],
        "operating_resistance_ohms": load_v / current,
        "operating_resistance_bounds_ohms": [load_low / current_high, load_high / current_low],
        "load_power_w": load_v * current,
    }


def parse_worksheet(stream, shunt_ohms, resistor_error):
    reader = csv.reader(stream, strict=True)
    if tuple(next(reader, ())) != FIELDS:
        raise ValueError("use the exact worksheet headings")
    observations = {}
    for line, row in enumerate(reader, 2):
        key = tuple(row[:2])
        if len(row) != len(FIELDS) or key not in KEYS:
            raise ValueError(f"line {line}: invalid device, point or column count")
        if key in observations:
            raise ValueError(f"line {line}: repeated observation {key}")
        try:
            observations[key] = read_point(key, row[2:], shunt_ohms, resistor_error)
        except ValueError as error:
            raise ValueError(f"line {line}: {error}") from error
    if observations.keys() != KEYS:
        raise ValueError("need all eight observations")
    return observations


def segment_slope(left, right):
    i_left_low, i_left_high = left["estimated_current_bounds_a"]
    i_right_low, i_right_high = right["estimated_current_bounds_a"]
    v_left_low, v_left_high = left["load_v_bounds"]
    v_right_low, v_right_high = right["load_v_bounds"]
    delta_i_low = i_right_low - i_left_high
    delta_v_low = v_right_low - v_left_high
    if delta_i_low <= 0 or delta_v_low <= 0:
        return {"status": "positive slope unresolved", "estimate_ohms": None}
    delta_i = right["estimated_current_a"] - left["estimated_current_a"]
    delta_v = right["load_v"] - left["load_v"]
    return {
        "status": "positive slope resolved",
        "estimate_ohms": delta_v / delta_i,
        "bounds_ohms": [
            delta_v_low / (i_right_high - i_left_low),
            (v_right_high - v_left_low) / delta_i_low,
        ],
    }


def compare_endpoints(first, last):
    first_low, first_high = first["operating_resistance_bounds_ohms"]
    last_low, last_high = last["operating_resistance_bounds_ohms"]
    if last_low > first_high:
        return "increase resolved"
    if last_high < first_low:
        return "decrease resolved"
    return "change unresolved within supplied bounds"


def analyze(stream, shunt_ohms=10.0, shunt_error_percent=1.0):
    shunt_ohms = bounded(shunt_ohms, 9.90, 10.10, "shunt resistance in ohms")
    error = bounded(shunt_error_percent, 0.05, 2.0, "shunt error percent") / 100
    observations = parse_worksheet(stream, shunt_ohms, error)
    devices = {}
    for device in DEVICES:
        points = {point: observations[device, point] for point in RAILS}
        devices[device] = {
            "points": points,
            "operating_resistance_change": compare_endpoints(points["small"], points["high"]),
            "middle_to_high_segment_slope": segment_slope(points["middle"], points["high"]),
        }
    return {
        "assumed_shunt_ohms": shunt_ohms,
        "shunt_error_percent": shunt_error_percent,
        "devices": devices,
    }


def main():
    parser = argparse.ArgumentParser(description=__doc__)
    parser.add_argument("worksheet", type=Path)
    parser.add_argument("--shunt-ohms", type=float, default=10.0)
    parser.add_argument("--shunt-error-percent", type=float, default=1.0)
    args = parser.parse_args()
    try:
        with args.worksheet.open(encoding="utf-8-sig", newline="") as stream:
            result = analyze(stream, args.shunt_ohms, args.shunt_error_percent)
    except (OSError, ValueError, csv.Error) as error:
        print(f"Worksheet error: {error}", file=sys.stderr)
        return 2
    print(json.dumps(result, indent=2, allow_nan=False))
    return 0


if __name__ == "__main__":
    sys.exit(main())
Challenge. Two ratios agree at one point. What should you measure next?

Suppose a lamp and a resistor each have 50 Ω at one measured operating point. Do they therefore draw the same current at another voltage?

Take another settled voltage and current pair for each. A fixed resistor stays close to the original ratio under comparable conditions. The filament heats differently as power changes, so its ratio can move. A separate pair also lets you compare the slope over that interval.

The four-wire resistance experiment shows how moving the sense points changes what a low resistance reading includes. The analyzer's tests use synthetic readings. Enter your own readings in the blank sheet after running the fixture.

References