A 1:1 transformer has the same number of turns on its two windings. It can still deliver a different voltage when you add a load. Use a Triad TY-145P to measure the same low-voltage sine at 200 Hz, 1 kHz and 10 kHz, first with just a meter on the secondary and then with a 600 Ω resistor load. Measure both windings so you can tell how much of the change comes from the source delivering less voltage to the primary.
The primary is the driven winding and the secondary is the receiving winding. Their magnetic coupling carries the signal across separate windings without a wire between them. For the turns-ratio idea, start with What is a Transformer?. Here, the useful question is what happens when the receiving winding has to supply current.
Triad’s February 23, 2026 sheet labels both sides 600 Ω CT, where CT means center tap, an electrical connection partway through a winding. The 600 Ω primary impedance is specified to ±15% with a 600 Ω secondary load. That is alternating-current (AC) behavior. An unpowered resistance measurement of the copper windings should be in the neighborhood of the separate 44 Ω primary and 58 Ω secondary nominal values. Exact readings depend on the part and temperature. At 25 °C Triad lists a 1:1 turns ratio, 100 mW output and a frequency response within ±2 dB from 200 Hz to 15 kHz. This experiment uses three points in that band. The published response curve is a separate manufacturer result. Do not apply its specification to the meter-only condition or use these three points as a rated response test.
| Terminal | Connection |
|---|---|
| 1 | Primary high. Tektronix AFG1022 channel 1 (CH1) center conductor through 549 Ω, and the primary meter V input. |
| 2 | Primary center tap. Leave open. |
| 3 | Primary return. Generator coax shield and primary meter COM (common). |
| 4 | Secondary high. Secondary meter V input and one end of the optional load. |
| 5 | Secondary center tap. Leave open. |
| 6 | Secondary return. Secondary meter COM and the other end of the optional load. |
Find the dot above pin 1 on the transformer bobbin flange and use Triad’s mechanical drawing to match the physical pins. Check continuity on 1–3 and 4–6 with the transformer isolated from the rest of the circuit. A resistance reading between the two windings checks for an accidental connection at meter voltage only. Keep this a low-voltage signal experiment. Use no mains, insulation stress test, earth-grounded instrument or wired interface on the floating secondary. The AFG1022 output shield is chassis ground. Keep its protective earth connected, and keep the two winding returns separate.
Use a 549 Ω 1% resistor (at least ¼ W) between the AFG1022 center conductor and pin 1. The generator has a nominal 50 Ω output impedance, so the source seen from the fixture is about 599 Ω. Set its load display to High-Z, or high impedance. That setting changes the displayed amplitude while the generator keeps its physical 50 Ω output. Start with the output disabled, a sine wave, zero direct-current (DC) offset and 1.00 V RMS open circuit at most. RMS means root mean square, the AC voltage value that gives the same power as a DC voltage in a resistor. For a sine, 1.00 V RMS is about 2.83 V peak to peak. The nominal short-circuit current for this fixture is about 1.67 mA RMS. That is a design bound, not a test step.
Use two 300 Ω 1% resistors (¼ W each) in series for the load. Measure the disconnected pair with the meter in resistance mode and record its combined value. Use only this pair in the loaded run. Leave the secondary meter connected in both runs. Both Fluke 87V MAX meters must run from their own batteries, with leads in COM and the V jack. On the secondary, a bench scope ground clip or another earth-connected input can defeat the isolation even when the clip is called “ground.” Keep the secondary meter entirely floating. The meter-only condition still includes the meter input. Fluke specifies nominally 10 MΩ and less than 100 pF (picofarads) for AC volts. At 10 kHz, 100 pF would have a capacitive reactance magnitude of about 159 kΩ, much larger than the 600 Ω load, but finite.
meter_only in the worksheet.load_600. If you recalibrate the source setting or see drift, measure that row’s open baseline again with the primary temporarily disconnected. Always disable output before changing connections. Keep lead placement short and consistent.Fluke specifies the 600 mV and 6 V ranges through 20 kHz with larger count terms at 10 kHz. A count is one step of the last displayed digit. For example, in the 600.0 mV range at 10 kHz, a 0.4100 V reading has a stated allowance of roughly ±(2% of 0.4100 V + 20 × 0.0001 V), or ±0.0102 V, under the manual’s calibration and environmental conditions. The 6.000 V range uses 0.001 V per count and adds 12 counts below 10% of range at 5–20 kHz. A small apparent frequency trend can be smaller than the combined uncertainty of both meters. The supplied analyzer imposes a common 0.180 V minimum on winding readings so the input stays at or above 3% of either range. If a reading is lower, log it separately as out of this worksheet’s scope and diagnose the setup. Do not substitute a floor reading.
Compare Vp/Vopen to see the source divider change. Then compare Vs/Vp to see the transfer between transformer terminals. Their product is Vs/Vopen, the whole fixture’s voltage ratio. Convert a voltage ratio r to decibels (dB), a logarithmic comparison, using 20 log10(r). For the loaded row only, calculate resistor power as Vs2/Rload watts. The meter’s reading is across the load, so the calculation does not require a current measurement.
A first-pass model at midband treats the transformer as an ideal 1:1 pair and adds the nominal 44 Ω and 58 Ω winding resistances in series. With 600 Ω drive resistance, 600 Ω load and 1.000 V RMS open circuit, the model gives 1.000 × (600 + 44 + 58)/(600 + 44 + 58 + 600) = 0.539 V at the primary and 1.000 × 600/1302 = 0.461 V across the load. That is about 0.354 mW. Compare it with an ideal 1:1 transformer with zero winding resistance, where the same external source and load give 0.500 V on both windings. The real transformer also has current that builds the core’s magnetic field, some field that does not couple the windings, and stray capacitance between nearby conductors. This simple model leaves those effects out. These amplitude readings alone cannot determine phase, complex impedance or why a particular end of the band changes.
Start with the blank CSV worksheet. CSV means comma-separated values. Enter volts, not millivolts, and keep one row for each of the six condition/frequency pairs. Run python3 transformer_load.py observations.csv --load-ohm 600, substituting the measured load resistance. The result includes both raw ratios, changes on adding the load, resistor power and each condition’s overall voltage response relative to its own 1 kHz result. The input rejects missing pairs, duplicate rows and values outside the worksheet limits. Download the complete reader experiment for the program, blank data, constructed example, figures and command-line tests. The example is synthetic and is suitable for testing the workflow. No physical transformer performance has been measured for this article.
"""Analyze paired TY-145P readings. All input voltages are measured RMS."""
import argparse
import csv
import json
import math
from pathlib import Path
FREQUENCIES = (200, 1000, 10000)
CONDITIONS = ("meter_only", "load_600")
HEADER = ("condition", "frequency_hz", "source_open_vrms", "primary_vrms", "secondary_vrms")
def bounded(value, low, high, label):
try:
value = float(value)
except ValueError:
raise ValueError(f"{label} must be a number") from None
if not math.isfinite(value) or not low <= value <= high:
raise ValueError(f"{label} outside finite range {low:g} to {high:g}")
return value
def load_csv(path):
readings = {}
with path.open(encoding="utf-8-sig", newline="") as source:
reader = csv.reader(source, strict=True)
if tuple(next(reader, ())) != HEADER:
raise ValueError("CSV header must be " + ",".join(HEADER))
for row in reader:
line = reader.line_num
if not row:
continue
if len(row) != len(HEADER):
raise ValueError(f"line {line}: expected five fields")
condition, frequency_text, *voltages = row
if condition not in CONDITIONS:
raise ValueError(f"line {line}: unexpected condition {condition!r}")
if frequency_text not in tuple(map(str, FREQUENCIES)):
raise ValueError(f"line {line}: frequency must be 200, 1000 or 10000 Hz")
key = (condition, int(frequency_text))
if key in readings:
raise ValueError(f"line {line}: duplicate {condition}/{frequency_text}")
readings[key] = tuple(
bounded(token, low, high, f"line {line}: {name}")
for token, name, (low, high) in zip(
voltages, HEADER[2:], ((0.2, 1.0), (0.18, 1.2), (0.18, 1.2))
)
)
missing = [(condition, hz) for condition in CONDITIONS for hz in FREQUENCIES
if (condition, hz) not in readings]
if missing:
raise ValueError(f"missing readings: {missing}")
return readings
def db(ratio):
return 20 * math.log10(ratio)
def analyze(readings, load_ohm):
results = []
for hz in FREQUENCIES:
point = {"frequency_hz": hz}
for condition in CONDITIONS:
source, primary, secondary = readings[condition, hz]
mid_source, _, mid_secondary = readings[condition, 1000]
end_to_end = secondary / source
point[condition] = {
"source_open_vrms": source,
"primary_vrms": primary,
"secondary_vrms": secondary,
"source_to_primary_ratio": primary / source,
"terminal_ratio": secondary / primary,
"terminal_db": db(secondary / primary),
"end_to_end_ratio": end_to_end,
"end_to_end_db": db(end_to_end),
"relative_to_1khz_end_to_end_db": db(end_to_end / (mid_secondary / mid_source)),
"load_mw": 1000 * secondary ** 2 / load_ohm if condition == "load_600" else None,
}
for ratio_name in ("terminal", "end_to_end"):
field = f"{ratio_name}_db"
point[f"loaded_minus_meter_only_{field}"] = (
point["load_600"][field] - point["meter_only"][field]
)
results.append(point)
return {"load_ohm": load_ohm, "points": results}
def main():
parser = argparse.ArgumentParser(description=__doc__)
parser.add_argument("csv", type=Path, help="six paired true-RMS readings")
parser.add_argument("--load-ohm", default="600", help="measured secondary resistance in ohms")
args = parser.parse_args()
try:
load = bounded(args.load_ohm, 588, 612, "load resistance")
readings = load_csv(args.csv)
except (OSError, UnicodeError, csv.Error, ValueError) as error:
parser.error(str(error))
print(json.dumps(analyze(readings, load), indent=2, allow_nan=False))
if __name__ == "__main__":
main()
At 1 kHz, suppose the open-source voltage is 1.000 V. Your meter-only readings are 0.960 V primary and 0.950 V secondary. With 600 Ω fitted you read 0.540 V primary and 0.460 V secondary. Where did most of the overall level change occur?
The source-to-primary ratio went from 0.960 to 0.540, a change of about −5.00 dB. The terminal ratio went from 0.950/0.960 to 0.460/0.540, about −1.30 dB of additional change. Most of the overall −6.30 dB comes from the primary receiving less voltage. The loaded resistor receives about 0.353 mW.