---
Title: Behringer UFO202 – measured with AudioLab
---

# Behringer UFO202

A USB audio card with a stereo input switchable between PHONO (with RIAA correction) and LINE, a stereo output and a headphone output: a simple way to connect a turntable to a computer.

*How to measure it yourself, step by step: [the guide](step-usb-card.html). All protocols: [Measured devices](measured.html).*

![Illustration: the Behringer UFO202 – ground terminal, input L and R, output L and R, USB and headphone output with its volume](measured/behringer-ufo202/setup/device-illustration-studio.jpg)

| | |
| --- | --- |
| **Device** | Behringer UFO202 (Windows: "Microphone (USB Audio CODEC)") |
| **Measured** | 2–5 November 2026 |
| **AudioLab** | development build before 0.3.5 |
| **Interface** | MOTU M4, ASIO driver (plays the test signal) |
| **Wiring** | External ADC (input), External DAC (output) |
| **Sample rate** | 48 kHz |

## Summary

| What | Result | Note |
| --- | --- | --- |
| Line input, full scale | 1.45 V | clips between 1.23 and 1.55 V (02) |
| Line input, THD at 997 Hz, 245 mV | **0.0016 %** (−95.7 dB) | above 0.4 V the right channel rises to 0.06 % (02) |
| Line input, S/N | **90.5 dB(A)** | 87.6 dB unweighted (04) |
| Line input, response 20 Hz – 20 kHz | within 0.9 dB | +0.5 dB in the bass, −0.4 dB at 20 kHz (03) |
| Line input, crosstalk | −79.6 dB at 1 kHz | −53 dB at 20 Hz, −64 dB at 20 kHz (05) |
| Phono input, deviation from RIAA | −6.3 dB at 20 Hz, +1.3 dB at 10–20 kHz | relative to 1 kHz (06) |
| Phono input, full scale | 31 mV at 1 kHz, 19 mV at 315 Hz | (06, 08) |
| Phono input, S/N against 5 mV at 1 kHz | **67 dB(A)** | (09) |
| Output, full scale | 1.17 V | at 0 dBFS the card clips the sine digitally, THD+N −65 dB (11) |
| Output, THD+N at 997 Hz | **0.0074 %** (−82.6 dB) at 0.59 V | 0.029 % at −20 dBFS, set by the noise (10, 11) |
| Output, S/N | **95.0 dB(A)** | 91.3 dB unweighted (13) |
| Output, response 20 Hz – 20 kHz | within 0.3 dB | (12) |
| Output, crosstalk | −68.5 dB at 1 kHz | −37 dB at 20 Hz (14) |
| Headphone output, at maximum volume | 0.44 V open at full scale | clips above −6 dBFS open (15, 16) |
| Headphone output, impedance | **50 Ω** | −8.2 dB of level into 32 Ω (17) |
| Headphone output, into 32 Ω | **0.15 mW** clean, 0.0031 % THD | clips above −8 dBFS (18) |
| Headphone output, response into 32 Ω | −3 dB at 20 Hz | +0.4 dB at 16 kHz (19) |
| Headphone output, S/N into 32 Ω | 75.7 dB(A) | against the highest clean level, 68 mV (20) |
| Headphone output, crosstalk into 32 Ω | −33.5 dB | flat over the band (21) |

The line input and the output are clean and flat: distortion in thousandths of a percent, the noise within 7–8 dB of what 16 bits allow, the response within a dB. The phono input takes any MM cartridge without clipping, the tracking tracks of a test record included, and is quiet enough for it; it does not follow the RIAA curve closely, though: the bass falls away below 100 Hz and the treble lies about 1 dB high, so a record sounds a little thinner and brighter than it was cut. The output's channel separation is weak in the bass. The headphone output is the weak part: an output impedance of 50 Ω, 0.15 mW into 32 Ω before it clips, the deep bass 3 dB down and 33 dB of channel separation – clean below its limit, but meant for sensitive headphones of higher impedance.

## How it was measured

**The inputs (01–09), wiring External ADC.** AudioLab plays the test signal through the MOTU M4 (ASIO); its OUT 3 and OUT 4 feed the card's INPUT L and R, and the card records over USB as a Windows device ("Microphone (USB Audio CODEC)", WASAPI, exclusive mode). For LINE the cables go straight; for PHONO through a −40 dB attenuator, so that the interface's smallest clean levels become a cartridge's millivolts. The levels at the phono input are the generator's, less the attenuator's −40 dB. The Windows input volume is set to 0 dB: on this card it is a gain in the card, and at 100 it added about 30 dB.

![The signal path: AudioLab plays the test signal through the MOTU M4 (ASIO), OUT 3 and OUT 4 feed the card's input L and R – through a −40 dB attenuator for PHONO, straight for LINE – and the card records over USB, through a USB isolator, as a Windows device (WASAPI, exclusive mode); two clocks, so no reference](measured/behringer-ufo202/setup/wiring-external-adc.svg)

**The output (10–14), wiring External DAC.** AudioLab plays through the card as a Windows playback device ("Speakers (USB Audio CODEC)", exclusive mode); the card's OUTPUT L and R go to the MOTU M4's line inputs In 3 and In 4, calibrated in volts. The Windows playback volume is at 100 (0 dB), the audio enhancements are off and the card's input is left open: at volume 54 the output played 5 dB lower, and with the phono input connected it was 20 dB noisier (10).

![The signal path for the output: AudioLab plays the test signal through the card over USB, through a USB isolator (WASAPI, exclusive mode), the card's OUTPUT L and R feed the MOTU M4's inputs 3 and 4, and the interface records over ASIO; two clocks, so no reference](measured/behringer-ufo202/setup/wiring-external-dac.svg)

**The headphone output (15–21)** is wired as the output, the 3.5 mm jack through a stereo-to-two-mono cable to In 3 (tip, L) and In 4 (ring, R), the headphone volume at maximum. For the load, 32 Ω resistors from L and from R to ground, measured with an ohmmeter (L 31.9 Ω, R 31.4 Ω) and entered in AudioLab, both channels loaded at once.

**For all:**

- **A USB isolator** between the computer and the card. The card and the interface share their ground over the cables and the computer; without the isolator the loop added mains hum and noise – 11 dB on the line input, 27 dB on phono – and false THD. The earlier runs without it are kept as the "b" files.
- **997 Hz instead of 1 kHz** for distortion, as AES17 recommends. The card has lines of its own at whole kilohertz, where a 1 kHz tone's harmonics fall; at 997 Hz the harmonics lie beside them.
- **48 kHz**, the MOTU M4's ASIO buffer at 1024 samples (at 128 the recording had gaps), the M4's generator below 0 dBFS.
- AudioLab watches every block for gaps in the signal and measures a step again when it finds one.

What the setup itself limits: the card and the interface run on two clocks, so there is no reference – no phase, and Gain & phase is off; the response comes from Stepped sine, the input's levels in the card's own dBFS. The output's noise is read through the M4's inputs; AudioLab did not find it limited by their floor.

## Measurements

### 01 – THD+N at 997 Hz, line input

**Why:** the first reading of the card's input: how loud a known voltage comes in, and how clean it is.

**Settings:** THD+N meter · 997 Hz (996.83 Hz, the nearest bin) · generator −20 dBFS = 245 mV from the MOTU M4 (outputs 3–4, calibrated) · UFO202 switch LINE · Windows input volume 0 dB · the card through a USB isolator · 48 kHz · FFT 65536, Blackman-Harris, 8 averages · 20 Hz – 21.6 kHz, unweighted.

![THD+N meter, line input, 245 mV at 997 Hz: the tone at −15.5 dBFS in both channels, H2 near −116 dBFS and the other harmonics at the noise, a spur near 215 Hz at −99 dBFS, the noise floor near −130 dBFS per bin](measured/behringer-ufo202/screenshots/01-thd-n-line.png)

**Result:**

| | L | R |
| --- | --- | --- |
| Tone in the card | −15.45 dBFS | −15.45 dBFS |
| THD | **0.0016 % (−95.7 dB)** | 0.0015 % (−96.6 dB) |
| THD+N | 0.025 % (−72.0 dB) | 0.025 % (−72.0 dB) |
| H2 / H3 | −100.5 / −106.9 dBc | −101.5 / −108.6 dBc |
| Noise | −87.5 dBFS | −87.5 dBFS |

245 mV reads −15.5 dBFS, so the line input reaches full scale at about **1.45 V** (+3.3 dBV); the maker gives 2 dBV as the maximum input level. The distortion is very low; THD+N is set by the noise. The two channels, each with its MOTU output, match within 0.01 dB. · Data: [01-thd-n-line.csv](measured/behringer-ufo202/csv/01-thd-n-line.csv)

**Why 997 Hz and an isolator.** A first reading at 1 kHz, with the card and the interface on the same computer, gave 0.030 % THD and the noise at −78 dBFS ([01c](measured/behringer-ufo202/csv/01c-thd-n-line-1khz-ground-loop.csv)). The card has lines of its own at whole kilohertz (04), where a 1 kHz tone's harmonics fall, and they count as distortion; at 997 Hz the harmonics lie beside them. The ground the two USB devices share over the cables added hum and noise; a USB isolator between the computer and the card removes it. All the input and output measurements here are made so.

### 02 – THD vs level at 997 Hz, line input

**Why:** how the distortion changes with the level, and where the input clips – its full scale in volts.

**Settings:** THD vs level · 997 Hz · generator −60 to −2 dBFS in 2 dB steps = 2.45 mV to 1.95 V (output limit raised for this sweep) · LINE · Windows input volume 0 dB · USB isolator · 2 averages · FFT 65536, Blackman-Harris · MOTU buffer 1024 samples.

![THD and THD+N of both channels against the voltage at the card's input, 2.5 mV to 2 V: THD falling to −98 dB (left) and −95 dB (right) at 0.3 V; above 0.4 V the right channel rises to −63 dB while the left stays near −80 dB; both clip above 1.23 V](measured/behringer-ufo202/screenshots/02-thd-vs-level-line.png)

**Result:**

| Input | In the card | THD L / R | THD+N L / R |
| --- | --- | --- | --- |
| 24.5 mV | −35.5 dBFS | −77.0 / −70.0 dB | −52.1 / −52.0 dB |
| 245 mV | −15.5 dBFS | −96.4 / −95.0 dB | −72.0 / −72.0 dB |
| 309 mV | −13.5 dBFS | **−97.9 / −95.3 dB** | −74.0 / −73.9 dB |
| 490 mV | −9.5 dBFS | −87.1 / **−66.3 dB** | −77.6 / −66.0 dB |
| 776 mV | −5.5 dBFS | −79.1 / −64.6 dB | −77.2 / −64.5 dB |
| 1.23 V | −1.5 dBFS | −82.5 / −69.1 dB | **−80.7** / −68.9 dB |
| 1.55 V | +0.4 dBFS (clipped) | −32.4 / −32.4 dB | −32.3 / −32.3 dB |

- **0 dBFS = 1.45 V** at the line input; it clips between 1.23 V and 1.55 V.
- Up to about 0.3 V the distortion is very low: **0.0013 % THD (−98 dB)** in the left channel. The noise stays at −87.5 dBFS throughout, so THD+N falls 2 dB for every 2 dB of level, to **0.0093 % (−80.7 dB)** at 1.23 V in the left channel.
- Above about 0.4 V the **right channel** distorts far more than the left: up to 0.06 % (−63 dB) between 0.5 and 1 V, against 0.01–0.04 % in the left. Swapping the cables at the card's input left it in the right channel, so it is the card's.

Data: [02-thd-vs-level-line.csv](measured/behringer-ufo202/csv/02-thd-vs-level-line.csv)

### 03 – Frequency response, line input

**Why:** how evenly the card records the band from 20 Hz to 20 kHz.

**Settings:** Stepped sine · 31 frequencies 20 Hz – 20 kHz · generator −20 dBFS = 245 mV · LINE · Windows input volume 0 dB · 2 averages, settle 750 ms (the card's delay of 250 ms measured and added). The gain is the tone in the card against the generator level: the card runs on its own clock, so there is no reference and no phase.

![Gain of both channels from 20 Hz to 20 kHz: 5.03 dB at 20 Hz falling to 4.55 dB by 300 Hz, flat to 2 kHz, a ripple of a few hundredths of a dB between 5 and 10 kHz, 4.14 dB (L) and 4.25 dB (R) at 20 kHz](measured/behringer-ufo202/screenshots/03-stepped-response-line.png)

**Result:**

| | L | R |
| --- | --- | --- |
| Gain at 1 kHz | +4.55 dB | +4.55 dB |
| 20.5 Hz | +0.48 dB | +0.48 dB |
| 20 kHz | −0.40 dB | −0.31 dB |
| 20 Hz – 20 kHz within | 0.90 dB | 0.79 dB |

The response is even: a rise of half a dB towards the bass, below about 300 Hz, and a fall of 0.3–0.4 dB at 20 kHz, with a small ripple from the converter's filter between 5 and 10 kHz. No −3 dB point within the band. The channels agree within 0.01 dB up to 2 kHz and within 0.1 dB at 20 kHz. · Data: [03-stepped-response-line.csv](measured/behringer-ufo202/csv/03-stepped-response-line.csv)

### 04 – Noise and S/N, line input

**Why:** the noise of the card's input on its own – the floor every recording through it lies on.

**Settings:** Noise & S/N · generator silent, the card's input on the MOTU M4's outputs (a low source impedance) · LINE · Windows input volume 0 dB · USB isolator · band 20 Hz – 21.6 kHz · unweighted, A-weighted and ITU-R 468 · FFT 65536, 2 blocks averaged.

![Noise density of both channels from 10 Hz to 20 kHz: flat near −130 dBFS/√Hz, a single spur near 220 Hz at −102 dBFS/√Hz, small lines at 2 and 7 kHz; no mains hum](measured/behringer-ufo202/screenshots/04-noise-line.png)

**Result:**

| | L | R |
| --- | --- | --- |
| Noise, unweighted | −87.6 dBFS | −87.6 dBFS |
| Noise, A-weighted | −90.5 dBFS | −90.4 dBFS |
| Noise, ITU-R 468 | −81.2 dBFS | −81.1 dBFS |
| S/N against full scale (1.45 V) | 87.6 dB, **90.5 dB(A)** | 87.6 dB, 90.4 dB(A) |

With full scale at 1.45 V the noise is about 61 µV unweighted. A 16-bit converter alone would allow about 98 dB; the input comes within 8 dB of it. The spectrum is flat, without mains hum; the spur near 220 Hz and a small bump near 7 kHz came with the USB isolator and were not there without it. Both channels are alike.

Without the isolator the noise was 11 dB higher, −76.7 dBFS (79.9 dB(A) S/N), with mains hum at 50 Hz and its harmonics and a row of lines at whole kilohertz. · Data: [04-noise-line.csv](measured/behringer-ufo202/csv/04-noise-line.csv), without the isolator: [04b-noise-line-ground-loop.csv](measured/behringer-ufo202/csv/04b-noise-line-ground-loop.csv)

### 05 – Crosstalk, line input

**Why:** how much of one channel leaks into the other – the stereo separation of the input.

**Settings:** Crosstalk · one channel driven at a time · 31 frequencies 20 Hz – 20 kHz · generator −20 dBFS = 245 mV · LINE · Windows input volume 0 dB · USB isolator · the undriven input on the MOTU M4's silent output.

![Crosstalk L→R and R→L from 20 Hz to 20 kHz: −53 dB at 20 Hz, improving 6 dB per octave to −80 dB at 1 kHz, rising again to −64 dB at 20 kHz; both directions alike](measured/behringer-ufo202/screenshots/05-crosstalk-line.png)

**Result:**

| | L → R | R → L |
| --- | --- | --- |
| 20 Hz | −52.7 dB | −52.6 dB |
| 100 Hz | −65.0 dB | −64.9 dB |
| 1 kHz | **−79.6 dB** | **−79.3 dB** |
| 10 kHz | −69.0 dB | −69.3 dB |
| 20 kHz | −63.6 dB | −64.4 dB |

The separation is best around 1 kHz, about 80 dB, and falls by 6 dB per octave towards both ends of the band: 53 dB at 20 Hz, 64 dB at 20 kHz. Both directions are alike. The maker's −77 dB at 1 kHz is met.

Without the isolator the 1 kHz reading was hidden by the card's line at 1 kHz (only an upper bound, −75 dB) and the treble read up to 6 dB better (−70 dB at 20 kHz); the bass was the same. · Data: [05-crosstalk-line.csv](measured/behringer-ufo202/csv/05-crosstalk-line.csv), without the isolator: [05b-crosstalk-line-ground-loop.csv](measured/behringer-ufo202/csv/05b-crosstalk-line-ground-loop.csv)

### 06 – RIAA accuracy, phono input

**Why:** how closely the phono input follows the RIAA playback curve – how a record will sound tonally through it.

**Settings:** Stepped sine · 31 frequencies 20 Hz – 20 kHz · generator −30 dBFS through a −40 dB attenuator = 0.78 mV at the card's input, the same at every frequency · switch PHONO · Windows input volume 0 dB · ANALYSIS → Compare with phono curve: RIAA, so the chart shows the deviation from the ideal curve, 0 dB at 1 kHz.

![Deviation from RIAA of both channels from 20 Hz to 20 kHz, 0 dB at 1 kHz: −6.3 dB at 20 Hz, −1.5 to −1.9 dB at 100 Hz, rising through 0 at 1 kHz to +1.3 dB above 10 kHz; the right channel up to 0.4 dB below the left in the bass, the two alike above 1 kHz](measured/behringer-ufo202/screenshots/06-riaa-phono.png)

**Result** (deviation from RIAA, relative to 1 kHz):

| | L | R |
| --- | --- | --- |
| 20 Hz | **−6.3 dB** | −6.4 dB |
| 50 Hz | −2.6 dB | −3.0 dB |
| 100 Hz | −1.5 dB | −1.9 dB |
| 2 kHz | +0.3 dB | +0.3 dB |
| 10 kHz | +1.3 dB | +1.3 dB |
| 20 kHz | **+1.3 dB** | +1.4 dB |

The phono input does not follow RIAA closely: the response tilts over the whole band, about 3 dB from 100 Hz to 10 kHz, and the bass falls away below 100 Hz, 6.3 dB down at 20 Hz. A record played through it sounds thinner in the bass and a little brighter than it was cut. Above 1 kHz both channels agree within 0.05 dB; in the bass the right one lies up to 0.4 dB below the left.

At 1 kHz the input reaches full scale at about **31 mV**, some 33 dB more gain than the line input: an MM cartridge giving 5 mV at 1 kHz plays at about −16 dBFS.

Two runs agree within 0.1 dB except at 50 Hz (0.2 dB): there the tone shares its frequency with the mains hum, which the phono input raises with the bass. At a quarter of the level (0.25 mV) the 50 Hz step varied by up to 0.9 dB between runs, the 2 kHz step by 0.5 dB, where the card has a line of its own (04). · Data: [06-riaa-phono.csv](measured/behringer-ufo202/csv/06-riaa-phono.csv)

### 07 – THD vs level at 997 Hz, phono input

**Why:** how much signal the phono input takes before it distorts, and how clean it is at a cartridge's levels.

**Settings:** THD vs level · 997 Hz · generator −40 to −2 dBFS in 2 dB steps through the −40 dB attenuator = 0.25 to 19.5 mV at the card's input · PHONO · Windows input volume 0 dB · distortion limit 1 % · USB isolator.

![THD and THD+N of both channels against the voltage at the phono input, 0.25 to 19.5 mV: THD falling to −92 dB at 6–10 mV and rising slightly to −87 dB at 19.5 mV; THD+N falling 2 dB for every 2 dB of level to −66 dB](measured/behringer-ufo202/screenshots/07-thd-vs-level-phono.png)

**Result:**

| Input | In the card | THD L / R | THD+N L / R |
| --- | --- | --- | --- |
| 0.78 mV | −32.2 dBFS | −75.9 / −77.3 dB | −38.7 / −38.6 dB |
| 2.45 mV | −22.2 dBFS | −85.8 / −86.5 dB | −47.9 / −48.5 dB |
| 7.8 mV | −12.2 dBFS | **−91.6 / −91.6 dB** | −58.6 / −58.4 dB |
| 19.5 mV | −4.2 dBFS | −86.5 / −87.6 dB | **−66.6 / −65.9 dB** |

- The input takes **24.5 mV at 997 Hz without clipping** (full scale about 31 mV, see 06): five times what an MM cartridge gives at 5 cm/s.
- The distortion is very low: **0.0026 % THD (−92 dB)** at 6–10 mV, rising a little towards full scale. THD+N is set by the noise (−71 dBFS) all the way: 0.05 % at 19.5 mV, 0.12 % at 7.8 mV.

**How the setup changed the reading.** At 1 kHz through the isolator the same input read 0.005 % THD ([07c](measured/behringer-ufo202/csv/07c-thd-vs-level-phono-1khz-isolator.csv)); at 1 kHz without it, 0.33 % ([07b](measured/behringer-ufo202/csv/07b-thd-vs-level-phono-ground-loop.csv)): mains hum and the card's lines at whole kilohertz, raised by the ground loop and the phono gain, fall on the harmonics of a 1 kHz tone and count as distortion. · Data: [07-thd-vs-level-phono.csv](measured/behringer-ufo202/csv/07-thd-vs-level-phono.csv)

### 08 – THD vs level at 315 Hz, phono input

**Why:** the tracking tracks of a test record are cut at 315 Hz, where the RIAA correction already raises the gain. Whether the phono input takes them without clipping decides whether the card serves to set up a turntable.

**Settings:** THD vs level · 315 Hz · generator −40 to 0 dBFS in 2 dB steps through the −40 dB attenuator = 0.25 to 24.5 mV at the card's input (output limit raised to 0 dBFS) · PHONO · Windows input volume 0 dB.

![THD and THD+N of both channels against the voltage at the phono input at 315 Hz, 0.25 to 24.5 mV: THD falling to −65 dB at 6 mV and rising slowly to −57 dB at 15.5 mV; above that the left channel clips at 19.5 mV, the right one at 24.5 mV](measured/behringer-ufo202/screenshots/08-thd-vs-level-phono-315.png)

**Result:**

| Input at 315 Hz | In the card L / R | THD L / R |
| --- | --- | --- |
| 2.45 mV | −17.7 / −18.1 dBFS | −52.4 / −52.1 dB |
| 6.2 mV | −9.7 / −10.1 dBFS | **−64.9 / −64.5 dB** |
| 15.5 mV | −1.7 / −2.1 dBFS | −57.0 / −56.8 dB |
| 19.5 mV | +0.3 dBFS, clipped / −0.1 dBFS | −37.4 / −55.6 dB |
| 24.5 mV | clipped / clipped | −19.7 / −21.5 dB |

- At 315 Hz the input reaches full scale at about **19 mV**: the left channel clips at 19.5 mV, the right one just above. At 1 kHz it was 31 mV; the RIAA correction gives 315 Hz about 4 dB more gain here.
- Below that the distortion stays low, **0.06 % THD at 6 mV**, 0.14 % at 15.5 mV – lower than at 997 Hz (07).

**For the tracking tracks:** a track of 100 µm at 315 Hz moves the stylus at about 14 cm/s in each channel. A cartridge rated 4 mV at 5 cm/s then gives about 11 mV, 4–5 dB below the card's clipping; one rated 6 mV gives about 17 mV, just below it. A typical MM cartridge reads the Ortofon tracking tracks through the card unclipped; with a hot one the 90 and 100 µm tracks come close to the limit, and AudioLab marks a track that clips the input.

Data: [08-thd-vs-level-phono-315.csv](measured/behringer-ufo202/csv/08-thd-vs-level-phono-315.csv)

### 09 – Noise, phono input

**Why:** the noise and hum of the phono input on its own, against what a cartridge gives.

**Settings:** Noise & S/N · the card's PHONO input closed by the attenuator's output resistance, **the interface disconnected** · Windows input volume 0 dB · band 20 Hz – 21.6 kHz · unweighted, A-weighted and ITU-R 468.

![Noise density of both channels from 10 Hz to 20 kHz: about −90 dBFS/√Hz below 50 Hz, falling with the RIAA correction to −135 dBFS/√Hz at 20 kHz; no mains hum, single lines at 1, 2 and 3 kHz](measured/behringer-ufo202/screenshots/09-noise-phono.png)

**Result:**

| | L | R |
| --- | --- | --- |
| Noise, unweighted | −73.3 dBFS | −73.9 dBFS |
| Noise, A-weighted | −83.4 dBFS | −84.0 dBFS |
| Noise, ITU-R 468 | −77.9 dBFS | −78.4 dBFS |
| S/N against 5 mV at 1 kHz (−16 dBFS) | 57 dB, **67 dB(A)** | 58 dB, 68 dB(A) |

The noise falls with frequency as the RIAA correction shapes it: the input's own noise raised in the bass and cut in the treble. No hum is left, only the card's lines at 1, 2 and 3 kHz. With an MM cartridge giving 5 mV at 1 kHz the input plays 67 dB(A) above its noise.

**With the interface connected** through the attenuator (all other phono measurements) the noise was 27 dB higher, −46.6 dBFS unweighted, −52.6 dBFS A-weighted: mains hum at 50 and 60 Hz and their harmonics, and the card's lines at whole kilohertz, raised by the phono gain. The hum comes in through the ground the two USB devices share over the attenuator – the setup's: with the card alone (above) it is not there. Without it, measurement 07 read 36 dB more THD at 1 kHz: the hum's and the lines' harmonics fall on the tone's. A USB isolator between the computer and the card removes it with the interface still connected (07 was measured so, at 997 Hz). · Data: [09-noise-phono.csv](measured/behringer-ufo202/csv/09-noise-phono.csv), with the interface connected: [09a-noise-phono.csv](measured/behringer-ufo202/csv/09a-noise-phono.csv)

### 10 – THD+N at 997 Hz, line output

**Why:** the card's output, the converter that plays: its level for a known digital signal and how clean it is.

**Settings:** THD+N meter · wiring External DAC: AudioLab plays through the card as a Windows playback device (exclusive mode, Windows volume 100), the card's OUTPUT L and R go to the MOTU M4's line inputs In 3 and In 4 (calibrated) · the card's input left open · USB isolator · 997 Hz at −20 dBFS · 48 kHz · FFT 65536, Blackman-Harris, 8 averages.

![THD+N meter, line output, 997 Hz at −20 dBFS: the tone at 0.118 V in both channels, odd harmonics H3, H5, H7 and H9 between −120 and −130 dBFS, the noise floor near −150 dBFS per bin, a trace of 50 Hz in the right channel](measured/behringer-ufo202/screenshots/10-thd-n-output.png)

**Result:**

| | L | R |
| --- | --- | --- |
| Output at −20 dBFS | 0.1181 V | 0.1181 V |
| THD | 0.0082 % (−81.7 dB) | 0.0082 % (−81.7 dB) |
| THD+N | 0.029 % (−70.7 dB) | 0.029 % (−70.8 dB) |
| H2 / H3 / H5 | −102.4 / −83.9 / −88.5 dBc | −105.9 / −84.0 / −88.6 dBc |
| Noise | 0.033 mV | 0.033 mV |

−20 dBFS gives 0.118 V, so full scale is about **1.18 V** (+1.4 dBV); the maker gives 2 dBV (1.26 V) as the maximum output level. The distortion is almost all odd – H3, then H5, H7, H9 – while H2 lies 20 dB lower. THD+N is set by the noise, 0.033 mV. Both channels agree within 0.03 dB. The tone arrives 31 ppm high: the card's clock runs that much faster than the interface's – relative to each other; which of the two is nearer the true 48 kHz is not known.

**What else changed the output.** A first run on this day, with the Windows playback volume at 54, audio enhancements on and the phono input still connected through the attenuator, read the output 5 dB lower (0.066 V at −20 dBFS, full scale 0.66 V) and the noise 20 dB higher, −87 dBFS, rising towards the bass like the phono input's noise and with mains hum at 50 Hz. With the volume at 100, the enhancements off and the input disconnected, both were gone. The 5 dB match the volume (it acts in the card, in exclusive mode too, as on the input); which of the changes removed the noise was not tried apart – its shape points to the connected phono input. The output is measured at volume 100, enhancements off, the input open. · Earlier run: [diag-output-thdn-volume-or-input.csv](measured/behringer-ufo202/csv/diag-output-thdn-volume-or-input.csv) · Data: [10-thd-n-output.csv](measured/behringer-ufo202/csv/10-thd-n-output.csv)

### 11 – THD vs level at 997 Hz, line output

**Why:** how the output's distortion changes with the level, up to digital full scale.

**Settings:** THD vs level · 997 Hz · −60 to 0 dBFS in 2 dB steps into the card (output limit raised to 0 dBFS for this sweep) · External DAC, OUTPUT → M4 In 3 / In 4 (calibrated) · Windows playback volume 100, enhancements off, the card's input open · USB isolator · 2 averages.

![THD and THD+N of both channels against the digital level: THD+N falling 2 dB for every 2 dB to −82.6 dB at −6 dBFS, THD falling to −89.8 dB at −10 dBFS and rising again to −78 dB at full scale; at 0 dBFS THD+N jumps back to −65 dB](measured/behringer-ufo202/screenshots/11-thd-vs-level-output.png)

**Result:**

| Digital level | Output L / R | THD L / R | THD+N L / R |
| --- | --- | --- | --- |
| −60 dBFS | 1.16 / 1.16 mV | −40.1 / −40.3 dB | −29.7 / −30.1 dB |
| −40 dBFS | 11.8 / 11.8 mV | −60.3 / −60.8 dB | −49.9 / −50.3 dB |
| −20 dBFS | 118 / 118 mV | −81.8 / −81.5 dB | −70.7 / −70.8 dB |
| −10 dBFS | 373 / 373 mV | **−89.8 / −88.9 dB** | −80.5 / −80.5 dB |
| −6 dBFS | 0.592 / 0.592 V | −86.3 / −85.5 dB | **−82.6 / −82.2 dB** |
| 0 dBFS | 1.170 / 1.170 V | −78.7 / −78.2 dB | −65.0 / −65.0 dB |

- **Full scale gives 1.17 V** (+1.4 dBV). The analog stage does not clip; at 0 dBFS the card clips the sine slightly in its digital path (below), and the level rises 1.92 dB from −2 to 0 dBFS instead of 2. The maker gives 2 dBV (1.26 V).
- The lowest THD, **0.0032 % (−89.8 dB)**, is at −10 dBFS (0.37 V); towards full scale it rises again to 0.012 %, mostly H3.
- THD+N follows the noise (−107 dBFS at the interface input, 0.033 mV) up to −6 dBFS: best **0.0074 % (−82.6 dB)** at 0.59 V. At 0 dBFS the residual without the harmonics rises by 25 dB at once (−106.6 to −81.4 dBFS) while the THD stays where it was. A full-scale signal is the one level to keep away from; see below.
- The two channels agree within 0.02 dB in level.

**What happens at 0 dBFS.** A THD+N reading at 0 dBFS shows a row of lines that are not harmonics of the tone: the largest at 844.5 Hz (−68 dBc), others at 1149 Hz, 152 Hz and spaced across the band.

![THD+N meter, line output at 0 dBFS, 997 Hz: besides the harmonics H2–H10, a row of lines at 844 Hz (−84 dBFS), 1149 Hz, 690 Hz, 535 Hz and further, spread over the whole band at −100 to −130 dBFS](measured/behringer-ufo202/screenshots/11a-thd-n-output-0dbfs.png)

They lie exactly where the tone's high harmonics fold back at a 48 kHz sample rate: 49 × 996.83 Hz = 48 845 Hz appears at 844.5 Hz, the 47th at 1149 Hz, the 48th at 152 Hz. Harmonics that high, folded so, arise when the sine is clipped while still at 48 kHz – in the digital path, before the converter's oversampling filter. AudioLab sends the samples within full scale, clamped to it, so the clipping is in the card; with it the level comes 0.08 dB short.

At −1 dBFS none of it is there: the level rises exactly 1.00 dB from −2 dBFS, THD+N is −80.1 dB (0.0098 %), the largest line besides the harmonics lies at −108 dBc. At −2 dBFS the reading is the sweep's (THD+N −81.0 dB). Near full scale the distortion is H2: −80.8 dBc at −1 dBFS, −82.0 dBc at −2 dBFS, −83.2 dBc at −3 dBFS (THD+N −81.7 dB), against −102 dBc at −20 dBFS (10), where H3 led. The clipping therefore sets in between −1 and 0 dBFS. · Data: [11a-thd-n-output-0dbfs.csv](measured/behringer-ufo202/csv/11a-thd-n-output-0dbfs.csv), [11a-thd-n-output-1dbfs.csv](measured/behringer-ufo202/csv/11a-thd-n-output-1dbfs.csv), [11a-thd-n-output-2dbfs.csv](measured/behringer-ufo202/csv/11a-thd-n-output-2dbfs.csv), [11a-thd-n-output-3dbfs.csv](measured/behringer-ufo202/csv/11a-thd-n-output-3dbfs.csv)

Without the isolator, at 1 kHz, the best THD+N was 8 dB worse (−78.7 dB), the noise 7 dB higher. · Data: [11-thd-vs-level-output.csv](measured/behringer-ufo202/csv/11-thd-vs-level-output.csv), earlier: [11b-thd-vs-level-output-1khz-ground-loop.csv](measured/behringer-ufo202/csv/11b-thd-vs-level-output-1khz-ground-loop.csv)

### 12 – Frequency response, line output

**Why:** how evenly the card plays the band from 20 Hz to 20 kHz.

**Settings:** Stepped sine · 31 frequencies 20 Hz – 20 kHz · −20 dBFS into the card · External DAC, OUTPUT → M4 In 3 / In 4 · Windows volume 100 · 2 averages. The gain is the level at the interface input against the digital level; only its course over frequency counts here.

![Gain of both channels from 20 Hz to 20 kHz, relative: flat within a few hundredths of a dB from 50 Hz to 5 kHz, −0.11 dB at 20 Hz, a rise to +0.19 dB at 12.6 kHz and back to −0.04 dB at 20 kHz](measured/behringer-ufo202/screenshots/12-stepped-response-output.png)

**Result** (relative to 1 kHz):

| | L | R |
| --- | --- | --- |
| 20 Hz | −0.11 dB | −0.09 dB |
| 12.6 kHz | +0.19 dB | +0.19 dB |
| 20 kHz | −0.04 dB | −0.04 dB |
| 20 Hz – 20 kHz within | **0.30 dB** | 0.28 dB |

The output is flatter than the input (03): within 0.3 dB over the whole band, with a gentle rise of 0.2 dB around 12 kHz from the converter's filter. The right channel plays 0.02 dB louder than the left throughout. · Data: [12-stepped-response-output.csv](measured/behringer-ufo202/csv/12-stepped-response-output.csv)

### 13 – Noise and S/N, line output

**Why:** the noise the card plays when it plays nothing, against its full scale.

**Settings:** Noise & S/N · a 1 kHz tone at 0 dBFS first finds the full scale at the output (1.17 V), then the card plays digital silence · External DAC, OUTPUT → M4 In 3 / In 4 · Windows volume 100 · USB isolator · band 20 Hz – 21.6 kHz · unweighted, A-weighted and ITU-R 468.

![Noise density of both channels from 10 Hz to 20 kHz: −150 to −155 dBFS/√Hz above 200 Hz, a little higher towards the bass; a line at 35 Hz in the left channel and a bump near 12 Hz in the right, single small lines at 105 Hz, 1 kHz, 2 kHz and near 20 kHz, no mains hum](measured/behringer-ufo202/screenshots/13-noise-output.png)

**Result:**

| | L | R |
| --- | --- | --- |
| Noise, unweighted | 0.032 mV | 0.027 mV |
| Noise, A-weighted | 0.021 mV | 0.021 mV |
| S/N against full scale (1.17 V) | 91.3 dB, **95.0 dB(A)** | 92.7 dB, 95.1 dB(A) |
| S/N, ITU-R 468 | 85.9 dB | 86.1 dB |

The output is about 4.5 dB quieter than the input (04: 90.5 dB(A)) and comes within 7 dB of what 16 bits allow (98 dB unweighted). The spectrum has no mains hum; what stands above the noise are a line at 35 Hz in the left channel, a bump near 12 Hz in the right and a few small single lines, all below −115 dBFS/√Hz.

Without the isolator the same output read 88.6 dB(A), 6.4 dB worse, with hum at 50 and 100 Hz. · Data: [13-noise-output.csv](measured/behringer-ufo202/csv/13-noise-output.csv), without the isolator: [13b-noise-output-ground-loop.csv](measured/behringer-ufo202/csv/13b-noise-output-ground-loop.csv)

### 14 – Crosstalk, line output

**Why:** how much of one played channel appears in the other.

**Settings:** Crosstalk · one channel played at a time · 31 frequencies 20 Hz – 20 kHz · −20 dBFS into the card · External DAC, OUTPUT → M4 In 3 / In 4 · Windows volume 100 (0 dB) · USB isolator.

![Crosstalk L→R and R→L from 20 Hz to 20 kHz: −37.4 dB at 20 Hz, improving 6 dB per octave to −68.5 dB at 1 kHz, levelling off at −75.7 dB (L→R) and −78.3 dB (R→L) above 5 kHz](measured/behringer-ufo202/screenshots/14-crosstalk-output.png)

**Result:**

| | L → R | R → L |
| --- | --- | --- |
| 20 Hz | −37.4 dB | −37.2 dB |
| 100 Hz | −50.3 dB | −50.1 dB |
| 1 kHz | −68.5 dB | −68.5 dB |
| 10 kHz | −75.6 dB | −77.7 dB |
| 20 kHz | −75.7 dB | −78.3 dB |

The output separates the channels worst in the bass: the leak falls 6 dB per octave, from −37.4 dB at 20 Hz to −68.5 dB at 1 kHz, and levels off in the treble, at −75.7 dB from left to right and −78.3 dB from right to left. Both directions are alike up to 1 kHz. The maker gives −77 dB at 1 kHz without saying whether for the input or the output: the input meets it (05), the output reads 8.5 dB worse. Every leak stands at least 20 dB above the background, so the readings are the card's.

Without the isolator the curve was the same within 1 dB: the output's crosstalk does not come from the ground loop. · Data: [14-crosstalk-output.csv](measured/behringer-ufo202/csv/14-crosstalk-output.csv), without the isolator: [14b-crosstalk-output-ground-loop.csv](measured/behringer-ufo202/csv/14b-crosstalk-output-ground-loop.csv)

### 15 – THD+N at 997 Hz, headphone output

**Why:** the headphone output plays the same converter through its own amplifier and volume control: its level and how clean it is, first without a load.

**Settings:** THD+N meter · External DAC, the headphone jack through a 3.5 mm stereo to two 6.3 mm mono cable to M4 In 3 (tip, L) and In 4 (ring, R), calibrated · the headphone volume at maximum · no load · Windows playback volume 0 dB, the card's input open · USB isolator · 997 Hz at −20 dBFS · FFT 65536, Blackman-Harris, 8 averages.

![THD+N meter, headphone output at maximum volume, 997 Hz at −20 dBFS: the tone at 44 mV in both channels, odd harmonics H3, H5, H7 and H9 between −128 and −137 dBFS, a small line at 50 Hz in the left channel, the noise floor near −155 dBFS per bin](measured/behringer-ufo202/screenshots/15-thd-n-headphone.png)

**Result:**

| | L | R |
| --- | --- | --- |
| Output at −20 dBFS | 44.3 mV | 44.7 mV |
| THD | 0.0087 % (−81.2 dB) | 0.0087 % (−81.2 dB) |
| THD+N | 0.042 % (−67.6 dB) | 0.042 % (−67.6 dB) |
| H2 / H3 / H5 | −98.3 / −83.6 / −88.5 dBc | −97.8 / −83.7 / −88.2 dBc |
| Noise | 0.018 mV | 0.018 mV |

At its maximum the headphone output plays 8.5 dB below the line output (10: 0.118 V), so digital full scale gives about 0.44 V. The harmonics are those of the line output within 0.5 dB – H3, H5, H7, H9 – so they come from the converter, not from the headphone amplifier. The noise is lower than the line output's, 0.018 against 0.033 mV, but so is the level: THD+N is 3 dB worse. Both channels agree within 0.1 dB. · Data: [15-thd-n-headphone.csv](measured/behringer-ufo202/csv/15-thd-n-headphone.csv)

### 16 – THD vs level at 997 Hz, headphone output

**Why:** how far the headphone amplifier goes before it clips, without a load.

**Settings:** THD vs level · 997 Hz · −60 to 0 dBFS in 2 dB steps into the card (output limit raised to 0 dBFS for this sweep) · headphone volume at maximum · no load · otherwise as 15 · 2 averages.

![THD and THD+N of both channels against the output voltage, 0.43 mV to 0.33 V: THD+N falling 2 dB for every 2 dB to −79.5 dB at 0.18 V, THD to −93 dB; at 0.22 V THD rises to −82 dB, at 0.27 V the output clips and the distortion jumps to 3.6 %](measured/behringer-ufo202/screenshots/16-thd-vs-level-headphone.png)

**Result:**

| Digital level | Output L / R | THD L / R | THD+N L / R |
| --- | --- | --- | --- |
| −40 dBFS | 4.4 / 4.5 mV | −60.6 / −60.3 dB | −47.2 / −47.2 dB |
| −20 dBFS | 44.3 / 44.7 mV | −81.4 / −81.5 dB | −67.6 / −67.6 dB |
| −8 dBFS | 0.176 / 0.178 V | **−92.9 / −93.3 dB** | **−79.5 / −79.6 dB** |
| −6 dBFS | 0.222 / 0.224 V | −81.9 / −81.2 dB | −78.8 / −78.4 dB |
| −4 dBFS | 0.273 / 0.275 V | −28.9 / −28.5 dB (clipped) | −28.9 / −28.5 dB |
| 0 dBFS | 0.329 / 0.329 V | −14.7 / −14.2 dB (clipped) | −14.5 / −13.9 dB |

- **At maximum volume the headphone amplifier clips between −6 and −4 dBFS**, at about 0.25 V: the converter's last 5 dB do not get through. Above −8 dBFS the distortion starts to rise; at −4 dBFS it is 3.6 %, at full scale 18 %.
- Below that it is clean: **0.0023 % THD (−93 dB)** and **0.011 % THD+N (−79.5 dB)** at 0.18 V, set by the noise as on the line output.
- Both channels agree within 0.05 dB.

At maximum volume, music that reaches full scale clips in the headphone output; a volume about 6 dB lower keeps the whole digital range clean, without a load. · Data: [16-thd-vs-level-headphone.csv](measured/behringer-ufo202/csv/16-thd-vs-level-headphone.csv)

### 17 – Output impedance, headphone output

**Why:** headphones draw current; how much the output's own impedance lets the level drop under them, and how much it lets the headphones' impedance shape the sound.

**Settings:** Output Z · 997 Hz at −20 dBFS · headphone volume at maximum, untouched between the readings · the level read open (Load: Nothing), then with 32 Ω resistors from L and from R to ground, both channels loaded at once (measured with an ohmmeter: L 31.9 Ω, R 31.4 Ω) · Z = R · (U₀ / U_R − 1).

![The level with the load against without, per channel: −8.245 dB in the left channel, Z 50.5 Ω; −8.256 dB in the right, Z 49.8 Ω](measured/behringer-ufo202/screenshots/17-output-impedance-headphone.png)

**Result:**

| | L | R |
| --- | --- | --- |
| Level open / with 32 Ω | −44.60 / −52.84 dBFS | −44.56 / −52.82 dBFS |
| Drop under the load | 8.24 dB | 8.26 dB |
| **Output impedance** | **50.5 Ω** | **49.8 Ω** |
| Damping factor at 32 Ω | 0.63 | 0.63 |

The headphone output has an output impedance of about **50 Ω**, more than most headphones themselves. With 32 Ω headphones it loses 8.2 dB of level: of the 0.44 V it gives open at full scale, about 0.17 V reaches them. With 300 Ω headphones the loss would be 1.3 dB. Since the headphones and the output divide the voltage between them, headphones whose impedance changes with frequency – most do, at their bass resonance – play with a response shaped by it: the usual guideline asks for an output impedance below an eighth of the headphones'. The open level matches 15 within 0.01 dB, so the volume had not moved. · Data: [17-output-impedance-headphone.csv](measured/behringer-ufo202/csv/17-output-impedance-headphone.csv)

### 18 – THD vs level at 997 Hz, headphone output into 32 Ω

**Why:** how much power the headphone output gives 32 Ω headphones before it clips.

**Settings:** THD vs level · 997 Hz · −60 to 0 dBFS in 2 dB steps into the card (output limit raised to 0 dBFS) · headphone volume at maximum · 32 Ω resistors on both channels (L 31.9 Ω, R 31.4 Ω), Load set to 32 Ω so the X axis is power, V² / R · otherwise as 15 · 2 averages.

![THD and THD+N of both channels against the power into 32 Ω, 1 nW to 0.4 mW: THD+N falling 2 dB for every 2 dB to −73 dB at 146 µW, THD to −90 dB at 92 µW; above 146 µW the output clips and the distortion jumps to 3 %](measured/behringer-ufo202/screenshots/18-thd-vs-level-headphone-32ohm.png)

**Result:**

| Digital level | Voltage L / R | Power L | THD L / R | THD+N L / R |
| --- | --- | --- | --- | --- |
| −20 dBFS | 17.1 / 17.3 mV | 9.2 µW | −80.9 / −80.2 dB | −61.2 / −61.2 dB |
| −10 dBFS | 54.2 / 54.6 mV | 92 µW | **−90.2 / −89.8 dB** | −71.1 / −71.2 dB |
| −8 dBFS | 68.3 / 68.7 mV | **146 µW** | −85.4 / −83.7 dB | **−72.9 / −72.9 dB** |
| −6 dBFS | 84.5 / 84.8 mV | 224 µW | −30.6 / −29.8 dB (clipped) | −30.6 / −29.7 dB |
| 0 dBFS | 112 / 111 mV | 0.39 mW | −12.6 / −12.1 dB (clipped) | −12.5 / −12.0 dB |

- **Into 32 Ω the output gives about 0.15 mW clean** (68 mV, −8 dBFS) and clips by −6 dBFS: 2 dB earlier than open (16). The digital range above −8 dBFS does not reach 32 Ω headphones clean at maximum volume.
- Below that the distortion is as low as open: 0.0031 % THD at 92 µW. THD+N is set by the noise; at 146 µW 0.023 % (−72.9 dB).
- With 32 Ω headphones of 100 dB SPL/mW, 0.15 mW would give about 92 dB SPL of sine; quieter headphones play correspondingly less.
- Both channels agree within 0.1 dB.

Data: [18-thd-vs-level-headphone-32ohm.csv](measured/behringer-ufo202/csv/18-thd-vs-level-headphone-32ohm.csv)

### 19 – Frequency response, headphone output into 32 Ω

**Why:** whether the headphone output plays the whole band evenly into a headphone load.

**Settings:** Stepped sine · 31 frequencies 20 Hz – 20 kHz · −20 dBFS into the card · headphone volume at maximum · 32 Ω resistors on both channels · Compare with phono curve: None · otherwise as 15 · 2 averages. Only the course over frequency counts.

![Gain of both channels into 32 Ω from 20 Hz to 20 kHz: flat within 0.1 dB from 150 Hz to 3 kHz, falling below 100 Hz to 3 dB down at 20 Hz, rising 0.4 dB towards 13–16 kHz; the right channel 0.05 dB below the left](measured/behringer-ufo202/screenshots/19-stepped-response-headphone-32ohm.png)

**Result** (relative to 1 kHz):

| | L | R |
| --- | --- | --- |
| 20 Hz | **−2.99 dB** | −3.06 dB |
| 50 Hz | −0.75 dB | −0.79 dB |
| 100 Hz | −0.22 dB | −0.24 dB |
| 16 kHz | +0.44 dB | +0.39 dB |
| 20 kHz | +0.34 dB | +0.31 dB |

Into 32 Ω the headphone output loses the deep bass: **−3 dB at 20 Hz**, −0.75 dB at 50 Hz, while the line output, unloaded, is within 0.1 dB there (12). A roll-off that appears with the load is what a capacitor in series with the output gives; the card was not opened to check. With headphones of higher impedance it would move lower. In the treble it rises 0.4 dB towards 16 kHz. · Data: [19-stepped-response-headphone-32ohm.csv](measured/behringer-ufo202/csv/19-stepped-response-headphone-32ohm.csv)

### 20 – Noise and S/N, headphone output into 32 Ω

**Why:** the hiss 32 Ω headphones hear from the output when nothing plays, at maximum volume.

**Settings:** Noise & S/N · headphone volume at maximum · 32 Ω resistors on both channels · the card plays digital silence · band 20 Hz – 21.6 kHz · unweighted, A-weighted and ITU-R 468 · otherwise as 15. AudioLab first plays a 1 kHz tone at 0 dBFS for the reference; into 32 Ω it clips (18), so the S/N below is taken against the highest clean level instead, 68.3 mV (146 µW) at −8 dBFS.

![Noise density of both channels into 32 Ω from 10 Hz to 20 kHz: about −145 dBFS/√Hz in the bass falling to −160 dBFS/√Hz above 1 kHz; a single line at 35 Hz in the right channel, no mains hum](measured/behringer-ufo202/screenshots/20-noise-headphone-32ohm.png)

**Result:**

| | L | R |
| --- | --- | --- |
| Noise, unweighted | 14.9 µV | 14.8 µV |
| Noise, A-weighted | 11.2 µV | 11.1 µV |
| S/N against 68.3 mV, the highest clean level | 73.2 dB, **75.7 dB(A)** | 73.3 dB, 75.8 dB(A) |
| S/N, ITU-R 468 | 66.7 dB | 66.8 dB |

The noise itself is low – 11 µV A-weighted, about 4 pW into 32 Ω, some 16 dB SPL in headphones of 100 dB SPL/mW – and without hum; only a line at 35 Hz in the right channel stands out. The S/N is short because the clean level is: into 32 Ω the output stays clean only up to 68 mV, against the line output's 1.17 V. The S/N the export gives, 80 dB(A), is against the clipped 0 dBFS tone (0.111 V) and overstates the clean range by 4 dB. · Data: [20-noise-headphone-32ohm.csv](measured/behringer-ufo202/csv/20-noise-headphone-32ohm.csv)

### 21 – Crosstalk, headphone output into 32 Ω

**Why:** how much of one channel the other headphone hears.

**Settings:** Crosstalk · one channel played at a time · 31 frequencies 20 Hz – 20 kHz · −20 dBFS into the card · headphone volume at maximum · 32 Ω resistors on both channels · otherwise as 15.

![Crosstalk L→R and R→L into 32 Ω from 20 Hz to 20 kHz: flat at −33.5 dB from 80 Hz to 5 kHz, rising to −32.1 dB at 20 Hz and −33.2 dB at 20 kHz; right to left 0.1 dB lower](measured/behringer-ufo202/screenshots/21-crosstalk-headphone-32ohm.png)

**Result:**

| | L → R | R → L |
| --- | --- | --- |
| 20 Hz | −32.1 dB | −32.3 dB |
| 100 Hz | −33.5 dB | −33.6 dB |
| 1 kHz | **−33.5 dB** | **−33.6 dB** |
| 10 kHz | −33.4 dB | −33.5 dB |
| 20 kHz | −33.2 dB | −33.3 dB |

Into 32 Ω the separation is only about **33.5 dB**, and the same at every frequency – 35 dB less than the line output's at 1 kHz (14). The leak does not depend on frequency; the rise at 20 Hz comes from the driven channel, which plays 2.9 dB lower there than at 100 Hz under the load, while the leak stays put. A flat crosstalk of this size is what a ground return of about 0.7 Ω shared by both channels gives against 32 Ω. Whether it lies in the card or in the plug, cable and adapter of this setup was not taken apart; headphones add the ground wire of their own cable to it. · Data: [21-crosstalk-headphone-32ohm.csv](measured/behringer-ufo202/csv/21-crosstalk-headphone-32ohm.csv)

### 22 – IMD SMPTE vs level, line output

**Why:** two tones at once, as in music: how much a strong bass tone modulates a treble one.

**Settings:** IMD vs level · SMPTE, 60 Hz + 7 kHz at 4 : 1 · −40 to −6 dBFS in 2 dB steps, the level being the peak of the two tones together · External DAC, line output → M4 In 3 / In 4 · Windows playback volume 0 dB · USB isolator · 2 averages. (The export names a 32 Ω load: the setting was left from 21, no resistor was on the line output – its level matches 10 to 0.02 dB.)

![IMD and its d2 and d3 parts of both channels against the generator level, −40 to −6 dBFS: IMD falling from 0.18 % to 0.0073 % at −10 dBFS, d3 following it; d2 falling to 0.002 % at −14 dBFS and rising to 0.010 % at −6 dBFS, where it takes the lead](measured/behringer-ufo202/screenshots/22-imd-smpte-output.png)

**Result:**

| Generator | IMD L / R | d2 L / R | d3 L / R |
| --- | --- | --- | --- |
| −40 dBFS | 0.18 / 0.19 % | −67.3 / −67.2 dB | −55.4 / −54.6 dB |
| −20 dBFS | 0.018 / 0.016 % | −87.4 / −89.9 dB | −75.3 / −76.1 dB |
| −14 dBFS | 0.0091 / 0.0092 % | −94.7 / −91.8 dB | −81.2 / −81.3 dB |
| −10 dBFS | **0.0073 / 0.0080 % (−82.8 / −82.0 dB)** | −88.6 / −85.8 dB | −84.7 / −84.8 dB |
| −6 dBFS | 0.011 / 0.012 % | −80.1 / −79.3 dB | −88.0 / −88.1 dB |

- The lowest IMD is **0.0073 % (−82.8 dB) at −10 dBFS**. Below it the reading falls 2 dB for every 2 dB of level: d3 lies at the noise, as THD+N does (11).
- Above −14 dBFS d2 grows, 1 dB per dB, and leads at −6 dBFS – the same even distortion that H2 shows near full scale (11a).
- Both channels agree within 1 dB.

Data: [22-imd-smpte-output.csv](measured/behringer-ufo202/csv/22-imd-smpte-output.csv)

### 23 – IMD CCIF vs level, line output

**Why:** two treble tones close together: what the output makes of them below the band – the difference at 1 kHz and the products beside them – where the converter's filter and the analog stage work hardest.

**Settings:** IMD vs level · CCIF, 19 kHz + 20 kHz at 1 : 1 · −40 to −6 dBFS in 2 dB steps, the peak of the two tones together · otherwise as 22 (the export again names the 32 Ω load left in the setting; no resistor was on).

![IMD and its d2 and d3 parts of both channels against the generator level, −40 to −6 dBFS: IMD falling steadily from 0.089 % to 0.0021 % at −6 dBFS with d3; d2, the 1 kHz difference tone, falling from 0.025 % to 0.0005 %](measured/behringer-ufo202/screenshots/23-imd-ccif-output.png)

**Result:**

| Generator | IMD L / R | d2 (1 kHz) L / R | d3 (18 and 21 kHz) L / R |
| --- | --- | --- | --- |
| −40 dBFS | 0.089 / 0.085 % | −72.1 / −72.0 dB | −61.4 / −61.8 dB |
| −20 dBFS | 0.0089 / 0.0085 % | −91.3 / −90.6 dB | −81.5 / −82.0 dB |
| −10 dBFS | 0.0030 / 0.0029 % | −102.7 / −100.6 dB | −90.8 / −91.2 dB |
| −6 dBFS | **0.0021 / 0.0022 % (−93.5 / −93.1 dB)** | −106.0 / −105.3 dB | −93.7 / −93.3 dB |

The reading falls 2 dB for every 2 dB of level all the way to −6 dBFS: it is the noise beside the products, not the products. The difference tone at 1 kHz lies at −106 dB at the top. The output handles two treble tones at full level cleanly; the even distortion of 22 belongs to the bass and the higher levels. · Data: [23-imd-ccif-output.csv](measured/behringer-ufo202/csv/23-imd-ccif-output.csv)

### 24 – IMD SMPTE vs level, line input

**Why:** a strong bass tone and a weak treble one into the input, as in music: how much the bass modulates the treble.

**Settings:** IMD vs level · SMPTE, 60 Hz + 7 kHz at 4 : 1 · generator −40 to −6 dBFS in 2 dB steps, the level of a sine with the same peak (24.5 mV to 1.23 V) · otherwise as 25.

![IMD and its d2 and d3 parts of both channels against the generator level, −40 to −6 dBFS: IMD falling from 0.045 % to 0.005 % at −18 dBFS; above it the right channel rises to 0.20 % at −12 dBFS, mostly d2, the left one to 0.034 % at −10 dBFS](measured/behringer-ufo202/screenshots/24-imd-smpte-line.png)

**Result:**

| Generator | IMD L / R | d2 L / R | d3 L / R |
| --- | --- | --- | --- |
| −40 dBFS (24.5 mV) | 0.046 / 0.043 % | −74.1 / −71.4 dB | −71.9 / −74.4 dB |
| −18 dBFS (0.31 V) | **0.0053 / 0.0050 % (−85.4 / −86.0 dB)** | −86.9 / −87.9 dB | −95.3 / −92.5 dB |
| −14 dBFS (0.49 V) | 0.011 / **0.11 %** | −80.4 / −62.3 dB | −85.3 / −64.1 dB |
| −12 dBFS (0.62 V) | 0.015 / **0.20 %** | −84.8 / −56.0 dB | −80.4 / −59.2 dB |
| −10 dBFS (0.78 V) | 0.034 / 0.15 % | −72.9 / −57.2 dB | −73.1 / −64.6 dB |
| −6 dBFS (1.23 V) | 0.019 / 0.085 % | −76.9 / −67.8 dB | −79.6 / −67.8 dB |

- The lowest IMD is **0.005 % (−86 dB) at 0.31 V**. Below it the noise sets the reading; the d2 part levels off near −87 dB from about −26 dBFS on, so some even modulation is there at every level.
- Above 0.4 V the **right channel** rises steeply, to **0.20 % (−54 dB)** at 0.62 V, mostly d2; the left one to 0.034 % at 0.78 V. The right channel's weakness of 02 and 25 is plainest here, with the strong bass tone.
- In use: line signals up to about 0.3 V pass cleanly; between 0.4 and 1 V the right channel adds modulation products of up to 0.2 %.

Data: [24-imd-smpte-line.csv](measured/behringer-ufo202/csv/24-imd-smpte-line.csv)

### 25 – IMD CCIF vs level, line input

**Why:** two treble tones close together into the input: the difference at 1 kHz and the products at 18 and 21 kHz, where the input's anti-alias filter and analog stage work hardest.

**Settings:** IMD vs level · CCIF, 19 kHz + 20 kHz at 1 : 1 · generator −40 to −6 dBFS in 2 dB steps, the level of a sine with the same peak (24.5 mV to 1.23 V) · External ADC, MOTU OUT 3 / OUT 4 → the card's INPUT L / R · LINE · Windows input volume 0 dB · USB isolator · 2 averages.

![IMD and its d2 and d3 parts of both channels against the generator level, −40 to −6 dBFS: IMD falling from 0.01 % to 0.0009 % at −18 dBFS; above it the right channel rises to 0.036 % at −12 to −10 dBFS and falls back to 0.023 %, the left one only to 0.006 %](measured/behringer-ufo202/screenshots/25-imd-ccif-line.png)

**Result:**

| Generator | IMD L / R | d2 (1 kHz) L / R | d3 (18 and 21 kHz) L / R |
| --- | --- | --- | --- |
| −40 dBFS (24.5 mV) | 0.010 / 0.012 % | −82.5 / −80.7 dB | −82.9 / −81.6 dB |
| −18 dBFS (0.31 V) | **0.00095 / 0.00093 % (−100.4 / −100.7 dB)** | −102.2 / −105.4 dB | −105.1 / −102.4 dB |
| −14 dBFS (0.49 V) | 0.0021 / **0.017 %** | −96.8 / −80.2 dB | −96.2 / −77.4 dB |
| −10 dBFS (0.78 V) | 0.0057 / **0.036 %** | −94.1 / −70.7 dB | −85.4 / −73.2 dB |
| −6 dBFS (1.23 V) | 0.0058 / 0.023 % | −91.0 / −73.0 dB | −85.8 / −89.8 dB |

- Up to 0.31 V the reading falls with the level, set by the noise: the input mixes two treble tones very cleanly, **0.0009 % (−100 dB)**. That is 1–2 dB above the MOTU loop alone at the same level (0.0008–0.0009 %, its protocol M08): the card may be cleaner still, the measuring path does not show it.
- Above 0.4 V the **right channel** turns worse – 0.036 % (−69 dB) at 0.6–0.8 V, d2 and d3 alike – while the left stays below 0.006 %. It is the same range and the same channel in which THD rose in 02: the right channel's distortion above 0.4 V shows with two tones too.

Data: [25-imd-ccif-line.csv](measured/behringer-ufo202/csv/25-imd-ccif-line.csv)

### 26 – THD+N at 997 Hz, line output, 44.1 kHz

**Why:** music on CD and most files is at 44.1 kHz; whether the output plays it as cleanly as 48 kHz.

**Settings:** as 10, at **44.1 kHz** (CONNECTIONS → Sample rate, for the interface and the card together) · 997.26 Hz at −20 dBFS · band 20 Hz – 19.8 kHz.

![THD+N meter, line output at 44.1 kHz, 997 Hz at −20 dBFS: the tone at 0.115 V, odd harmonics H3 at −114 dBFS, H5, H7, H9 and H11 (at 10.97 kHz, marked as a spur) between −118 and −126 dBFS, the noise floor near −150 dBFS per bin](measured/behringer-ufo202/screenshots/26-thd-n-output-44k1.png)

**Result:**

| | 44.1 kHz L / R | 48 kHz (10) L / R |
| --- | --- | --- |
| Output at −20 dBFS | 0.1152 / 0.1152 V | 0.1181 / 0.1181 V |
| THD | **0.017 % (−75.4 dB)** / 0.017 % | 0.0082 % (−81.7 dB) / 0.0082 % |
| THD+N | 0.034 % (−69.4 dB) / 0.034 % | 0.029 % (−70.7 dB) / 0.029 % |
| H3 | −77.6 / −77.5 dBc | −83.9 / −84.0 dBc |
| Noise | 0.034 / 0.033 mV | 0.033 / 0.033 mV |

At 44.1 kHz the output distorts **twice as much** as at 48 kHz: the same odd row H3, H5, H7, H9, each about 6 dB higher, and H11 at 10.97 kHz (−89 dBc, just above the H10 the meter counts, so marked as a spur). The level is 0.22 dB lower, full scale about 1.15 V; the noise is the same. THD+N, set mostly by the noise, rises by 1.3 dB. · Data: [26-thd-n-output-44k1.csv](measured/behringer-ufo202/csv/26-thd-n-output-44k1.csv)

### 27 – Frequency response, line output, 44.1 kHz

**Why:** at 44.1 kHz the band edge, 20 kHz, lies close to the Nyquist frequency of 22.05 kHz: whether the converter's filter still leaves the top of the band alone.

**Settings:** as 12, at 44.1 kHz · Compare with phono curve: None.

![Gain of both channels from 20 Hz to 20 kHz at 44.1 kHz, relative: flat within a few hundredths of a dB from 40 Hz to 2 kHz, −0.11 dB at 20 Hz, a ripple of up to +0.23 dB between 3 and 16 kHz, back to −0.01 dB at 20 kHz; the right channel 0.04 dB below the left](measured/behringer-ufo202/screenshots/27-stepped-response-output-44k1.png)

**Result** (relative to 1 kHz):

| | L | R |
| --- | --- | --- |
| 20 Hz | −0.11 dB | −0.13 dB |
| 4 kHz | +0.15 dB | +0.15 dB |
| 12.6 kHz | +0.23 dB | +0.23 dB |
| 20 kHz | −0.01 dB | −0.01 dB |
| 20 Hz – 20 kHz within | **0.33 dB** | 0.36 dB |

At 44.1 kHz the output is as flat as at 48 kHz (12: within 0.30 dB): the converter's filter leaves 20 kHz untouched though it lies 2 kHz below Nyquist, and its ripple between 3 and 16 kHz stays below a quarter of a dB. · Data: [27-stepped-response-output-44k1.csv](measured/behringer-ufo202/csv/27-stepped-response-output-44k1.csv)

## Limits

- **One unit.** Another UFO202 may differ, the right channel's distortion above 0.4 V in particular (02).
- **48 kHz, with 44.1 kHz checked on the output** (26, 27). The card's input records at 16 and 48 kHz only – Windows lists no other rate for it, and at 44.1 kHz it refuses to start – so a 44.1 kHz recording through it is resampled by Windows in shared mode; the other measurements and other rates were not repeated.
- **The phono levels** rest on the attenuator's −40 dB, so the millivolts carry its tolerance; the RIAA deviation is relative to 1 kHz and does not depend on it.
- **The phono input with a cartridge** – its input capacitance and how it loads a real cartridge – was not measured: a resistive source stands in for the cartridge.
- **The USB isolator** brought small spurs of its own: near 215–220 Hz and 7 kHz on the line input (01, 04). The treble crosstalk of the line input read up to 6 dB worse with it (05); the reason was not found.
