Your sound board with repo is not working because the return/repo signal path is being defeated somewhere between the board's output stage and the amplifier input. The three usual culprits are a bus assignment that sends the channel only to the main L/R instead of the repo bus, a TRS insert jack that bleeds signal to ground when nothing is plugged in, or a ground-loop or impedance mismatch that mutes the return channel. Each of these can make the repo pathway read as dead while the main mix still passes audio normally.
How the Repo Bus Actually Routes Signal
A repo (return, or "remote") bus on a professional sound board is a summing amplifier input, not a pass-through wire. Every channel strip splits its post-fader signal into a mix bus matrix. When you press the repo assign button, a field-effect transistor or voltage-controlled amplifier gate opens and shunts a copy of that channel's signal into the repo summing node. That node feeds a discrete op-amp stage, then the repo master fader, then the balanced output driver.
The failure point most operators miss: the repo assign is often post-fader by default. If the channel fader sits at minus infinity (bottom), the repo bus receives silence even though the assign LED is lit. This is a design choice, not a fault. On most analog consoles, the repo send taps the channel after the fader and after the mute circuit.
The output driver is the second gate. A balanced repo output uses a differential pair, typically a 5532 or similar dual op-amp, driving tip and ring through 47-ohm to 100-ohm buildout resistors. If the tip or ring leg opens, or if the driver's negative rail collapses, the output can float at ground potential. A voltmeter reads signal at the summing node but nothing at the connector. That is why board metering can look healthy while the repo jack is functionally dead.
Signal flow, in plain terms: the channel makes audio, the fader decides how loud, the repo button decides whether that audio joins the return mix, and the output driver pushes the combined mix down the cable. Break any link and the chain stops.
The Numbers Behind "Working" and "Not Working"
Audio levels on a repo bus are measured in dBu, a voltage reference where 0 dBu equals 0.775 volts RMS. A healthy line-level repo output should deliver plus 4 dBu nominal, roughly 1.23 volts RMS, with 20 dB of headroom before clipping at about plus 24 dBu (12.3 volts RMS). If your meter shows the repo master at minus 20 dBu, the connected amplifier likely needs 26 dB or more of gain to reach full output, which often exceeds its input sensitivity.
Impedance matters here. A balanced line output expects a load of 10,000 ohms or higher. Connect it to a 600-ohm input and the driver may current-limit, distort, or sag. Modern gear is bridging, so this is rare, but older installed systems still use 600-ohm matching. A mismatch of this kind reads as a weak or intermittent repo channel, not a hard failure.
Cable capacitance is a quieter killer. A 100-foot unbalanced run with 30 picofarads per foot carries about 3 nanofarads. Against a 10,000-ohm load, the RC rolloff corner sits near 5.3 kHz. The repo signal then sounds dull and low, which operators describe as "not working" when they mean "sounds wrong." Balanced runs tolerate far more length because the differential receiver rejects the common-mode noise that capacitance amplifies.
One threshold matters more than any other: the noise floor. A repo bus with 60 dB of gain and a 1 millivolt hum pickup produces an audible buzz. Ground loops inject 60 Hz (or 50 Hz outside North America) at levels that swamp quiet passages. The board is passing signal, but the hum masks it.
Edge Cases That Break the Return Path
Insert jacks are the most common silent killer on analog boards. A TRS insert on a repo master, when nothing is plugged in, relies on internal normalling contacts to bridge tip and ring. Those contacts oxidize. After 10 to 15 years in a humid room, a normalling contact can develop hundreds of ohms of resistance or open entirely, muting the repo path while the main output stays clean. Plugging a cable into the insert "fixes" it because the plug forces contact.
Digital boards add a different failure class. A repo output assigned to a Dante or AES50 network stream will not pass audio if the sample-rate clock is not locked. A clock mismatch of even one sample per frame produces dropouts, not silence, so operators hear stuttering and assume a bad cable. Check the clock master before replacing hardware.
Phantom power is a documented edge case. Engaging 48-volt phantom on a repo output that is not designed for it can damage the output capacitor or the driver. The repo dies quietly days later. Never route phantom to a line-level return unless the manufacturer rates the output for it.
Another counterintuitive detail: some boards mute the repo bus when the main L/R fader is pulled down, because the repo master is derived from the same post-main node. The repo "works" only when the main is up. This is rare but it exists on budget mixers, and it fools technicians who test the repo bus in isolation.
Red flag: repo LED lights but no output at the connector — suspect the output driver, not the assign. Red flag: repo output drops only in humid weather — suspect oxidized insert normalling contacts. Red flag: repo output buzzes at 60 Hz when a laptop charger connects — suspect ground loop. Red flag: repo stutters on a digital board — check clock lock before swapping cables. Red flag: main meters active, repo meters dead — verify the assign is post-fader and the channel fader is up.
Misconceptions That Send People Down the Wrong Path
The first misconception is that the repo button is the only switch that matters. On most desks, the repo send passes through the mute circuit, the fader, and sometimes a subgroup. A muted channel sends nothing to the repo bus even with the assign engaged. Check the full strip, not just the button.
The second is that a lit signal LED means audio is present. The LED usually taps a point before the fader. A channel can show full green metering and send nothing to the repo bus because the fader is down. Metering location varies by model and is worth confirming in the manual.
The third is that cables fail suddenly. Copper rarely fails without mechanical stress. More often, a cable's shield contact loosens inside the XLR shell, producing an unbalanced signal that hums. The repo bus is fine; the connector is not.
The fourth misconception concerns "repo" itself. In live-sound and broadcast contexts, repo can mean return, remote, or a specific manufacturer's monitor bus label. Confirm which bus your model calls repo before chasing signal through the wrong schematic. A board where repo means "remote control port" will never pass audio through that jack, and no amount of troubleshooting changes that.
Professionals also warn against testing a return path with a single source. A repo bus that passes a line-level sine wave can still fail under a low-impedance dynamic microphone because the source cannot drive the load. Test with the actual source when possible.
Comparison of Return-Path Architectures
| Architecture | Typical signal level | Failure mode | Diagnostic clue | Best-fit use |
|---|---|---|---|---|
| Analog balanced repo output | +4 dBu nominal | Insert normalling contact opens | Buzz or silence only at the master | Installed analog consoles |
| Digital networked repo stream | 0 dBFS digital | Clock unlock dropouts | Stutter, not silence | Dante/AES50 systems |
| Unbalanced repo RCA | -10 dBV | Ground loop hum | 60 Hz buzz with laptop charger | Home and small venues |
| Transformer-isolated repo | +4 dBu | Core saturation at high level | Distortion on loud peaks | Broadcast and mobile rigs |
| Impedance-matched 600-ohm | 0 dBm | Driver current limiting | Weak or clipped output | Legacy installed systems |
| Mic-level repo return | -50 dBu | Insufficient gain staging | Very low level, hiss | Talkback and intercom |
The transformer-isolated row hides a caveat. Isolation solves ground loops but saturates when driven above about 1 volt at low frequencies. A repo channel that sounds fine on speech can distort on kick drum.
Diagnosing the Signal Path, Step by Step
Start at the connector and work backward. Set the source to a known good signal, such as a 1 kHz tone at 0 dBu from a signal generator or an interface. Measure at the repo output jack with a meter or an oscilloscope. If you read the tone, the board is fine and the problem lies downstream in the cable or amplifier.
If the jack is silent, measure at the repo master insert point. Signal there but not at the connector confirms a dead output driver. No signal at the insert confirms the problem is upstream, in the bus assignment or channel strip.
Next, confirm fader and mute states. Pull the channel fader to unity and mute off. Recheck the assign LED. If the bus still reads dead, swap the channel to a different repo input and observe whether the fault follows the channel or the bus.
For digital boards, check clock status first. A "lock" or "sync" indicator should be steady, not blinking. Confirm the repo output is mapped to the correct physical or network channel, because misrouted Dante subscriptions show as silence with no error message.
Here is a practical troubleshooting sequence:
Verify source signal at the connector with a meter; record the reading for comparison. Check the repo assign is post-fader and the channel fader is above minus 20 dB. Exercise each insert jack by plugging and unplugging a cable to wipe normalling contacts. Confirm clock lock and channel mapping on any digital console before touching hardware. Lift the ground on the amplifier side using an isolation transformer to test for loops. Re-terminate suspicious XLR or TRS connectors with fresh solder before replacing gear.
A quiet, useful threshold: if the repo output measures below minus 10 dBu with a plus 4 dBu source, treat the path as degraded and inspect the driver. Above plus 20 dBu, check for clipping on the next device.
Limits of What This Diagnosis Can Prove
Bench testing a repo bus proves the board and the immediate cable. It does not prove the amplifier's input stage is healthy, the room wiring is correct, or the venue's electrical ground is sound. A repo channel can measure perfectly and still hum in the room because the fault exists in the building's ground reference.
Aging electrolytic capacitors in the output stage degrade over 15 to 25 years, raising distortion and lowering headroom gradually. Bench numbers may still pass while real-world performance sags. Only a distortion analyzer or a careful listening test catches this.
No single article can substitute for the service manual of a specific model. Routing, mute logic, and metering points vary by manufacturer, and repo definitions vary by market. Treat the thresholds here as starting calibration points, not specifications for your console.
This article provides general educational information about audio signal-path troubleshooting, not a diagnosis or repair plan for any specific device. It does not promise that any outcome will be achieved. Consult a qualified audio technician or the manufacturer's service documentation before modifying or repairing equipment.