IR Troubleshooting: Learning Works but Playback Is Unreliable

If learning succeeds but playback works only sometimes, you’re usually close—but missing one of the “boring” IR details: the correct carrier frequency, the right repeat pattern, stable line-of-sight, or a quiet light environment.

1) Carrier frequency assumptions

Many consumer remotes are near 38 kHz, but not all. If your playback uses the wrong carrier, it can “sort of” work at close range and fail at distance.

• Try a couple common carriers: if your tool supports it, test 36/38/40/56 kHz.
• Compare multiple learns: inconsistent timing often signals noise rather than the true carrier pattern.

2) Repeats and “long press” frames

• Short press vs long press: some devices require a repeat frame (or repeated bursts) to register.
• Hold behavior: volume/channel often uses a different pattern than power.
• Too many repeats can be bad: flooding can confuse receivers; tune repeat count to the device behavior.

3) Placement, sensor location, and distance

• Find the receiver window: it may be off-center or behind tinted plastic.
• Start at 10–30 cm and increase distance gradually once reliable.
• Avoid extreme angles: reflections help sometimes, but don’t depend on them while troubleshooting.

4) Ambient light noise and interference

• Sunlight is loud: direct sun can raise the noise floor and cause intermittent misses.
• Some LEDs flicker: test in a darker room if reliability is inconsistent.
• Don’t learn in noise: a “successful” learn can still capture garbage that replays poorly.

5) A repeatable validation workflow

1. Learn the same button 3 times in a dim room, close range, stable aim.
2. Replay each learned sample at 10–30 cm and count success rate over 10 attempts.
3. If intermittent, adjust repeats/hold behavior and retest.
4. Increase distance only after close-range is stable.
5. Document carrier/repeat settings alongside the captured file so you can reproduce later.

Validation criteria

1. You can achieve consistent success at close range in a low-noise light environment.
2. Reliability remains high across multiple learned samples, not only one capture.
3. You can explain failures as carrier mismatch, repeat pattern, or environmental light—not “random.”