Passive Acoustic Monitoring (PAM)¶

This page collects field-tested practices for flying a Slocum quietly and for the mission design PAM work needs. It is platform-focused on Slocum; some underlying principles (reduce pump activity, silence the altimeter) apply to any glider.

Source

Paraphrased from the UG2 community Slack (which has a dedicated PAM channel) and the Teledyne Webb Research user forum. Reflects field experience from NOAA Fisheries, UVI, AMLR, Rutgers, WHOI, and others. This is a condensed community reference — it is not a manufacturer document. For sensor integration, work directly with Teledyne and your instrument vendor.

Flying quietly: the self-noise budget¶

The biggest wins come from reducing the glider's own noise. In order of impact:

1. Buoyancy pump — the loudest thing on the glider¶

Every inflection runs the pump, which dominates the spectrogram (the piston shows as bright broadband bars). To reduce it:

Minimize pump activity: fewer, deeper, slower inflections; use the minimum buoyancy drive that still flies. The same levers as the Power Saving page help here — quieter usually means lower power too.
Drift segments: where the science allows, drift-at-depth (wide deadband) keeps the pump off for long quiet stretches.

2. Pitch control — prefer "set once" over constant servo¶

Constantly servoing the pitch battery makes noise. A common PAM approach:

First let the glider find trim with autoballast on servo (optimal pitch and oil/ballast settings).
Then switch to battpos / "set once" so the battery moves once per half-yo to reach the target and then holds — much quieter than continuous adjustment.

Set-once pitch can miss the target angle

Operators report set-once dives/climbs coming in shallower than commanded (e.g. ~18° when 26° was requested) while still reporting "autoballast achieved." Tune on a real dive: have servo+autoballast find the working pitch/oil values first, then commit them to a fixed battpos. A wider pitch deadband is the alternative if you stay on servo.

3. Altimeter — turn it off when you can¶

The Slocum altimeter pings at ~170 kHz, but because the pulse is so sharp it smears across the whole spectrogram; in person it sounds like continuous quiet ticking. It is worst in shallow water near the bottom.

Turn the altimeter off when you don't need it for flying (deep water, known bathymetry), or restrict it to the dive only (not the climb), and push u_alt_min_depth deep. See the altimeter one-pager.
Keeping the altimeter off also helps the glider stay in low power mode (it's one of the things that blocks it).

4. Fin / rudder — reduce hunting (and roll)¶

Excess rudder movement and roll add noise. Reduce fin hunting with u_hd_fin_ap_deadband_reset 1 and wider heading deadbands (see the Fin / Digifin page). Excess roll also makes noise and muddies the data — trim it out where you can.

5. Pinger / transponder — weigh noise vs. recovery risk¶

A pinger or transponder adds periodic pings to your recordings. Disabling it cuts noise but removes an acoustic means of relocating a lost glider (and may violate insurance requirements). Decide deliberately.

Know your own acoustic signatures

For analysis, expect to see the piston (pump), altimeter pings, and any pinger in your spectrograms. Documenting these self-noise signatures up front makes the biological signal much easier to pick out later.

Mission design for PAM¶

Station-keeping near a mooring or slow feature: running yo's looping around a single waypoint with a tight waypoint distance works well for holding near a PAM mooring. To follow a drifting feature (e.g. a float), run an SFMC script that issues new commanded waypoints based on expected drift speed and direction.
Sampling strategy: consider downcast-only sampling and turning off instruments whose noise competes with PAM (notably a pumped CTD). Fewer active instruments also reduces the abort risk below.

The \"timed out waiting for science to stop logging\" abort

Complex PAM payloads push gliders toward the ~5-instrument threshold where a long-standing bug surfaces: at the surface the glider can hang stopping/starting science logging and aborts with MS_ABORT_BEH_ERROR (surface_x behavior → B_ERROR). It dates to 7.x, was mostly fixed by 8.0, and reappeared on G3S due to different CPU timing; newer 11.x releases claim a fix. Field mitigations: - Sequence the same mission 5–6 times (and set waypoints to continue past the last achieved one) so an overnight abort just resumes the mission. - Add a surface script that puts the glider back on mission on this abort. - Stagger sampling timing/states for the most complex instrument first (the DMON or a DVL is often the culprit) — e.g. sample on the dive but stop a few metres below the surface on the climb. - Review low power mode interactions, and report it to TWR (more reports raise its priority).