Batteries¶

Seaglider's main battery pack is not just an energy store — it is the trim mechanism. The large main battery, with a brass weight bolted to its underside, is the mass that the mass shifter slides fore/aft for pitch and rotates for roll. That double duty means every battery change touches almost everything else: vehicle mass, ballast, pitch/roll trim, and even the compass calibration.

Pack architectures by generation¶

On all generations the main pack carries an 1100 g brass weight on its bottom face — the axial asymmetry that makes roll control work when the pack is rotated (±80° on SGX, ±40° on SG).

A battery change is a trim change¶

Plan for the knock-on effects before swapping packs:

• Mass and ballast — new packs change vehicle mass and centre of gravity. Re-weigh the vehicle, redo the tank ballast / volmax estimate, and enter the new mass in sg_calib_constants.m before dive 1 — the Flight Model System bakes it in at the start of the mission. Weigh every pack and write the weight, serial number and installation date on it. Packs of the same type differ: in one change the new main pack was 158 g heavier than the old one (10.28 vs 10.12 kg) and the new secondary pack 16 g lighter.
• Pitch trim — $C_PITCH established on the old packs is unlikely to survive a battery change; expect to re-trim, and consider starting the mission with the Pitch Adjuster enabled (see Trim & Flight Model).
• Compass — the steel in battery cells carries its own local magnetic field, so swapping or even re-seating packs can shift the compass hard-iron calibration. Check the compass after battery work — see Compass Calibration. Steel parts can also pick up magnetisation from handling. Some teams have looked at checking packs with a gaussmeter, and degaussing them if needed, before they go in.
• Gauges and capacity — reset the fuel gauges and set the new capacity (see below).

A self-test right after a battery change may warn about roll current ("motor ran … measured … mA – problem with motor or ammeter?"). APL-UW describes this as an out-of-date check based on an arbitrary current threshold, and it can be ignored if the roll moves themselves are normal.

Capacity monitoring¶

Declared capacity and cutoffs live in glider parameters; consumption is tracked by on-board fuel gauges and, mission-long, in the vis energy and endurance plots (projected mission duration from the observed per-dive draw):

The quoted effective capacities (575 Ah SGX / 350 Ah SG) are already practical numbers, not nameplate cell capacity — endurance planning should still leave reserve for recovery operations at the end of a mission (a glider in recovery keeps calling on $T_RSLEEP until someone picks it up).

The fuel gauge is an estimate, not a meter¶

There is no coulomb counter on the battery. The glider keeps a running total for each subsystem and sensor: the time it was powered, multiplied by a preset current from the CURRENTS file (values measured in the lab). These totals are summed per battery and compared with $AH0_24V and $AH0_10V. The Batteries and fuel gauges menu shows the breakdown, device by device (pitch, roll, VBD at apogee and at the surface, Iridium, GPS, processor, each sensor), with a total for each bus and the percentage of capacity used.

That has two consequences:

• Anything that draws more than its preset current is invisible. A failing VBD pump that takes much longer than normal at apogee, or oil thickened by cold water, uses more energy than the gauge counts. So can a faulty component. The gauge then shows plenty left while the pack is running out.
• Pack age isn't counted. Lithium primary cells lose capacity in storage. The manufacturer quoted about 3% per year, and operators have assumed up to 5%. A pack built from old stock, or moved from another glider with an unknown history, can have much less than its label says.

In one case a glider hit its 24 V cutoff with the gauge showing more than half the pack left. Its 10 V pack was also falling quickly with a third supposedly remaining, and energy use per dive looked normal. The most likely explanation was an over-optimistic capacity: the 10 V pack had come from another glider and its age was unknown. The follow-up was to measure real current draw on the bench during simulated dives, compare it with CURRENTS, and start recording every pack's manufacture date.

For endurance estimates, use energy-per-dive figures from dives when everything was known to be healthy. For example, one team used only the dives before a pump anomaly, not the whole mission.

After a battery change¶

• Reset the fuel gauges (Batteries and fuel gauges → Reset). If you forget, the new pack inherits the old one's consumption. You can't recover the amp-hours used on the new pack before the reset, so estimate them from the bench and test time.
• Set the declared capacity. Set $AH0_24V (and $AH0_10V on 10/24 V gliders) to the new pack's capacity, less any margin you want to hold back. For example, one team set 310 Ah on a univolt glider rather than the quoted ~350 Ah. A value left over from the last mission may carry a different margin.
• On univolt gliders the packs are paralleled onto one bus. In one example the log reported all consumption on the 24 V line and 0 Ah on the 10 V line. The aggregate percentage is the number to watch.
• A self-test warning "no BATTERY file" has been seen once, as a transient, just before a GPS fix. The file was present afterwards, and the gauges logged consumption normally after the pump was run. Check the gauge output after the self-test rather than assuming it is working.

Pack voltage: what's normal¶

Lithium primary packs hold a nearly flat voltage for most of their life and then drop quickly near the end. A slow, steady decline is therefore not the normal pattern, and a sudden fall is the warning sign.

Where the voltage was measured matters. Readings dip during motor moves, especially VBD pumps. Dips under load matter less than a steady or sharp fall in the unloaded value. Each dive's log records the lowest voltage seen and the amp-hours used, in the $24V_AH and $10V_AH lines. Plot them over the mission rather than reading single values. Cold water also lowers the current a pack can deliver.

Voltage cutoffs and launch refusals¶

$MINV_24V and $MINV_10V are the lowest acceptable voltages. If a pack goes below its value during a dive, the glider goes into recovery with a code such as VOLTAGE_CUTOFF_24V.

The glider also refuses to launch if the minimum voltage seen since boot is below the cutoff ("minimum 24V voltage … is below cutoff value…", "Battery pack voltage too low; unable to launch!"). One short, meaningless dip during bench testing is enough. In one case a very short pump at 16.4 V, probably a relay or FET operating, was recorded as the minimum. Rebooting the glider clears the recorded minimum. Alternatively, lower $MINV_24V for the launch and raise it again straight after.

Spurious low readings¶

Several gliders have reported impossible values, such as a 24 V pack at 0.5–2 V, at random: during a motor move, or during a sampling period with no motor moving at all. The next reading was normal again. The pattern that identifies a measurement glitch rather than a failing pack:

• the values cluster at the normal voltage and around 1 V, with nothing in between;
• comm.log voltages and the other readings in the same dive look normal;
• there is no link to depth, motor current or pump rate.

The 10 V and 24 V measurements share a battery-monitoring relay, a buffer amplifier and the ADC. An intermittent contact or glitch in any of these can cause it. A fault on the high-voltage side of the main relay would be more serious, so check the dive data for any link to pump rate or current.

The only workaround at sea is to disable the check ($MINV_24V,0 is allowed) and watch the voltages yourself in every dive's log and in occasional capture files. Operators have kept gliders flying this way when help was far away. When a glider is sitting at the surface in recovery because of a spurious cutoff, $RESUME alone won't get it diving again. Disable the check for that dive, then turn it back on in the next cmdfile. Fix it properly at the next service.

Fuses and paralleled packs¶

Each pack is fused. Once packs are paralleled on the main board, there is no way to tell from the glider's readings that one of them has a blown fuse. The bus voltage looks fine, but capacity is quietly missing. The only check is to measure each pack on its own, at its connector, before connecting them. Do this after any electrical fault that could have overloaded the packs, such as a shorted main board. Never replace a pack fuse yourself (see safety).

Stretching a low battery¶

When the numbers say the pack is running out before the mission is over, the aim is to cut pumping first, since the VBD is where most energy goes:

• Fly slower. Lower $MAX_BUOY and lengthen $T_DIVE. The saving is larger than the loss of speed.
• Accept slower climbs without extra pumps. Widen $W_ADJ_DBAND and pump less when correcting ($DBDW).
• Fix a heavy trim. If the glider pumps on every climb, shifting $C_VBD a little toward buoyant saves those pumps.
• Use the boost pump only near the surface. Raise $D_BOOST, dive shallower (e.g. 100 m), and consider boost-only dives. See Energy on the VBD page.
• Turn off what isn't essential. Switch off PAM and optical sensors, or sample every other profile. Keep the CTD.

If the glider is to wait at the surface for a pickup, send fewer dive files (UPLOAD_DIVES_MAX,0) and call less often. Lengthen $T_RSLEEP, but check its units in your parameter reference first: operators have got this wrong.

When the glider stops because the gauge says the pack is exhausted, the cause is the declared capacity, not the voltage. One team raised the threshold to leave a 5% margin, flew energy-saving shallow dives toward the pickup point, and let the glider drift at the surface when the current was going the right way. They booked the recovery boat early.

Conversions and payload voltage¶

• 10 V to 24 V. Nothing stops a 10 V pack being rebuilt as 24 V to give more high-voltage capacity, but the pack loses amp-hours. One estimate for an aft pack was about 110 Ah down to 81 Ah. Such a pack is custom-built, not a stock part.
• Payload supply. Spare endcap ports can be powered from either the 10 V or the 24 V bus. This is chosen by a jumper on the main board, not by the .cnf file. The voltage in the .cnf only tells the fuel gauge which capacity ($AH0_10V or $AH0_24V) to charge the sensor's consumption to. Ask APL-UW or your service provider for the jumper positions.
• Minimum voltage for a payload. Check what the payload needs against the pack's end-of-life voltage, not its fresh one. For example, a pCO₂/CH₄ sensor that stops working below 11–12 V can't run from a 10 V pack at all.

Lithium-primary safety¶

Seaglider packs are high-energy-density lithium primary (non-rechargeable) cells. The label warnings are the whole story in miniature — never:

• short-circuit,
• charge,
• force over-discharge,
• overheat or incinerate (each cell is marked with a maximum temperature),
• crush, puncture, or disassemble (never open a pack or replace a blown fuse).

Handling practices (the largest single cause of field failures is accidental short circuit during handling):

• Insulate conductive work surfaces; keep sharp objects off them.
• No rings, watches, or other conductive jewelry while handling packs.
• Non-conductive (or covered) tools only; trim one lead or tab at a time.
• Move cells in trays on carts rather than carrying them by hand.
• Check open-circuit voltage against the label at incoming inspection.
• Never force a pack into or out of its housing.

Storage: original packaging, dry and ventilated, ideally ≤23 °C, segregated from flammables, fresh cells separated from depleted ones, sprinklers and suitable extinguishing means available.

Shipping¶

Glider lithium-metal packs ship as dangerous goods: UN3090 (cells/batteries packed alone) or UN3091 (contained in or packed with equipment — i.e. in the glider), under ICAO/IATA air and IMDG sea rules. The regulatory details, documentation, and vessel-carriage considerations are the same as for Slocum lithium packs — see the shared write-up on the Slocum primary batteries page rather than duplicating it here.

See also¶

• Trim & Flight Model — re-ballasting and re-trimming after battery work.
• VBD — where most of those amp-hours actually go.