A Combiner Box Is a Density Problem

Choosing a cable gland for a combiner box is not difficult in itself. Choosing ten of them, in a row, on the bottom face of a moulded enclosure that is 300 mm wide, is a different task — and it is the one that actually decides whether the box is buildable, serviceable and still sealed in year five.

The failure pattern on combiner boxes is consistent. The glands are ordered last, after the fuses, the isolator and the surge protection have taken the space inside. They arrive in one size because that is the size the string cable uses. They are fitted at whatever pitch the holes allow, and the last two are tightened by feel because a spanner will not fit between them. Six months later the box has water in it and nobody can say which entry let it in.

Everything below is about avoiding that sequence, starting before any part number is chosen.

Count the Entries Before Choosing Any Gland

The count is always higher than the number of strings. A combiner box has string inputs, a DC output, and then the entries nobody schedules: a monitoring or communication cable, a surge protection earth conductor, a heater or thermostat supply where the site is cold, and a spare position for the string that gets added in phase two.

Each of those has a different cable diameter and therefore a different gland size. A box with twelve strings and a single size of gland will usually have at least two sizes fitted by the end of commissioning, which is fine — provided the hole schedule was planned for it. It is not fine when the holes were drilled first and the extra cable had to be squeezed through a string gland with the insert removed.

The Entry Schedule

An entry schedule is one page and it prevents most of the rework. It is drawn as the bottom face of the box, to scale, with every hole positioned, sized and labelled. If the schedule cannot be drawn without glands touching, the box is too small for the string count and no gland selection will fix it.

Spacing: the Dimension Nobody Draws

A gland needs room to be tightened, and that room is set by the spanner across the body and the dome nut, not by the thread diameter. Two M25 glands whose centres are one thread apart will both fit the holes and neither will be tightenable — the second one will be tightened with pliers or left hand-tight, and hand-tight is where the leak starts.

The working rule is to leave clear space on each side of every gland equal to its spanner size, measured at the widest point of the body. On a crowded box that usually means two rows, offset from each other, rather than one continuous line. An offset pattern also improves access for a socket and gives the cables a natural route into the box without crossing each other.

Two further constraints are easy to forget. The first is the enclosure wall: a hole placed too close to a moulded corner or a rib cannot be sealed because the face is no longer flat under the washer. The second is the door or lid: a gland that protrudes into the swing of a hinged cover will foul it, and the usual field remedy — shortening the gland by leaving the dome nut loose — is not a remedy.

Knockouts vs Drilled Holes

Moulded combiner boxes often ship with knockouts or pre-scored positions on the bottom face. They are convenient and they are also a common source of leaks, because a knockout removes material in a step and leaves a rim that is not always flat.

Whichever method is used, the sealing face is what matters: flat, burr-free, and large enough for the washer to sit wholly on it. A deburring tool and thirty seconds per hole is the cheapest reliability measure on the whole box. Where a knockout leaves a raised rim, a sealing washer with a larger outside diameter than the rim is the correct answer rather than extra torque.

Thin Walls and Torque

A moulded box wall is thin, and a gland tightened into it pulls the wall into a cone. The washer follows the cone, the contact area reduces to a ring, and the joint either leaks immediately or leaks after the first thermal cycle. On a metal box the same torque is absorbed by a stiff wall; on plastic it is not.

Three measures work. Use a body with a large washer face so the load is spread over more material. Support the inside of the wall with a reinforcing plate or a gland plate where several entries are close together. And specify a torque that the wall can take rather than a torque the gland can take — the gland’s figure is a maximum, not a target. Where the wall cannot take any meaningful torque, the sealing element has to do the work through compliance rather than through pressure.

Sealing Many Cables at Once

Density creates a second-order problem: several cables in the same box heat each other, and the internal temperature of a combiner box in sun is well above ambient. Elastomer seals in that environment age faster, and the box’s own internal temperature becomes a selection input rather than a detail.

For the entries themselves, four practices make a difference. Size each gland to its own cable rather than to the largest cable in the box. Use blind plugs on spare positions, rated for the same IP as the glands, rather than leaving open holes. Where a single entry must carry two cables, use a purpose-made multi-entry or split gland instead of removing the insert from a standard one. And keep the insert in its gland: an insert that has been removed to pass a connector is a sealing element that will not go back correctly.

Isolation and Fusing: What the Gland Does Not Do

A combiner box contains fuses, an isolator and surge protection, and none of those functions is affected by the gland. It is worth stating because gland selection is sometimes used to solve an electrical problem it cannot solve: a metal gland does not provide DC isolation, and a plastic gland does not improve it.

Where the box is non-metallic and the requirement is to avoid a conductive path through the enclosure, a polyamide gland is the consistent choice, because it introduces no galvanic couple and no metal-to-metal contact with the wall. Where the box is metal and earthed, a metal gland with a proper earth path is consistent — and the earthing question is the one covered in the earthing and bonding guide, not something a plastic gland can substitute for.

Commissioning and Serviceability

A combiner box is re-entered more often than most enclosures: strings are added, fuses are replaced, monitoring is retrofitted. Every re-entry is a chance to damage a seal, so the design should assume the box will be opened.

The entries upstream of the box follow a different logic — UV, jacket hardness and thermal cycling at the module end — which is set out in the solar cable gland selection guide.

Three things make that cheap. Label both the entry and the cable so the right gland can be identified without tracing. Leave at least one plugged spare position per row, so a new string does not require drilling a hole in a live box. And record the gland sizes on the entry schedule inside the lid, so the next technician fits the same part rather than the nearest one in the van.

When a box does leak, the diagnosis is a separate exercise: reading the water to find which entry failed is faster than replacing glands at random, and on a box with twelve entries that difference is most of the labour.

Questions From Combiner Box Builders

Can I use one gland size for every entry?

Only if every cable is within that gland’s range, which is rare once monitoring and earth conductors are counted. Two or three sizes is normal and simplifies spares.

Do I need IP68 glands on an outdoor combiner box?

Usually not. IP65 or IP66 covers rain and hose spray, which is the real exposure. IP68 is relevant where the box can actually be submerged or jet-washed at close range.

Can I drill extra holes after commissioning?

Yes, if the face is flat where you drill, the hole is the correct size, and the inside of the box is protected from swarf. Drilling a live DC box is a separate safety question and usually means isolating the string first.

Why do the glands on the bottom row leak first?

Because the bottom face collects water, and because those entries are the hardest to reach with a spanner. Access difficulty and leak rate correlate strongly on dense boxes.

Should spare entries be plugged or left with a gland fitted?

Plug them. A blind plug is cheaper, sealed to the same rating, and leaves the position clean for the gland that will eventually be needed.

Send the string count, the cable list and the box drawing, and we will produce an entry schedule with gland sizes, hole dimensions and a spacing check before anything is drilled. The polyamide range covers the black UV-stabilised sizes used on moulded boxes, and the hole size chart gives the drilling dimension for each thread.