A rust call on a driveway gate almost always arrives as a paint question. The gate has orange streaks running from the bottom rail, the black has gone chalky, and the owner wants a price for a cleanup and recoat. That is the wrong question, because paint is the last and least consequential decision in a rust repair. The first is how much steel is left, and on a coastal gate that answer is rarely what the surface suggests.
Rust does not eat a gate evenly, and here it does not work from the outside in. It works from wherever water gets in and cannot get out, and on a welded steel leaf there are four such places. The underside of the bottom rail, where grit and clippings hold a damp poultice against the coating long after rain stops. The punched sockets where pickets pass through that rail, each a small cup with a picket standing in it. The toes of every weld, where spatter and heat scale left the coating thin. And the inside of the hollow tube, which breathes humid salt air at night and keeps what the moisture brought with it.
Because the damage runs inside out, visible bloom is a late symptom rather than the problem itself. That is why metal gate repair crews in mount pleasant open a rust call with a hammer, a scraper and a caliper rather than a wire brush, and why a quote written from a walk past the gate is a guess. A leaf that looks alarming is often sound under a failed topcoat, and a leaf that looks merely tired can have a bottom rail with no measurable wall left in it.
Surface rust and section loss are two different jobs
The distinction that governs everything downstream is whether the steel has lost thickness or only lost its finish.
Surface rust is a film. It comes off with a wire wheel and leaves sound, faintly pitted grey metal behind, and the wall still measures what the fabricator bought. It is a finishing problem.
Section loss is not. Steel converting to iron oxide swells to several times its original volume, so the first sign is not a hole but a bulge: a rail that has grown, a seam opening along a weld, paint lifting in a blister that feels spongy. Scrape it and the flakes keep coming, and every layer is wall thickness already gone.
Two field checks separate them. Tap along the member with a hammer handle, because sound tube rings and scaled tube answers dull and flat. Then probe that band hard with an awl, since sound steel stops the tip while compromised steel dents or lets it through, and that is the real boundary of the repair whatever the paint above it looks like.
After that, measure, because a repair decision needs a number. Grind a patch to bright metal and put a caliper on a cut edge, or use an ultrasonic thickness gauge. Ornamental gate frames here generally run between a sixteenth and an eighth of an inch in the wall, and whether a tenth of that is gone or half of it decides everything. Run a magnet along the rail too, since filler over old rust is common and sands beautifully.
The loss matters differently depending on the member
A gate leaf is not a panel. It is a frame carrying its own weight out from two hinges, holding its shape because the members and the welds between them resist being pushed into a parallelogram.
That makes the perimeter frame, particularly the bottom rail and the hinge stile, structural. Loss there shows up as behaviour rather than appearance: the leaf racks and the latch stile drops relative to the strike. Measure corner to corner both ways, since on a square leaf those numbers agree, and half an inch of difference across a twelve foot leaf is a frame that has already moved.
Infill is judged more leniently. A picket carries almost nothing, so one that has lost a third of its wall is cosmetically poor and structurally irrelevant, until it works loose and starts enlarging its socket on every swing. The sockets are where the two questions meet: six rotted ones in a bottom rail is not six small repairs, it is a rail reduced to a strip of steel with a row of enlarged holes through it.
Cutting out and welding in new steel
Section loss gets removed and replaced. No coating, converter or encapsulating product restores thickness, and anything applied over active scale is a lid on a pot that keeps cooking.
Brace before cutting. The frame is what holds the leaf square, so the moment a length of bottom rail comes out the geometry is held by nothing. Take the diagonals first so there is a target to return to, then tack a temporary diagonal across the leaf. Skipping it is how a sound repair ends with a gate that no longer meets its partner.
Cut back past the damage rather than to it. The visible edge of the rot is not the edge of the thin steel, so the cut goes an inch or more into metal that rings and measures. A bead run against a rust thinned edge looks fine and has almost nothing to fuse into.
Match the replacement. Older gates do not always agree with current stock, and being an eighth out shows as a step along the rail the first time low sun hits it. Wall thickness should equal or exceed the original, because lighter puts the next failure exactly where the repair is.
How the joint is made depends on access. A splice mid run is usually built over an internal sleeve, a short length of smaller tube slid inside both ends so the weld has backing and the pieces cannot wander out of line. Where the member runs corner to corner, replacing it whole puts the new welds at the original joints instead of adding two seams mid span.
Thin, pitted tube welds badly at ordinary settings, since heat runs away into the pitting and blows through. Lower current, stitches rather than a continuous bead, and working around the joint keeps burn through and distortion in hand. Old coating and galvanising burn off in the arc, so grind to bright metal first. Then dress the welds flush and knock the spatter down, which on a gate is not cosmetic: weld toes and spatter peaks are where the next coating sits thinnest.
When the honest answer is a new leaf
Repair stops making sense at a point that is easy to recognise once someone names it.
When the rot is not local. A bottom rail is a repair. A bottom rail plus both stiles plus a dozen sockets is a new leaf with the old one paying for the labour to dismantle it. When the frame has racked and stayed racked, because pulling it square after the steel has yielded rarely holds. And when the pitting is general, since a surface with unreliable remaining thickness only gets thinner when it is blasted back to something coatable.
A new leaf is not a new installation, which is the part worth hearing. Posts, footings, hinges, the operator and usually the pattern carry over, and a fabricator can match the surviving leaf closely enough that the pair reads as original. That flips only where the design uses cast collars or scrolls nobody makes now.
Drain holes are part of the repair, not an upgrade
Any member being opened is a chance to fix the reason it failed.
A hollow section in this climate needs to drain, which means a hole at the genuine low point of each closed member, big enough that it cannot bridge over with paint or pack solid with dirt. Plenty of driveway gates are raked to follow a slope, so that low point is a corner rather than the middle of a rail.
The sockets need a deliberate decision. Either they are seal welded all round so nothing enters the rail past the picket, or they are left with a clear path through to a drain below. What fails is the common middle state, half closed by weld and half by paint, holding water against both parts. Open tube ends want caps that shed water rather than plugs that trap it.
Refinishing on site, or taking the leaf to a shop
Once the steel work is done there are two genuinely different paths, and the trade matters.
On site keeps the gate hung and the driveway usable, which decides it for many properties. The work is mechanical removal with wire wheels, flap discs and needle scalers, or a small vacuum shrouded spot blaster, then a brushed or sprayed system. Site preparation realistically reaches a clean, adherent surface rather than a blasted one, so the coating has to tolerate that: zinc rich primer on bare steel, cold galvanising compound over the repair welds, then an epoxy and urethane build. Brushing beats spraying around ornament and into pitting because it works coating in instead of bridging it. The frame interior cannot be treated, older coatings may carry lead so debris has to be contained, and the schedule is hostage to humidity and dew point.
A shop buys preparation site work cannot reach. The leaf comes off its hinges, so the operator is disconnected and hardware freed or cut, it travels supported flat so it cannot flex, and every reachable surface is blasted to bright metal before recoating. What that costs is days with an open driveway and whatever temporary security it needs, plus a real chance of extra work, because blasting finds every thin spot the hammer missed. Re-galvanising an old welded gate carries its own condition: moisture trapped in a closed section turns violent in a zinc bath, so vent holes are mandatory and some galvanisers decline used weldments outright.
Telling a year later whether the repair held
The useful check is not how the gate looks, since fresh coating looks good regardless of what is under it.
Look into a drain hole and see whether water runs out and whether it brings rust with it. Re-take the two diagonals and compare them with the numbers from the day of the repair. Tap the repaired rail for the ring, then tap the sections nobody touched, because those are now the oldest steel on the gate. And look hard at the toes of the new welds, because a repair that will fail again usually says so there first.
