Spring charts work backwards from one number, and that number is the weight of the door hanging in the opening on the day somebody measured it. Everything else follows. The wire size, the length, the inside diameter and the number of turns put on the shaft are all derived from that single figure, and if the figure is right the door floats at any height and the opener barely works at all.
What nobody writes on the chart is that the figure has an expiry date. A door is not a sealed object. It absorbs water, it gets painted, it gets panels added to it, it gets a section swapped after somebody reverses into it, and every one of those events moves the weight the spring is fighting. The spring does not get told. It keeps delivering exactly the torque it was wound for, which was correct for a door that no longer exists.
By the time somebody is standing under a broken spring with the door flat on the floor and the opener making an unhappy noise, the panels overhead often weigh noticeably more than the pair above them was ever sized to lift. The failure looks sudden and it reads as bad luck. It usually is not. It is a part that spent several years working above its design load because the thing it was lifting changed and the counterbalance never did.
The counterbalance is matched to a weight, not to a model number
A torsion spring does not know what brand of door is below it. It stores a twist and gives it back. The only question that matters is whether the torque it gives back at any point in the travel equals the weight it has to hold at that point.
Get that right and the door is balanced. It sits where it is left at knee height and at shoulder height, the opener only has to overcome friction and get it moving, and the spring works through its stress range exactly as the fatigue rating assumed.
Get it wrong by even ten or fifteen percent and nothing announces itself. The door still goes up. The opener still closes it. What has changed is that the spring is now being twisted slightly further past its design point on every single cycle, and fatigue life falls away much faster than the weight error suggests. A modest overload does not shave a modest slice off the service life. It takes a serious bite out of it.
That is why a spring on a door that has gained weight breaks early, and why replacing it with the same part breaks early again.
Where the weight actually comes from, by material
This is where the panel material stops being a cosmetic choice and starts being an engineering one.
Wood that has been living in coastal air
Solid and panelled wood doors are the clearest case, because wood changes weight two ways at once.
The first is moisture. Wood is hygroscopic, and a door standing in humid onshore air sits at a higher equilibrium moisture content than the same door would inland. The swing is not trivial across a big slab of timber, and it is highest exactly where the door spends most of its life facing the street with weather on one side and a dry garage on the other. A door that was measured in a factory and shipped dry is heavier once it has settled in. This is routine in the beach cities of Orange County, where garages in Huntington Beach and Newport Beach take salt-laden marine layer most mornings and never quite dry out at the bottom rail.
The second is paint. A wood door gets repainted because that is how a wood door survives, and every coat adds weight. One repaint is barely measurable. Four or five decades of repaints, each one a full primer and topcoat over the previous film, build a layer that is no longer a finish but a laminate. Nobody strips a garage door back to bare timber before repainting it, so the coats accumulate.
Put the two together and the arithmetic goes one direction only. A tired, often repainted wood door in a damp garage weighs more than the spec sheet it was sprung from, and the gap has been widening quietly since the day it was hung.
Steel that has had things added to it
Steel doors are stable in themselves. The weight drift comes from what gets bolted to them.
The common one is insulation. A single skin steel door is light, and an owner tired of a garage that reads like an oven in an Anaheim or Irvine summer buys a retrofit kit: rigid foam panels or batts cut to each section and held in with clips or adhesive. It works, it is a sensible purchase, and it is also pounds per section added to a door whose spring was sized without any of it. Nothing about fitting a kit prompts anyone to rewind the shaft.
The second is appearance. Carriage house overlays, applied trim boards, decorative hardware and faux hinge straps are all fastened to the face of an existing door. Hardware is light. A full overlay package is not, especially a timber or composite one. The door looks like a different door afterwards and it weighs like one too.
A single section replaced in whatever was available
Dented sections get swapped, and the replacement is matched on appearance and fit rather than on mass.
A steel section ordered in a lighter gauge because the original is no longer made, an insulated section fitted into an uninsulated door because that was what the supplier had, a composite bottom section put under steel upper sections to resist the wet corner of the garage: each of those is a reasonable repair and each of them changes the total. The imbalance is also distributed oddly, which is its own problem. Weight concentrated in the bottom section behaves differently through the curve of the track than weight spread evenly, and the door can end up fighting the drums at the one point in travel where it was previously fine.
Glass, which weighs far more than it looks
Swapping solid stiles for a row of window inserts, or putting full-view glazing into the top section, is one of the biggest single weight changes anyone makes to a residential door. Glass and its frames are dense compared with the thin steel or thin wood panel they replace, and insulated double glazed inserts more than double that again.
Owners who do this usually treat it as a trim change. It is not. It is a specification change, and the spring needs recalculating the same afternoon.
Composite faced with real timber
Composite and fibreglass doors are often chosen as the low maintenance answer to the wood door problem, and then faced or clad with genuine timber to get the look back. That is a composite core carrying a timber skin, sitting in the same humidity the plain wood door was struggling with, and the cladding takes on moisture even when the core does not. The result sits somewhere between the two materials, and almost never where the original spring chart put it.
Why the same part number breaks again
All of that converges on one decision made at repair time, and it is the decision that quietly separates a spring job that lasts from one that repeats.
The replacement spring has to be sized to the door as it weighs today. Not to the sticker inside the jamb, not to the old spring's markings, and not to what the manufacturer's data says a door of that size and build should weigh. Those are all records of the original door. On a door that has been repainted, insulated, overlaid, glazed or part replaced, they are records of something else.
A same part number swap on a changed door leaves the pair under sprung from the first turn. The symptoms are familiar to anyone who has lived with one. The door is heavy on the manual release and heaviest near the floor. It drifts down from half height instead of staying put. The opener strains, slows, trips its force setting on cold mornings, and eventually strips the plastic drive gear in the head unit, because an opener is designed to move a balanced door and not to lift an unbalanced one. Cables and drums take more load than they were chosen for as well.
The honest route is measurement rather than inference. The technician unwinds the springs with winding bars, lets the door rest on the floor with no counterbalance on it, and takes the weight directly. That is a two minute step inside a job that is happening anyway, and on a two spring door both springs are replaced as a matched pair regardless, because the second one has run the identical cycle count in the identical air and a new spring beside an aged one splits the load unevenly from the start.
What the homeowner knows that the chart does not
A technician can weigh the door. What a technician cannot see is its history, and the history is what explains the number.
Worth saying out loud before the parts are ordered: how many times the door has been repainted, and roughly over how long. Whether an insulation kit was ever fitted, and whether it is still all there. Whether any decorative overlay or trim was added after purchase. Whether any section has been replaced, which one, and whether it came from the original manufacturer. Whether windows or glass inserts were put in later. Whether the original door was insulated or single skin, if anyone remembers.
If the last spring job lasted noticeably less time than the one before it, say that too. A shortening interval between spring failures on the same door is the clearest single signal that the door got heavier somewhere along the way, and it points at the weight before anyone has touched a tool.
Diagnosing the counterbalance from the panel material out is slower than reading the old spring and ordering its twin. It is also the only version that ends with a door that stays where it is put.