Jacksonville averages roughly 68 thunderstorm days a year. That figure gets used badly — usually attached to a dramatic claim about direct strikes. Direct strikes on a specific house are rare, and a strike severe enough to destroy a refrigerator control board usually destroys several other things at the same time, which makes it easy to identify and easy to insure against.
The failures we actually see are not like that.
What a long convective season does
From June through September, a thunderstorm somewhere on the grid is an almost-daily event. Each one produces switching activity on the distribution system: reclosers operating, brief sags as a fault clears, momentary interruptions short enough that a clock does not reset but long enough that a switching power supply does.
A refrigerator is one of the few appliances in a house that is energised continuously for ten or fifteen years. It takes every one of those events. The damage is cumulative rather than catastrophic — electrolytic capacitors degrade, a solder joint that has been thermally cycled develops a crack, an input stage weakens.
Then one afternoon a perfectly ordinary storm passes and the panel goes dark. The homeowner remembers the storm and blames it. The storm was the last event, not the only one.
Why we do not sell a board on the first look
A dead panel looks like a dead board, and the substitution is easy to sell. It is also frequently wrong, so the supply side gets proven first:
Voltage under load, not at rest. A reading taken with nothing running says almost nothing. We read while the compressor starts. A weak neutral, a shared circuit or a marginal subpanel feeder only shows up under load, and a board browning out on every start presents exactly as an intermittent panel.
The receptacle. Loose back-stab connections heat, oxidise and drop voltage. This is common in kitchens that were remodelled once and never inspected since.
Ground reference. A control reads its sensors against a reference. A poor ground gives readings that wander, and the cabinet then holds temperature perfectly at 2am and badly at 6pm.
The sensors themselves. Every temperature sensor is measurable. A sensor out of range means the board is doing precisely what it was told with bad information, and replacing the board changes nothing at all.
Only when the supply, ground, harness and sensors are proven do we agree that the board is the fault — and then we say so with the measurements attached.
The advice that costs us money
A surge protective device at the service panel costs a fraction of a control board for built-in refrigeration. Point-of-use protection on the refrigerator circuit is the cheaper alternative, provided the device is rated for a motor load; a strip intended for a computer is not the right device on a compressor circuit.
We recommend this on the first visit regardless of whether it earns us anything, because the alternative is the same failure two summers later, and we would rather have a customer who trusts the recommendation than a second board sale.
One more thing to check afterwards
When a board does fail, we look for what took it. A control that failed because a fan motor was drawing hard will fail again if the motor stays in place. Diagnosing only the visible casualty is how one repair becomes two.