Skip to content
System / Voidz

Board repair · fault patterns

Board Works Cold and Fails Once It Is Hot

A board that works first thing and quits by the afternoon is one of the few patterns that is almost always the board. Heat does two things to a circuit that nothing else does: it raises the internal resistance of ageing capacitors until the supply rail sags under load, and it expands and contracts solder joints on heavy components until one of them cracks. Both faults disappear the moment the cabinet cools, which is exactly why they are maddening in the field and straightforward on the bench.

Book a board inRates and terms

$85

Bench diagnosis

Credited when the repair is approved

$1,200–8,000

Typical replacement

What the OEM board costs new

10–25%

Repair cost

Of the replacement price

1–3 business days

Turnaround

Bench time, plus transit. 90-day warranty.

Before you pull the board

Cheaper things that produce the same symptom

Work through these first. Each one is faster and cheaper to check than a board is to ship, and each one is a real reason a working board arrives on the bench.

  1. Airflow, and whatever is making the cabinet hot in the first place

    A blocked condenser, a collapsed filter, or a failed cabinet fan raises the temperature around the board well past what it was designed for. Fix that and the board may have years left — and if it does not, at least the replacement is not going into an oven.

  2. High-limit and thermal cutouts

    A unit shutting down when it gets hot may be a safety doing precisely its job. That is not a board fault, and a new board will shut down the same way. Confirm the limit is not tripping before you look further.

  3. Run and start capacitors on the motors

    Motor capacitors degrade with heat and age, and a weak one produces a unit that starts cold and refuses hot. It presents almost identically to a board fault from the outside and it is a fraction of the cost to test.

  4. Connectors, by reseating them

    Thermal cycling backs connectors out and oxidises terminals. Unplug and reseat every connector on the board cold, then run it. If the fault moves or clears, the board was never the problem.

Then it is the board

What actually implicates it

With the list above cleared, these are the signs that point at the board itself. None of them is proof — that is what the diagnosis is for — but together they are enough to make shipping it worthwhile.

  • It fails after a fairly predictable run time and works again once it has cooled
  • The fault moves or clears when the board is flexed, or when a connector is pressed
  • Cooling one area with freeze spray restores normal operation immediately
  • The low-voltage rail measures correctly cold and sags under load once hot
  • It behaves erratically rather than stopping — random resets, outputs chattering, a display that garbles

Failure modes

What is physically wrong, and why

The cause matters as much as the fault — a board replaced into the condition that killed it fails again.

Supply capacitors with risen ESR

An electrolytic capacitor loses capacitance and gains internal resistance as it ages, and heat accelerates both. Cold it still holds the rail; hot, under switching load, the rail sags far enough for the processor to brown out. It measures fine on a meter and fails under the only conditions that matter.

Cracked solder joints on heavy components

Relays, transformers, inductors and connectors have mass, so they move relative to the board through every heating and cooling cycle. The joint eventually cracks, and the crack opens as the board expands — which is why the fault arrives with temperature and why flexing the board reproduces it.

Regulators without a working thermal path

A linear regulator relies on its tab and the copper around it to shed heat. A dried thermal pad, a cracked solder fillet under the tab, or years of dust and it goes into thermal shutdown, cycling on and off as it cools and reheats.

Processor brownout, which looks like everything else

A sagging rail does not stop a microcontroller cleanly; it makes it misbehave. Outputs fire in the wrong order, counters reset, communications fail. This is the fault most often diagnosed as a bad sensor, a bad thermostat, or a bad motor before anyone measures the rail hot.

Mail-in

Getting the board to the bench

Book it in first. An RMA number comes back the same business day, and the board arrives with the paperwork already attached to it.

Ship to

System Voidz — Repair IntakePO Box 33 (USPS)413 Railroad Ave (UPS, FedEx)Watrous, NM 87753

Use the PO Box for USPS and the street address for UPS or FedEx — a PO Box cannot accept courier deliveries.

Wait for the RMA number before shipping.

Tell us this much on the form

  • Which of the checks above you have already cleared. It is the difference between starting the diagnosis and repeating your work.
  • Photograph the board in place, connectors attached. It is the single most useful thing you can send.
  • The equipment make and model, and the number printed on the board itself.
  • Pack it in an anti-static bag. Bubble wrap alone is not enough.
Book a board in

Questions

Works cold, fails hot — common questions

Can you reproduce an intermittent fault on the bench?

Heat-related ones, usually yes — that is what makes this pattern worth sending in rather than guessing at. The board is run under load and brought up to temperature until it misbehaves, which turns an intermittent field fault into a measurable one.

Say on the intake form that it is heat-related. It changes what gets tested first and it is the sort of detail that saves a day.

It only does it in summer. Is it worth sending now?

Yes, and the winter is the better time. A board with no thermal margin left is already failing; the cabinet is simply not hot enough yet to show it. The fault is the same in January on the bench as it is in July on the roof.

A board sent in a slow month also comes back faster than one sent the week everything breaks.

Should I just replace the capacitors myself?

Sometimes that is genuinely the whole fault, and on an older board it is a reasonable thing to try if you are equipped for it. Two cautions, honestly: lifting a pad on a multi-layer board turns a repair into a scrap board, and capacitors are only one of the four causes above, so a recap that does not fix it has cost you the board.

If it is a board you cannot be without, send it. If it is a spare and you want the practice, go ahead — and if it does not come back to life, it is still repairable from there.

Equipment

Where this pattern turns up most

The same fault, on the equipment it is most often seen on. Each page covers what that class of board does and what else goes wrong with it.