Guide
How to Calibrate an Oven, and What Calibration Cannot Fix
Calibration corrects exactly one fault: a steady gap between the number you set and the average the cavity holds. Measure that average properly first, because a swinging oven, a slow oven and a hot-spotted oven all feel wrong and none of them are calibration.
Before you change any setting, establish the oven’s average — not a reading. Calibration adjusts one thing only: a consistent gap between the temperature you ask for and the temperature the cavity actually holds over time. It cannot correct an oven that swings widely, an oven that takes forever to get there, or an oven that browns one corner of a tray and not the other. Those are three separate faults that all feel like “the oven is off”, and an offset applied to any of them just moves the problem somewhere else.
So the work is in the measuring. The adjustment itself takes under a minute.
What the adjustment actually does
Your oven decides what to do from a single temperature sensor in the cavity. The control reads that sensor, compares it to your set point, and switches heat on and off to hold the two together. A calibration offset is a number stored in the control that says: whatever the sensor reports, treat it as being so many degrees higher or lower before you act.
That is the whole mechanism, and it explains the limits. An offset shifts the target. It cannot make the cavity heat faster, cannot make it hold more steadily, and cannot change where the hot corner is. If your complaint is about anything other than the average being wrong, calibration is the wrong tool.
Measure the average, not a moment
An oven does not sit at one temperature. It falls a little, calls for heat, overshoots a little, and drifts down again, endlessly. A thermometer glanced at once will catch a peak or a trough and give you a number that is real but meaningless. What you want is the middle of that oscillation.
- Use a probe thermometer with a lead and an outside display if you have one. It is the single biggest improvement to this test, because opening the door to read a dial dumps the cavity and ruins the measurement you were taking.
- Put the sensing tip roughly where the food goes — centre of the middle rack, clear of the walls, the floor and the racks themselves. Metal touching the probe reads as metal, not as air.
- Set 350°F and leave the door shut. Ignore the preheat tone; it reports that the air arrived, not that the cavity walls, the racks and the door glass have caught up. Give it a further twenty minutes.
- Now record a reading every three minutes for half an hour, still without opening the door.
- Average the highest and lowest numbers you wrote down. That average is what your oven means by 350.
Test it the way you cook, too. A cavity holding a stone or a heavy tray swings less than an empty one, and if everything you bake goes in on a stone, the empty-oven result is not the one that matters.
Reading the pattern you wrote down
A steady gap
Every reading sits low, or every reading sits high, and the spread between them is unremarkable. This is the case calibration exists for. The cavity is behaving correctly and simply aiming at the wrong number.
A wide swing
Readings scattered a long way apart, with the average roughly right. Nothing here is a calibration fault. The control is letting the cavity fall too far before it acts, or the heat is arriving in bursts that are too coarse — a sensor whose reading has become erratic, a relay on the control board, or on a gas oven an igniter no longer drawing enough current to open the safety valve cleanly.
A slow climb, or a peak that fades
The oven gets there eventually, or gets there and then loses ground. That is a heat-production question — a cracked bake element, an igniter on its way out, a door that no longer seals, or a cavity fighting a vent it should not be fighting. Adding an offset makes the display more wrong, not less.
Where the adjustment lives
On an electronic oven it is a button sequence, usually a press-and-hold on the bake or temperature control that puts the display into an adjustment mode. The sequence differs by manufacturer and by series, and the use and care guide for your exact model gives it. The available range is deliberately limited in both directions, which is itself a hint about how large a correction is reasonable.
On an older gas oven with an analogue thermostat there is no software involved. Pull the knob straight off; behind it, on the back of the dial skirt, sit two small screws and a moveable pointer against a notched scale. Loosen, move by one notch, tighten, and re-measure. The manual states what one notch is worth on that thermostat — do not assume, because it varies.
Either way, re-run the whole measurement afterwards. An adjustment you have not verified is a guess with more confidence attached.
Four things this will not fix
- Hot spots. Uneven browning across one tray is airflow and radiant geometry, not the set point. Cover a sheet with slices of plain bread, bake until they colour, and photograph the result — you now have a map of your oven and can rotate accordingly.
- Preheat time. How long the cavity takes to arrive is about how much heat the appliance can make and how much mass it has to warm. The target has nothing to do with it.
- Browning and broiling. Broil is radiant heat from directly above, largely outside the thermostatic loop. An oven that bakes correctly and browns poorly is describing its broil element or its rack height.
- A door that no longer seals. A gasket that has hardened, or a hinge that has let the door sit proud, leaks heat continuously. The oven compensates by running more, the average may even come out right, and the food still behaves oddly.
When the correction you need is too large to be one
If the measured average is out by more than the adjustment range allows, or the size of the error changes depending on how long the oven has been running, stop. That is a sensor that has drifted or a control that is misreading it, and dialling in an offset to cover it buries a fault you will meet again at a worse moment. A cavity sensor is a resistance device whose value climbs with temperature; a technician checks it against the resistance table in the service literature for that model, which settles the question in minutes.
The same applies to anything that smells hot, trips a breaker, or has started behaving differently since a self-clean cycle. Self-clean does not shift calibration, but the temperatures it reaches are a stress test that finds whatever was already marginal, and thermal fuses and tired sensors are the usual casualties.
Where this bench stops, and where it does not
To be straight with you: we do not repair cooking appliances. This bench works built-in refrigeration — refrigerators, freezers, ice makers and wine cabinets — and an oven call is one we turn down rather than take badly. The guide above is here because the question deserved a proper answer, not because it is a service we are selling.
If the appliance actually misbehaving is the cold one, that we can help with. The diagnosis costs $95 flat, and the figure is taken back off the bill if you decide to have the work done. Phone or book from the header of this page; what happens on a refrigeration call is written out in detail elsewhere on the site.
Asked on the phone
The thermometer reads low but everything I bake comes out right. Should I still adjust it?
No. An oven is calibrated correctly when the food is correct, not when a dial in the cavity agrees with the display. Recipes are written for real ovens and most of them tolerate a modest offset without anyone noticing. Adjust when something has changed — bakes that used to work stopped working, or a new appliance behaves unlike the one it replaced. Chasing a number for its own sake usually ends with two errors instead of one.
The oven is fine on bake and wrong on convection. Is that a calibration problem?
Almost never. Convection is a separate heat source and a separate fan, so it has its own ways of being wrong: a convection element that has failed while bake carries on, a fan running slow enough to stir the air without moving it, or a recipe that has not had the usual convection reduction applied. Test the two modes separately and treat them as two results, because one offset setting cannot correct a fault that only appears in one of them.
How often is it worth re-checking?
Not on a schedule. Check after the appliance has been moved, after a sensor or control board has been replaced, and whenever something you have baked for years suddenly needs a different time. Those three moments are when a real shift tends to have happened. Between them, an oven that is producing the food you expect is telling you everything a thermometer would.