Lid thermometer vs grate probe: the gap is not a number you can subtract
Google's own answer box gives the gap between a lid dial and a grate probe as fifty to a hundred degrees, twice, with the sign reversed. Six results, three different signs. That is not a hard question. It is a question with no fixed answer, being answered as though it had one.
Google’s own answer panel on this search asks how accurate a grill lid thermometer is, and answers that dome dial readings run fifty to a hundred degrees above the actual grate temperature. Two questions further down the same panel it asks whether grate temperature is the same as dome temperature, and answers that grate temp is often fifty to a hundred degrees higher than dome temp. Same number. Opposite sign. Same box, four inches apart.
The results underneath do not settle it either. A second Facebook group says the built-in reads about fifteen to twenty degrees lower than grate level. A thread on The Virtual Weber Bulletin Board, still ranking after seventeen years, opens by saying the lid temperature is warmer than the cooking grate temperature, because obviously heat rises. A Big Green Egg thread reports the dome running up to seventy degrees higher, then adds that the longer the cooker sat, the smaller that difference became.
Six sources, three signs, and every one of them stated as a fact about cookers in general. What all of them are doing is handing you a number to subtract. There isn’t one. The gap between a lid dial and a grate probe is not an offset, it is a quantity that moves in four independent ways during a single cook, and a page that gives you one figure for it has told you about one moment on one machine.
What is tested here, and what is not
Analyzed. How each instrument responds is standard instrument behaviour, not an opinion: a bimetallic dial reacts to its whole immersed stem, a thermocouple or thermistor reacts at its tip. The degree figures quoted above are other people’s, reproduced as they were published and attributed rather than absorbed.
Tested. Nothing yet. The run described below shares one afternoon and one set of probes with the measurements on preheating, fire management, tuning plates and the water pan. It needs nothing bought that the protocol did not already require. Method at How We Test.
They are not two versions of the same instrument
A grate probe reads a point. The junction at the tip of a thermocouple, or the bead in a thermistor, is a few millimetres of material sitting in the gas, and what it reports is the temperature of that spot and nothing else.
A lid dial is a coil of two bonded metals on a stem, and the coil winds or unwinds according to the temperature of the stem along its whole immersed length. That length is short, it is usually an inch or two, and one end of it is bolted through the lid into a piece of steel with outside air on the far side. A dial does not read the chamber. It reads an average along a line, and one end of that line is being cooled by the weather.
Which is why asking which one is more accurate is the wrong question, and why it never gets a stable answer. They are not competing measurements of the same thing. One is a point reading in the gas, the other is a line average that includes the vessel wall. A dial can be perfectly calibrated in ice water and boiling water and still be forty degrees away from the grate, because calibration is about whether the instrument reports its own stem correctly, not about whether its stem is anywhere useful.
This also explains a complaint that gets blamed on cheap manufacture. If the same dial reads lower on a windy afternoon, or seems to have drifted once the weather turned, nothing has drifted. Colder air on the outside of the lid pulled the cold end of the stem further down, and the average moved with it. A longer stem reaching further into the chamber suffers less of this, which is one of the few things worth paying for in a dial and one of the few things nobody puts on the box.
Why the sign flips between one thread and the next
Most of the arguing happens over kettles, bullets and kamados, and on those cookers the fire is directly underneath the food. Put a grate probe in the column of rising gas above the coals and it reads hot, hotter than the dome. Move the same probe four inches sideways to the part of the grate with no fire under it and it reads cool, cooler than the dome. The sign of the gap is decided by where the probe went, which is exactly what one of the Facebook answers says out loud: it varies greatly depending on the orientation of the thermometer relative to the coals, and when the probe is not over coals most people say about fifty degrees.
An offset is a different geometry and inherits a different problem. The fire is not below the food, it is beside it, and the gas arrives sideways before it turns for the stack. The vertical gap between grate and lid is smaller than on a kettle for that reason. In its place a horizontal one opens: the firebox end runs hotter than the stack end for the whole cook, by a margin large enough that people buy steel to close it.
So an offset owner is inside a field with two gradients in it, one up the chamber and one along it, and the lid dial samples that field at exactly one point, usually the middle of the lid because that is where the hole was drilled. No single number can describe a surface. That is not a criticism of the dial, it is arithmetic.
The gap shrinks while you watch, and that part is just the warm-up
The Big Green Egg thread contains the most useful observation on the whole page and nobody in it seems to notice: the dome ran up to seventy degrees above the grate, the difference got smaller the longer the cooker was left, and it never closed below about twenty.
That is not the thermometer settling. That is the cooker’s steel coming up to temperature. Early in a fire the gas is already hot while the metal is still cold, so an instrument sitting up in the flow runs far ahead of one sitting near cold mass, and the two converge as the mass soaks. The residual twenty degrees is the real geometric gap for that cooker. Everything above it was a preheat that had not finished.
Which means most of the figures being traded on those forums are undated. A gap measured forty minutes after lighting and a gap measured three hours in are two different quantities, and almost nobody quoting a number says which one they took. A large part of the disagreement in the search results is not disagreement at all. It is people reading the same instrument at different points of the same curve.
The dial also hides the swing, and this is the claim worth testing
An offset burning wood does not hold a temperature, it cycles: it climbs after each split goes on and slides back before the next one, and the size of that sawtooth is set by four things, only some of which you control on the day.
Now put the two instruments in front of that cycle. The thermocouple is a wisp of wire with almost no mass and almost nothing between it and the gas. The dial is a metal coil, at the end of a solid steel stem, bolted into a lump of lid. Heat has to travel down that stem before the coil knows anything happened, and the stem and the coil both have to be warmed before they will move. That is the description of a filter. Fast movements get damped out of it.
The prediction that falls out of this is specific: over a split cycle, the dial should lag the grate probe by minutes and it should show a smaller swing than actually happened. Not a different average. A smaller amplitude, on the same fire, in the same chamber, at the same time.
If that holds, it explains something the forums never resolve. The owner watching a dial reports a steady pit. The owner watching a grate probe reports a pit that swings sixty degrees. Both are honest, both are looking at the same machine, and the instrument is doing the disagreeing for them. It also means every complaint about an offset being unstable is arriving from the people who fitted probes, and every reassurance is arriving from the people who did not.
There is a practical edge to it too. The nastiest moment in a cook with a pan of water in the chamber is the minute it runs dry, when the ceiling holding the pit down disappears while the fire carries on at the same rate. A grate probe shows that climb starting. A lid dial, lagging and flattening, shows it once it has already happened.
| What you are really asking | Lid dial | Grate probe |
|---|---|---|
| Is the meat sitting at cooking temperature right now | No. Twenty to seventy degrees out, and the sign depends on the cooker | Yes, and only for the spot it is clipped to |
| Is the fire climbing or dying at this moment | Late, and it under-reports how far it went | Yes, and this is what it is for |
| Is one end of the chamber cooking faster than the other | Cannot answer it. One point does not report a spread | Yes, by being moved, or by using two |
| Has the pit finished warming up | Yes, and it is genuinely good at this | Only alongside the dial. The answer is in the gap between them |
Row four is the one that gets left out of every teardown of lid thermometers. The dial is a poor thermometer and a good soak indicator: when the gap between the two stops shrinking, the steel has come up and the pit is ready. Nothing else on the cooker tells you that.
What we are measuring
Two of the claims above are reasoning, not readings. That the gap collapses during the warm-up and then settles is borrowed from someone else’s thread and needs confirming on this pit. That the dial under-reports the amplitude of the swing is a prediction from how the instrument is built, and it has never been put in front of a thermometer here.
Both have a shape that one afternoon can confirm or destroy, and neither needs anything bought. The probes are already on the rig because every other protocol on this site requires them.
Lid dial against grate probe, through the warm-up and across four split cycles: planned, not yet recorded
Protocol. Empty pit, the lid dial as fitted and uncalibrated, five probes in the positions used for the grate-spread run with one of them clipped directly beneath the dial so the two instruments are looking at the same few inches of chamber. Logged every two minutes from first light, through the whole warm-up and on across four full split cycles, so the same afternoon yields the gap while the steel is still cold and the gap once it has soaked. Recorded: the gap at ten minute marks through the warm-up, whether it stops shrinking and at what value it settles, then the peak-to-trough swing measured at the grate against the peak-to-trough swing displayed on the dial over the same four cycles, and how many minutes the dial runs behind the probe at each turn. Recorded as well because the probes are already down: the gap under the dial against the gap at the firebox end and at the stack end, which decides whether a dial can stand in for a probe at one spot on this cooker or at none of them. The dial is then checked in ice water and boiling water, at the end and not the start, to separate a badly placed instrument from an inaccurate one.
The figures and the photograph are published once this protocol has been run on the bench. Until then this page carries the method, not the numbers.
What to do instead of subtracting
Stop looking for your offset. If you insist on having one number, take it at steady state after the warm-up has finished, with the probe in the position you actually cook in, and treat it as valid for that position and that fire size only. Move the meat and the number is no longer yours.
Do not buy a better dial to fix this. The dial’s problem is where it is, not what it costs, and a more accurate instrument in the same hole reports the same wrong place more precisely. This is the same conclusion reached from the other direction on the tuning plates page: nothing bolted into the chamber repairs an instrument that is in the wrong position to begin with.
Run the ice water and boiling water check anyway, but understand what it buys. It tells you whether the dial reports its own stem honestly. It says nothing about whether its stem is somewhere you care about. A perfectly calibrated thermometer in the lid still reads the lid.
And keep the dial. It is the cheapest soak indicator on the cooker, it is the one reading you can take from thirty feet away, and once you know your own settled gap it is a serviceable glance-check between probe readings. It stops being useful only when it is the sole instrument on the pit.
Where it fits
A lid thermometer and a grate probe do not disagree because one of them is wrong. They measure different things in different places with different response times, and the gap between them is a moving quantity rather than a constant you can correct for. What the warm-up does to that gap is on how to use an offset smoker. Why the reading cycles at all instead of holding still is on offset smoker fire management. The gap along the chamber rather than up it, and the steel sold to close it, is on tuning plates, and the cheaper damper that hides its own worst moment from a lid dial is the water pan. What you burn is on wood for smoking, and the cut most often blamed on a thermometer when the real fault was timing is brisket. A cooker that reads steady because it genuinely is steady is a pellet grill, which buys that with a different set of compromises. How this site keeps measurement apart from reasoning is on How We Test, and who is running the pit is on About.
That same grate-level gradient has a practical use beyond explaining a thermometer disagreement: how long to smoke a spatchcock chicken uses it to decide which way to face a flattened bird on the grate.
