Scope. This page provides a controlled reader-run protocol and does not claim site-run measurements. For cooking, cooling, storage and reheating, follow the current official guidance linked on the page.
Most instructions that involve holding food at a steady low temperature come with an implied shopping list. Somewhere in the method there is a circulator, and the reader who does not own one closes the tab and cooks the thing hot instead.
The machine is not the technique. The technique is a large mass of water sitting at a known temperature, and you can assemble that from a pot, a hob and a thermometer you probably already own. What you cannot assemble for free is the stability a circulator provides, and the honest version of this article is about exactly how much stability you give up, where that matters, and where it does not.
What a water bath is doing that a pan is not
A pan transfers heat to food through a metal surface that is much hotter than the food you want. The base of a pan on a low setting can sit far above the temperature of whatever is in it, and the difference between the two is where scorching, curdling and overshoot come from. You manage that gap by stirring, by moving the pan on and off the heat, and by paying attention.
A water bath removes the gap. Food suspended in water cannot get hotter than the water, no matter how long it sits there, because there is no surface in contact with it that is hotter than the water. That single property is what makes the technique forgiving. The consequence is worth stating plainly: in a water bath, time stops being a risk and becomes only a cost. Leaving something in for longer than intended changes the result far less than leaving it in a pan for longer than intended.
Water also has a large heat capacity, which is the second half of the trick. A few litres of water at a target temperature resist change. Add cold food and the temperature falls, then recovers. Nudge the hob up and the temperature climbs, slowly. That sluggishness is annoying when you want to change something and extremely useful when you do not.
The related technique, holding a pan rather than a bath at a low temperature, is covered separately in holding a low temperature on an unwilling hob. A bath is the easier of the two problems, which is why it is the one to start with.
The setup
Nothing here is special equipment, but two of the choices matter more than people expect.
- The pot. The widest heavy-based pot you own, with a lid. Width gives you room to keep food off the base; mass gives you thermal inertia; the lid cuts the evaporative losses that otherwise drag the surface temperature down and make the hob work harder.
- The water volume. This is the variable most guides omit entirely, and it is the one that decides whether the method works. More water means slower drift, slower recovery after you add food, and slower response to the hob. Any drift figure quoted without a volume attached is meaningless, so record yours with the volume beside it.
- Something between the food and the base. A folded tea towel, a metal rack, a trivet, an upturned saucer. The base of the pot is the only part of the system that is hotter than the water, so nothing you are cooking should touch it.
- Probe placement. One probe in the water, mid-depth, away from both the base and the wall. If you can manage a second probe in the food, use it, because water temperature and food temperature are different questions and the second one is the one you actually care about.
- The heat source. The lowest setting your hob will hold, plus a diffuser if the lowest setting still overshoots. On some equipment the lowest setting is genuinely too much, and that floor is a property of the appliance rather than of your technique.
The working method is simple and slightly counter-intuitive: heat the water past your target, then turn the heat off and let it coast down to the target before the food goes in. Chasing a target upwards means overshooting it. Approaching from above means you arrive at the number with the heat already off, which is the state you want to be in.
Holding the target: what the drift actually looks like
This section is a validation protocol, not a claim that an improvised bath is suitable. Map the temperature at the centre and edges with a calibrated thermometer, record the largest deviation and abandon the setup if it cannot follow the validated recipe's required conditions.
The protocol is set. A stated volume of water in a stated pot on each hob type, brought to a target, held with the smallest possible input, logged continuously for an hour with a calibrated probe, then logged again through the disturbance of adding food of a stated mass and starting temperature. Record it in the fields below.
Worksheet fields: pot and lid; water volume; hob type and setting; target temperature; highest and lowest reading across the hour; food mass and starting temperature; lowest reading after the food went in; time to recover to the target. Copy these labels into a notebook or spreadsheet and fill one row per run.
Two things to expect, stated in advance so that your log can contradict them: that a lid changes the numbers more than the hob setting does, and that the free method holds a band rather than a number. A circulator advertises a set point. A pot on a hob gives you a range you are inside, and the size of that range is the entire question. Record the range, including the runs where it was embarrassingly wide.
The jobs this is good enough for
A band is sufficient whenever the food's own tolerance is wider than the band. That covers a great deal of everyday cooking:
- Melting and holding chocolate or butter without splitting or seizing.
- Warming a finished sauce so it is hot at the table rather than tepid, without cooking it further.
- Infusing dairy or oil with aromatics at a controlled temperature instead of on a guessed low flame.
- Gently poaching fruit, so it softens without collapsing.
- Bringing eggs, butter or dairy to a working temperature quickly before baking.
- Holding a custard base steady while you assemble everything else.
- Softening dense vegetables in fat at a temperature that stays well below browning.
These share a shape. In each case, being a little above or a little below the target changes the rate at which something happens rather than whether it happens. Those are the jobs a pot of water does well, and they are more of what home cooks actually need than the sous vide marketing suggests.
The jobs where the drift matters
The band becomes a problem when the food's tolerance is narrower than the band, which happens in three situations.
When the target window is narrow. Egg proteins set across a range that is small enough for a few degrees of drift to move you from one texture to another. If the drift band is wider than the doneness window, the method cannot hold the window, and no amount of attention fixes that.
When the hold is long and unattended. Drift over an hour is one thing. Drift overnight, with nobody watching and the kitchen cooling around it, is another. The free method assumes a person in the room.
When the holding time itself is the point. Any application where the duration at a temperature is doing work that matters for safety rather than for texture is outside what we will improvise here. We do not publish holding times, pasteurisation values or cooling rules from our own reasoning. Those come only from named food-safety authorities, and for anything other than a short, attended, straightforwardly-cooked job you should follow a validated method from one of them rather than a pot on a hob.
That boundary is not a legal disclaimer. A pot on a hob that holds a band rather than a number is genuinely unsuitable for a technique whose whole safety basis is holding a specific number for a specific duration, and the honest thing is to say so rather than to imply a workaround.
When buying a circulator becomes rational
The purchase makes sense when at least two of these are true, and rarely otherwise:
- You want a set point, not a band, because the window you are working in is narrow enough that the difference is visible on the plate.
- The cook is long or unattended, so there is nobody to nudge the hob.
- You want the same result repeatedly, and "the same" means the same number rather than the same rough region.
- You are running the technique often enough that the setup effort of the free version becomes the real cost.
If none of those apply, a pot of water and a probe will do the job and the money is better spent on the probe. We take that argument apart properly in whether you need a circulator at all, and the three approaches are set side by side as a protocol in sous vide against a low oven and a pot of water. The general question of where equipment money actually changes a result is on what's worth paying for.
To be explicit, because this is the failure mode of every budget-alternative article: this is not a controlled circulator. It is a way of getting most of the benefit for none of the money on a subset of jobs, and it is a way of finding out whether you like the technique before spending anything on it.
In Plain English
A water bath is a pot of water held at roughly the temperature you want, with the food sitting in it and off the base. It works because nothing in the pot is hotter than the water, so the food cannot run away from you the way it can in a pan.
Who it helps: anyone who has read a low-temperature method, seen a piece of equipment in the ingredient list, and given up. Also anyone who owns a circulator and has never checked what the free version would have done instead.
When to use it: melting, warming, infusing, softening, holding, and gentle poaching. Anything where a few degrees either side changes the speed but not the outcome.
When not to: narrow doneness windows, long unattended holds, and anything where the holding time is doing safety work. For that last category, follow a named food-safety authority's current guidance rather than improvising.
What to do next: get a probe you have calibrated, put a measured volume of water in your heaviest pot, and find out what your lowest hob setting actually holds it at. That single experiment tells you more about your own kitchen than any article can, including this one. The rest of the technique library assumes you have done it.
Next Step
Run this protocol on each hob type you use and keep the drift logs in full, including runs where the free method failed to hold anything useful. This page supplies the complete protocol rather than promising later site results.
Related technique: The Four Temperature Bands Every Cook Should Know.