A chamber can hold its setpoint perfectly and still be several degrees out where the load sits. How to specify uniformity rather than setpoint, why the load changes the chamber, what feed water does to an autoclave, and what validation actually involves.
Three of the most ordinary machines in a research laboratory are also the three most often bought on a single number. Somebody writes down a temperature, a volume and a price, and an autoclave, an oven or an incubator arrives. Two years later a result will not reproduce, a sterility check fails, or an auditor asks for the mapping report, and the department discovers that the number on the front of the machine was never the number inside it.
Setpoint is what the controller is aiming at, as seen by one sensor in one place. Stability is how much the temperature at that place wanders over time once the chamber has settled. Uniformity is the spread between the hottest and coldest points inside the working volume at the same instant. They are independent, and only the third one tells you what your sample experienced.
A chamber can be perfectly stable at setpoint at the control sensor and several degrees different in a bottom corner. That is not a fault, it is physics, and every manufacturer knows it. What varies is whether they publish it. When comparing quotations, ask four questions and put the answers in the tender.
Write the required uniformity into the purchase specification as a number with the measurement conditions attached, and make acceptance conditional on demonstrating it on site. That single clause converts a marketing figure into a contractual one.
Every uniformity figure you were given assumes an empty chamber. The moment a load goes in, the airflow pattern changes and the thermal mass changes, and the effect is different for each machine type.
In a forced convection oven, the fan sets up a circulation that the load can block. Trays packed edge to edge, or a single large tray across the return path, create a dead zone whose temperature has nothing to do with the display. Leave the manufacturer's stated clearances at the walls and between shelves, and load the same way every time so results are at least comparable.
In a gravity convection oven there is no fan, so the top of the chamber is warmer than the bottom by design. That is fine as long as everybody knows it and the samples always go on the same shelf.
In an incubator, the enemy is the door. Every opening dumps the chamber air and the recovery is slower than most people assume, particularly in a room whose own temperature is drifting. Stacking plates against the chamber wall makes it worse. If a protocol needs frequent access, two small incubators beat one large one, and the argument for that is throughput rather than capacity.
In an autoclave, the load does not merely disturb the chamber, it is the thing being sterilised, and a wrapped pack or a bottle of liquid reaches temperature long after the chamber does. Liquid loads also need a slow exhaust at the end of the cycle, otherwise the contents boil as the chamber pressure falls. A machine with only a fast exhaust cycle is not a machine for media preparation, and this is a specification to check before purchase rather than after the first flask of agar has emptied itself into the chamber.
Steam sterilises because condensing steam gives up a large amount of heat directly at the surface it touches. Air prevents that. A pocket of air trapped in a wrapped pack, an inverted container or a blocked chamber drain is a pocket that does not reach temperature, whatever the gauge says.
This is why chamber pressure alone proves nothing. A mixture of air and steam at a given pressure is cooler than pure saturated steam at that same pressure. If a machine reports only pressure, it is not telling you the sterilising condition. Insist on an independent chamber temperature reading, and understand which air removal method the machine uses: gravity displacement, which relies on steam pushing air down and out through a drain, or a vacuum assisted cycle, which pulls the air out first and is the only reliable option for wrapped porous loads. If the laboratory processes wrapped instrument packs, the air removal test appropriate to that machine type belongs in the daily start-up, not in an annual service visit.
An oven has the opposite problem. Dry heat sterilisation and drying are not the same process, and a drying oven is not a substitute for a steriliser. Dry heat needs a much higher temperature held for much longer, because there is no condensing steam to carry heat into the load. Departments that use a drying oven as a steriliser because both are hot boxes are running an unvalidated process. If dry heat sterilisation is genuinely required, specify a machine rated for it and validate it as such.
On the teaching side, the EDIBON BIEV computer controlled steam steriliser is built to make this visible: the cycle is instrumented and logged, so students see the relationship between pressure, temperature and time rather than being told about it. The same logic runs through the AUHT/CTS ultra high temperature unit in the dairy laboratory, where the sterilising condition is a process parameter students set and measure.
The most common cause of premature autoclave failure in Bangladesh is not the machine. It is what goes into it. Mains water in most of the country carries enough dissolved solids to scale a heating element and to leave deposits in the chamber, the trap and the valves within a couple of years of daily use.
Follow the feed water table in the manufacturer's manual rather than a general rule, because the requirement runs in both directions. Water that is too hard scales. Water that is too aggressively pure attacks some chamber and pipework materials. The manual states a range for conductivity and for specific ions, and matching a treatment train to it is a solvable problem: our guide to choosing the grade, the treatment train and the consumables covers how those systems are sized and what they cost to run.
Two practical additions. Budget for the treatment consumables in the same line as the autoclave, not separately, because a cartridge that is never replaced makes the treatment ornamental. And ask at quotation stage where the heating element, the door gasket and the safety valve are stocked, since those three are what an autoclave consumes.
Autoclaves and large ovens are the items most likely to be delivered to a room that cannot take them, because the difficulty is in the services rather than the footprint.
Validation is not a certificate that arrives with the machine. It is a set of records the laboratory generates, and its structure is the same for an autoclave, an oven or an incubator.
The independent sensors are the point that gets skipped, and it is the one that matters. A machine's own probe is the thing being checked. Sensors used for mapping need their own calibration certificates with a traceability chain, which is the subject of building a laboratory record that survives an audit.