Every test engineer knows what they need in the lab. The chamber. The thermocouples. The data logger. The fixture. The test plan. These are not the things on this list.
This list is about the other things — the ones that determine whether Tuesday goes smoothly or becomes a cautionary tale you tell at reliability conferences for the next decade. Some of them are physical objects. Some are habits. One is a specific phone number.
All of them are real.
1. A thermocouple you actually trust
Not the one that's been in the calibration-pending drawer since 2019. Not the one with the slightly bent tip that someone "straightened out" with pliers. Not the one whose label has been crossed out and rewritten twice.
A thermocouple you actually trust. Calibrated, traceable, with a certificate you can find. The chamber's internal sensor tells you what the air is doing. Your thermocouple tells you what the DUT is doing. These are not the same number, and the one that matters for your qualification is the second one. The auditor will ask. The answer cannot be "we assumed they were close."
2. A calibration certificate folder you can find in under 60 seconds
The auditor is already at the reception desk. You have until they sign in. That is not the moment to discover that the calibration certificates for the humidity sensor and the reference thermometer are in a folder called "FINAL_certs_v2_ACTUAL_USE_THIS_ONE" on a shared drive that nobody has touched since the last audit.
Paper or digital, it does not matter. What matters is that you could find it blindfolded, groggy, under time pressure, with someone watching. If you cannot, reorganise it now. Today. The audit is always sooner than you think.
3. The loaded ramp rate calculation — done before the chamber arrives
P = m × Cp × (dT/dt). Five minutes of arithmetic that determines whether your chamber can actually run your test profile with your DUT inside. Not the spec sheet ramp rate. The loaded one.
The spec sheet number was measured in an empty chamber at 23°C ambient. Your lab is 29°C in August. Your DUT is 4 kg of aluminium. These facts interact, and the result is a number that may be significantly lower than what is written on the datasheet. The engineers who discover this after the chamber is installed and the test plan is approved are the ones who tell cautionary tales at conferences.
The Loaded Ramp Rate Calculator does this in about 90 seconds. Use it before you sign anything.
4. The manufacturer's service contact — not the sales contact
You will learn the difference between these two people at approximately 11pm on a Sunday, when the chamber is showing an E-04 error code and the test restart window closes in six hours.
The sales contact will express sympathy and promise to escalate first thing Monday. The service contact will either pick up or call back. These are different outcomes.
Get the direct service number when the chamber is commissioned. Write it down somewhere that is not your work email — which you may not be able to access from home. Put it in your phone under "Chamber Service [manufacturer name]." You will use it exactly once, under circumstances that make you very glad you have it.
5. A deionised water supply with a conductivity meter on it
The humidity system runs on water. The water quality determines how long the humidity system works correctly before something starts depositing scale on the evaporator, the electrodes, or the tubing that connects them.
Most chambers specify deionised water with a minimum resistivity, typically 1 MΩ·cm or higher. Tap water has a resistivity somewhere between 0.0005 and 0.05 MΩ·cm. Using it is not immediately catastrophic. It is slowly catastrophic, in a way that expresses itself as gradual humidity accuracy drift six months later, during a test, when you are already committed to the run.
A conductivity meter costs less than one hour of chamber service labour. The comparison is instructive.
6. A door seal inspection on the maintenance schedule
The door seal is the most underdiscussed failure mode in any environmental test laboratory. It fails slowly. It fails silently. It expresses itself as slight humidity variations that stay just inside the alarm threshold, or as a cold-end performance that degrades a degree or two per month, or as condensation in places that shouldn't have condensation.
By the time the door seal is failing visibly — cracked, compressed flat, pulling away from the groove — it has been failing functionally for months. A quarterly visual inspection and an occasional smoke-pencil or thermal camera check around the seal perimeter is the kind of maintenance that never feels urgent until the day you realise a stability study has been running in an uncontrolled environment for six weeks.
7. An IQ/OQ that was actually run — not copy-pasted from the last one
Every laboratory has one. The Installation Qualification for Chamber 3 that is, on close inspection, the IQ for Chamber 2 with the serial number changed. The model number updated. The date adjusted. The rest left as-is, including the temperature uniformity data that was measured in a different room, at a different ambient, with a different load configuration.
It passes the first review because the person reviewing it is looking for completeness, not for whether the specific numbers are plausible for this specific chamber in this specific installation. It fails the second review, which happens to be a regulatory audit eighteen months later.
Run the IQ/OQ. Measure the actual chamber, in the actual installation, in the actual configuration you intend to use. The paperwork should describe what happened, not what you wished had happened.
8. A second thermocouple on the DUT — not just in the air
This appears twice on this list because it deserves to appear twice.
The chamber sensor measures the air. Your thermocouple on item 1 of this list measures the DUT — but only if you attach it to the DUT. This is a step that gets skipped when the test is familiar, when the DUT is small, when the schedule is tight, or when the previous five runs all passed and nobody thought to check.
The qualification is for the DUT reaching the specified temperature, not the air reaching it. JESD22-A104 says so explicitly. Your customer's reliability engineer may ask about it. The answer "we assumed the DUT temperature matched the air" has a specific effect on the credibility of your qualification data, and that effect is not positive.
9. A deviation log entry written before the excursion is flagged — not after
This is the one that seems like a documentation technicality until it is not.
A temperature excursion happens. The chamber controller logged it. The alarm fired. You responded, stabilised the condition, and documented it. So far, fine. The question is when you wrote the deviation entry — before anyone asked, as a routine part of the response, or after the sponsor noticed the anomaly in the data and sent an email.
These two documents look identical. They are not identical. One is a record of a process working correctly. The other is a record of a problem that someone had to notice before it was addressed. In a regulated environment, the tense of the documentation is part of the data. Write the entry when the excursion happens, not when the question arrives.
10. A copy of the actual standard — not someone's summary of it
Not a blog post about it. Not a LinkedIn carousel summarising the key points. Not the interpretation your predecessor wrote in the test plan seven years ago and nobody has checked since. The actual document.
Standards contain notes. Notes contain prohibitions. JESD22-A104 has a note that explicitly prohibits substituting thermal shock chambers for temperature cycling chambers — because the ramp rate of the DUT matters and a too-fast rate produces unrealistic damage. This is not an obscure detail. It is NOTE 2 of the standard. Most engineers working to JESD22-A104 have never read it, because they have been working from a summary of the standard that omitted it.
The standard is the primary source. Everything else is someone else's reading of it. In a qualification context, your reading is the one that gets audited.
None of these are expensive. All of them have been the difference between a test that passes and a story someone tells at a conference.
The full guides to the technical topics referenced here — writing a test plan that survives an audit, handling deviations and excursions, JESD22 and the DUT temperature requirement — go considerably deeper on each point. But the list above is the version you should be able to recite without looking anything up.