Technical Article
MIL-STD-202 Method 107 · The Fixed Lettered Conditions A Component Qualifies To, And How A Chamber Runs Them
MIL-STD-202 is the test rulebook for electronic and electrical components, the resistors, capacitors, connectors, and relays a designer buys by the reel, and its Method 107 is the thermal-shock test those parts qualify to. Where an equipment-level shock test tailors its extremes to a platform’s life, Method 107 does the opposite: it names a small set of lettered conditions, each with fixed temperature extremes and a set number of cycles, so that a part proven to a condition means the same thing to whoever bought it. The chamber’s task is narrow and exact, to reproduce a named condition, cycle after cycle, on a tray of small parts.

Discrete components qualified to a lettered shock condition
A component qualifies to a fixed letter, not a tailored spec.
MIL-STD-202 is the test book for electronic and electrical components, the parts a designer orders rather than builds, and Method 107 is its thermal-shock test. It exists to qualify a part’s construction against sudden temperature change before that part is trusted in anything.
The aim is not to study one assembly’s life but to certify a component class: that this family of capacitors, this line of connectors, this style of resistor can take the shock the standard names and keep working, so a designer downstream can specify it and move on.
The heart of Method 107 is a short list of lettered test conditions, and understanding why they are lettered explains the bulk of what the standard is. Each condition, A, B, and onward, names a fixed pair of temperature extremes and a fixed number of cycles, and the engineer running the test does not choose the numbers; the engineer chooses the letter, and the letter chooses the numbers. A milder condition might swing a part between a modest cold and a modest heat for a handful of cycles; a harsher one widens the extremes toward the deep cold and high heat a rugged part must take and may lengthen the count, each step up the alphabet a step up in severity. The reason for this rigidity is the opposite of arbitrary. A component is a commodity, a capacitor or a connector bought by the million and dropped into a thousand different designs by a hundred different customers, and none of those buyers can afford to invent a shock test of their own and then argue about whose is harder. A standardized condition cuts through that: when a maker says a part is qualified to a given condition, every buyer reads the same meaning, can set it beside a rival part qualified to the same letter, and can trust the qualification without repeating the work. For the chamber the consequence is concrete. It does not improvise; it holds the lettered conditions as recipes and reproduces the named one exactly, the same extremes, the same dwell, the same cycle count, run after run, so that a part tested in one lab to a condition has met the same trial as a part tested in another, and the entire meaning of the letter rests on every chamber reading it alike.
The damage itself is the same mismatch of expansion that any thermal shock exploits.

A lettered condition is not just a pair of temperatures; it spells out the whole trial. It fixes the high and the low the part will see, the number of times it will cross between them, the time it dwells at each end, and the limit on how long the transfer may take.
The cycle count is modest and set: a handful of complete crossings, often five, is enough to bring out a construction flaw, since a sound termination that survives a few hard transitions will survive many. The standard names the figure so no lab can quietly run fewer.
The dwell is sized to let the small part reach the extreme and settle, a span of minutes rather than the hours a heavy assembly needs, and the transfer between extremes is held to a short ceiling so the change stays sharp.
Every one of those numbers belongs to the condition, not to the engineer. To run the part to a letter is to accept the letter’s extremes, its cycles, its dwell, and its transfer as one package, and a test that altered any of them would no longer be that condition at all.
The conditions climb in severity as the letters advance. The gentler ones pair a moderate cold with a moderate heat and ask for a modest run of cycles; the harsher open the gap to a punishing cold and a punishing heat, and may demand more crossings to earn the grade.
A buyer reads the part’s intended world and picks the matching letter. A part bound for a benign indoor life qualifies to a mild condition; one bound for an engine bay, an avionics rack, or the open field is held to a wider, longer one.
The point of the ladder is comparison. Two parts qualified to the same condition have met the same trial, and a part qualified to a harsher one has plainly survived more, so the letter becomes a shorthand a procurement engineer can trust at a glance.
This is the deepest difference between a component standard like 202 and an equipment standard like the one that governs whole systems. The equipment standard tailors its test to a single platform’s mission, because that platform is one of a kind.
A component is the opposite of one of a kind. It is a commodity, made by the million and bought by customers who will never meet, each dropping it into a design the maker has never seen.
A test that bent to each of those uses would be useless, since no two buyers could compare results or trust a qualification they had not run themselves. The fixed condition is what lets a component be a known quantity across the whole market.
In trading flexibility for a common language, then, 202 does on purpose what an equipment test would never want. Its rigidity, which would be a flaw there, is its whole virtue here, because the value of the part rests on every buyer reading its qualification the same way.
A component rarely meets Method 107 by accident; it meets it because a document sent it there. The part’s procurement specification, its detail spec or its datasheet, cites the method and the condition by name, so the line in the paperwork that reads as a code is in fact the instruction to run this exact shock.
That citation is how the requirement travels. A designer who needs a rugged capacitor writes the condition into the part’s spec, the maker qualifies to it, and the qualification answers the spec, so the letter ties the buyer’s need, the maker’s test, and the chamber’s recipe into one unbroken reference.
A qualification speaks for more than the pieces that were tested. A sample is drawn from a lot, run to the condition, and judged against an acceptance rule, often that not one of the sample may fail, and the result is read as a verdict on the population the sample came from.
That is what makes the test economical at component scale: a maker need not shock every capacitor it ships, only enough to stand for the rest under the standard’s sampling, so a whole production run is qualified by the few that proved its construction.
Fast in, dwell, fast out, repeat.

For the chamber, a fixed standard means a narrow, exacting job. It is not asked to improvise a profile but to reproduce a named one, and reproduce it the same way every time.
It holds the lettered conditions as set recipes, each carrying the exact high and the exact low the letter names, so an operator selects a condition rather than dialling numbers that might wander from the standard by a careless degree.
It runs the cycle count the condition fixes, no fewer and no more, since a chamber that lost count or cut a cycle short would void the qualification it was meant to grant.
It hits each extreme and holds the dwell the condition sets, with the transfer between them kept short so the change stays a shock and does not soften into a slow lean.
And it does all of this run after run without drift, because a qualification rests on the promise that the part met exactly the named condition, and only a chamber that holds that condition the same way every time can carry the promise.
A component is small and light, and that changes the dwell. Where a heavy assembly lags the air through a long soak, a resistor or a chip capacitor reaches the extreme almost as fast as the air around it, so the hold at each end can be brief.
It also means a tray can carry many at once, hundreds of small parts shocked together to the same condition, which is how a lab qualifies a lot rather than a single piece.
Method 107 is generally run in air, the parts moved between a hot air zone and a cold one, which suits the general run of components well. A handful demand the fiercer liquid shock, the parts whose own standards call for it, but for ordinary mil-spec components the moving air of a two-zone chamber is the shock the method intends.
It helps to place Method 107 in the family it belongs to. It is the component-level shock test, sitting below the equipment-level standard that proves a finished system and apart from the microelectronics standard that governs bare die and packages.
A part clears 107 on its own, is soldered into a board, and the board is built into a product that must pass its own equipment-level shock; each level tests what the one below cannot, and 107 owns the rung where the loose component is proven before anything is built from it.
A part clears 107 alone; the product built from it still has its own shock to face later.
Passing Method 107 is a statement about how a part is built. The shock pulls on the joints where a component is weakest, its terminations, its seals, the bond between its body and its leads, and a part that comes through has shown those junctions hold.
It is a qualification, not a screen: the test proves a design and a construction can take the shock, rather than weeding weak units from a production run, and a part that earns its condition has earned the right to be specified by the letter alone.
Method 107 turns thermal shock into a contract written in a single letter. The condition fixes the extremes, the cycles, and the dwell; the chamber reproduces them faithfully; and the qualification that results means one thing to every buyer who reads it.
A component proven to its letter is a component a designer can trust without ever running the test again.