16 articles on the questions that come up before a seal fails and after it has: squeeze and gland depth, compound choice, extrusion, installation damage, cross-referencing 369 AS568 sizes against metric and military part numbers.
How to turn an AS568 cross-section into a working gland: squeeze targets by seal type, groove width from fill ratio, clearance limits and the tolerance stack that decides whether it seals.
Extrusion starts where the clearance is wider than the rubber can bridge. How to check a joint against the pressure-clearance limit, and how a back-up ring changes the calculation.
Roughness decides whether a seal conforms or leaks along the machining marks. What Ra and Rz actually measure, target values for static and dynamic grooves, and why too smooth is also a failure.
Compression set is the most common reason a used O-ring stops sealing. How it is measured, which limits actually apply, and how gland design and compound choice move it.
Nine failure modes, what each one looks like on the removed part, the root cause behind it, and the design change that actually fixes it. A working checklist for maintenance and QA.
In gas service an O-ring can be destroyed by pressure coming off, not going on. Why RGD blisters only appear after shutdown, which compounds resist it, and how to change the decompression profile.
A selection procedure that starts from the fluid, not from the datasheet. Where FKM earns its price, where EPDM is the only answer, and the three media that destroy the obvious choice.
Compatibility tables rate a compound against a fluid at one temperature for one duration. Four variables they quietly hide, why two suppliers disagree, and how to run a 70-hour immersion test that settles it.
VMQ and FVMQ are the only two compounds that combine a wide cold range with useful sealing behaviour. Where silicone wins, where fluorosilicone earns five times the price, and where both lose to FKM.
Shore A hardness controls extrusion resistance, seating force, low-temperature flexibility and friction — but not chemical resistance. What 50A, 70A and 90A each buy, and how to measure it.
Tg, TR-10 and brittleness are three different numbers, and none of them says your seal will work on a cold morning. How low-temperature performance is specified, ranked and designed around.
The cut, the spiral tear and the nibbled edge are not material failures. Six assembly mechanisms, the chamfer and radius dimensions that prevent them, and how to inspect for damage you cannot see.
Rubber ages in the store as well as in service. The storage conditions that matter, why ozone and tension destroy rings on a shelf, and how to re-qualify a batch that is past its date.
The two size systems are not translations of each other. How AS568 dash numbers and ISO 3601 metric sizes are defined, when a cross-reference is legitimate, and the tolerance difference that makes a 'match' leak.
Most wrong orders are not supplier errors; they are measurements of the outside diameter reported as the inside. How to measure a used or unknown ring, and when a caliper is the wrong tool.
Regulatory compliance for a rubber seal is a property of the recipe and the extraction test, not of the polymer. What each approval covers, why 'FDA-approved' is not a regulatory status, and what to ask a supplier for.
Every article is checked against the same data the rest of this site is built from: AS568 dimensions from our size table, military part numbers from our designator index, and material limits from compound datasheets. Where the published standard and common practice disagree, the article says so instead of picking one silently. Nothing here is a substitute for testing the actual parts in the actual application.