Life Testing Relays

Being mechanical devices which have moving parts and in particular including two metal surfaces (the contacts) which come together repeatedly to provide a signal path all mechanical relays have a finite lifetime. Solid state relays which have no moving parts do not have this limitation provided they are used within their ratings.

Relay manufacturers (including Pickering Electronics who manufacture our reed relays) typically quote a mechanical life and a full power life. The mechanical life reflects the life time limits created by the moving parts including the contact materials as they come together with a light signal load present. For reed relays the limits are imposed primarily by the contacts, for EMR's it can be a combination of the contacts and the other moving parts (many of which are plastic materials) which become worn with an increasing number of operations.

Load Conditions

Wherever a life time for a relay is quoted under hot switch conditions it is always quoted with a purely resistive load as these are the only conditions which can be reproduced consistently for test purposes. The load is created by a power supply with a series resistor that limits the hot switch current.

Normal Life Testing

The mechanical life testing typically takes a very long time to conduct. For a reed relay operated at say one operation per second then doing 1 billion operations is the sort of thing that can only be conducted very rarely - 300,000 hours is a long time. In most (but by no means all) systems relays are not switched that often, which suggests that most relay failures are not down to mechanical life issues.

For load tests the life is expected to be much longer, so the time taken to conduct a test is shorter. Even so it can be a long time, and it has to be expected to be the case if the relays are to be useful in a test system.

Accelerated Life Testing

It is common for a an accelerated life test to be conducted on relays by operating the relays at a much faster and continuous rate than would be expected in a test system. Increasing the operation rate to say 10 or even 100 operations (for reed relays) will significantly shorten the life test time.

Accelerated testing though does have its problems.

The mechanical life tests does not allow parts to fully settle and heat from friction other effects to dissipate. There is evidence based on Pickering Electronics experience that accelerated testing tends to produce mechanical life results which are pessimistic, not always in a predictable way. However, experience suggests that accelerated life testing is useful to compare relays that use a similar technology for mechanical life. It is on this basis that Pickering Electronics evaluates reed blades for inclusion in their reed relays, and how Pickering Interfaces evaluates EMR's to select the product with the best lifetime to use in its switching systems.

On loaded life time tests though a great deal more care is needed because there is a very strong tendency for the breaking and making of signal paths to create heat in each operation, and the faster the relay is switched the more unrealistic the test is. As the contacts make or (especially) break and arc tends to be generated on a voltage load and that arc creates a much higher concentration of power than the loss in a closed contact. As the heat builds then parts operate at unrepresentative temperatures which softens the material and creates more wear than expected.

Contacts in particular erode more readily, but the same issue can arise in the plastic materials that operate the moving contacts of EMR's. Many EMR's capable of switching high signal powers specifically state a maximum operating rate under full load - sometimes being the number of operations allowed per minute (inferring several seconds between hot switch events). The limitations are related to the thermal time constants of the parts and the amount of power dissipated in a typical hot switch event.

For the above reasons accelerated lifetime tests is useful on mechanical endurance tests for product comparison but not used on loaded life tests.

How did we do?
0 out of 0 people found this helpful.