1
\$\begingroup\$

I am building a system for measuring some parameters in a climate chamber. The test runs from -55 to +125°C. I need to switch some signals between multiple devices under test using relays. I want to avoid using exotic or expensive parts.

I can find "generic" relays rated from -40 up to +105°C. I will be switching low voltages (below 12 V) and low currents (below 100 mA) most likely with a relay rated for 230 Vac and 16 A. Switching speed is not important.

I know that the plastic can melt, but I wonder what else can happen to a relay when it is used at a higher and lower temperature than specified?

\$\endgroup\$
8
  • 3
    \$\begingroup\$ Do your relays need to be in the temperature chamber too? When I run stuff in a temp chamber, I typically run wires into the chamber carrying signals, but switch them with relays or something outside the chamber, unless those switches are part of the product. \$\endgroup\$
    – Bort
    Commented Oct 18, 2019 at 18:13
  • 1
    \$\begingroup\$ Yes, they need to be in the chamber, because otherwise I will need to route ~500 wires outside. \$\endgroup\$
    – filo
    Commented Oct 18, 2019 at 18:14
  • 1
    \$\begingroup\$ When the datasheet does not cover it, asking here is a good idea. But often you can just try it and see what happens. Don't be afraid to try things. \$\endgroup\$
    – user57037
    Commented Oct 18, 2019 at 18:19
  • 1
    \$\begingroup\$ @mkeith ABS's glass transition temperature is 105C. Coincidence? Nylon is 70C which is a lot lower than I thought it would be. I thought it was higher than ABS. \$\endgroup\$
    – DKNguyen
    Commented Oct 18, 2019 at 18:20
  • 2
    \$\begingroup\$ @mkeith Well, yeah it's close enough and his application his on-off so I would try it too as long as it's not nylon. The relay may be also be potted into the enclosure. Tougher to know what that is made of. He could also just go for the transparent relays. Those should be made out of either acrylic or polycarbonate. If it's polycarbonate he's good. If it's acrylic it's a bit more of a crapshoot since the Tg varies from 85C to 165C depending on composition. But either way you can see inside it to know if something is wrong. \$\endgroup\$
    – DKNguyen
    Commented Oct 18, 2019 at 18:31

1 Answer 1

Reset to default
3
\$\begingroup\$

I wonder what else can happen to a relay when it is used at a higher and lower temperature than specified?

The first thing that it will do is cease operating in a manner consistent with the datasheet. This means that the component is no longer guaranteed proper functionality. Many manufacturers also test components to the operating temperatures listed in the datasheet, and if you run it outside of that temperature, the results will be unpredictable.

Many epoxies and plastics start to degrade and deform at ~120C, the lose their properties. The metals will probably be fine.

For a relay, there is also additional power dissipated in the coil, which raises the temperature beyond the ambient (environmental) temperature, so I would imagine that most relays are also have a de-rated operating temperature, from the storage temperature for this reason.

If the design is tolerant for failure (meaning you don't mind failure and replacing the parts), then run it outside of the operating temperature range. If you need reliability, then don't.

\$\endgroup\$
3
  • 1
    \$\begingroup\$ The data sheet point is well taken, and users often don't understand it, or don't believe it. I'd add that another option is to "simply" test the part under the operating conditions, but suggest that would be a very cost prohibitive path to take. Finding the right part for the right job would be the right answer 99,999 times out of 100,000. \$\endgroup\$
    – Scott Seidman
    Commented Oct 18, 2019 at 18:41
  • \$\begingroup\$ Scott, if you have the relay and the thermal chamber and you are waiting with bated breath to find out if you can use it, then running a test is reasonable. If you have to buy parts anyway, might as well buy parts that are rated to 125 unless such are unobtainable. In OP's case, cost could factor into it. Mentioned 500 wires... So maybe it could be worth it to try the cheap parts first. Especially if they only have to hold up for one test run in the chamber. \$\endgroup\$
    – user57037
    Commented Oct 18, 2019 at 18:48
  • \$\begingroup\$ I've had to 'de-rate' parts from the datasheet once or twice, and ran them beyond their conditions, but usually it was for prototyping and we didn't care if we caused a failure. If at home or in the lab, one can tolerate a little failure, if going into a production environment, the last thing you want is failure. \$\endgroup\$
    – Voltage Spike ♦
    Commented Oct 18, 2019 at 20:15

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.