23

I currently work as a Test Engineer. I test (power) electronics.

One test I often perform is something called a "smoke test". Among engineers, this is a common term:

"The phrase smoke test comes from electronic hardware testing. You plug in a new board and turn on the power. If you see smoke coming from the board, turn off the power. You don't have to do any more testing."

Easy to understand, right? But not necessarily professional in a test report. One of my colleagues, the one having to deal with the customers, notified me we should probably change the wording in reports. He's afraid it will be considered pejorative. I don't necessarily agree with him, but sure, let's find an alternative. Except I can't think of an alternative to "smoke test" that's equally unambiguous and clear.

What would be a proper equivalent term for "smoke test"?

Target audience is engineers, salesmen and managers when addressed in written reports.

2
  • Comments are not for extended discussion; this conversation has been moved to chat.
    – tchrist
    Jul 8, 2017 at 14:09
  • This could be really simple as "PIT" for Plug In Test. Jul 9, 2017 at 5:31

4 Answers 4

33

Critical Power-On Failure Test

The system or component tested failed to meet basic power-on requirements, and was unable to be safely activated. No further tests were performed.

Being the son of a member of the USAF, working in an area where electronic systems were frequently tested for this, the best term and phrasing I can think of is the aforementioned. Basically, instead of calling it by the physical-result, you refer to the technical result.

We use the modifier Critical to indicate that it is a significant failure. The test itself is a power-on test, POST, whatever you want to call it. Essentially, we just refer to the idea that power-on failed, and it was a critical failure. This is also a safety test more than anything: you don't continue testing because it's unsafe to test something smoking.

You could also consider the modifier Basic, to indicate that the test truly does not expect much other than a simplistic ability to power the device on without shorting it, catching it on fire, or electrocuting yourself.

4
  • 3
    +1 especially for POST. Pointy-haired managers love acronyms. Jul 7, 2017 at 17:51
  • 4
    @MissMonicaE Thinking about it now, I am technically wrong about POST, because most smoke-tests are not even with the device in the "on" state, but just plugging it in and hoping you don't get zapped, though POST is the most relevant acronym I can think of. :) Jul 7, 2017 at 18:46
  • 3
    Well, POST is commonly meant to mean Power-On Self-Test. That's the basic test a computer for example performs / performed before trying to hand control over to the bootloader. Now leaving out the S for self would result in POT, and I don't know if calling it POT evokes quite the right image... Jul 8, 2017 at 13:14
  • 1
    @Deduplicator to confuse things further, POTS was the Plain Old Telephone System - so I read POT in context as Plain Old Telephone.
    – oerkelens
    Jul 8, 2017 at 13:27
6

Shakedown Test

(MW): a test under operating conditions of something new [...] for possible faults and defects

3
  • 2
    "Operating conditions" means a more involved test than just "does it catch fire when we turn it on?" though. Jul 7, 2017 at 17:50
  • Idiomatically, "Smoke Test" and "Shakedown Test" are synonymous, with "Shakedown" sounding less drastic or catastrophic. If the OP is explicitly talking about a test wherein the pass/fail criteria is "did it catch fire?" (which is not evident based on the wording of the question) then the only applicable phrase would be "Smoke Test'. If, however, it is more of a simple test to check for faults or defects under general operating conditions, I think the answer holds water.
    – vynsane
    Jul 7, 2017 at 18:00
  • 2
    Hm, it sounded to me like the OP is talking about a test under less than operating conditions. I'll ask. Jul 7, 2017 at 18:11
4

For something more complicated, like a factory or a power plant, initial tests like these are referred to startup and commissioning tests.

It seems like smoke test is a start-up test - you start the device, observe it for any obvious problems, and shut it down.

I believe start-up test is equally unambiguous and clear - it can be reasonably implied that the reason for the start-up test is to look for any obvious problems.

14
  • 2
    "start-up test" could also apply to testing that output (and internal) voltage levels stabilize where they should during/after an inrush of current when powered on. But a "smoke test" doesn't imply anything about checking for correction functioning, just absence of catastrophic failure / blowing a fuse. A power supply that keeps its output at 0 volts could pass a smoke test but fail any other kind of start-up test. So it's unfortunately more ambiguous than smoke test. For something with an embedded CPU, a "start-up test" would probably include having that boot up. Jul 8, 2017 at 12:22
  • @Peter Cordes as used this term covers a broad number of tests that include inspection, limited testing (such as confirming that each motor correctly responds to each control center icon), and, yes, actually starting. Some examples here. So the term includes any test in this group, some which may start the plant and many which don't. I understand a smoke test may involve starting something or may not. There doesn't seem to be more ambiguity in one term than the other.
    – Brillig
    Jul 10, 2017 at 14:23
  • @Peter Cordes your specific example was a power supply that keeps its output at 0 volts - there are many startup tests that power supply would pass, so you are not correct when you say but fail any other kind of start-up test. It obviously wouldn't pass a test of actually starting, but there are probably some smoke tests that it wouldn't pass either. It just depends what is involved in a smoke test. If it's, say, a coffee pot and you plug it in and see if it works without any obvious problems then that's a smoke test it fails if the power supply keeps its output at 0 volts.
    – Brillig
    Jul 10, 2017 at 14:27
  • The OP made it pretty clear that they were using "smoke test" to mean "plug it in and look for smoke" (and if so, no further testing is possible). You're talking about looking for smoke during testing of features / functionality, it seems. edit: seems you're using "smoke test" metaphorically. The OP is talking about literal smoke (or sparks or fuse-blowing), and excluding other kinds of basic testing from the definition. Jul 10, 2017 at 14:27
  • @Peter Cordes nevertheless no matter how they use the term internally it has a broadly recognized meaning. He also said it was meant to be interpreted by non-technical people like salesmen, so they would only know the general meaning of the word, not how he used it. If they knew how he used it, there would be no need to find a different term at all. Therefore, my comparison is between the general meaning of smoke test and the general meaning of startup test. Your desire to limit the meaning to his internal use is nonsensical since he's already said that's not how it's perceived.
    – Brillig
    Jul 10, 2017 at 14:30
-3

Sanity Test

Software sanity tests are synonymous with smoke tests.

https://en.wikipedia.org/wiki/Sanity_check

6
  • 5
    OP isn't testing software - and in the hardware world a sanity check is very different. Jul 7, 2017 at 22:20
  • Can you say more about what sanity check implies in hardware? My answer does not mean to say that software and hardware are the same, but that sanity testing is an alternative term in a closely related field and may be suitable outside of software testing.
    – cainlevy
    Jul 7, 2017 at 22:47
  • That wikipedia article should really put this part sooner: "Another, possibly more common usage of 'sanity test' is to denote checks which are performed within program code (e.g. assert()). A build sanity test is a smoke test, but "sanity test" without that qualifier isn't that specific. Jul 8, 2017 at 12:38
  • 1
    See also MikeJRamsey56's better definition of software smoke tests in a comment: short tests of a new build to check that the build succeeded. Wikipedia is claiming this is what "sanity test" means in software terminology, but most people would not agree. Jul 8, 2017 at 12:39
  • 1
    @PeterCordes I've only ever used 'sanity test' in the software world to indicate that it passes all basic UX requirements, like the 'Add' button is actually 'Add' and not 'Delete', for example. Never really to test functionality, just to test that things are mostly sensical. Jul 9, 2017 at 12:00

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