Why can't I test a smoke detector with smoke?

[Paul Uszak]

I just came across the user instructions for my smoke detectors and am surprised. I've attached a small section of the instructions. They're from a mainstream manufacturer and labelled with a British Kitemark. You can see that I'm not supposed to test it with flame, smoke nor heat, and the results might be misleading! How can this be?

The self test button clearly doesn't introduce ionised particles into the chamber of an ionisation type, lumps of smoke into the sight line of an optical type, nor heat up the sensor of a heat type detector. I assume that it only tests the output circutry and the thingie that makes the annoying noise when I cook. So if cooking sets it off, why can't I test it so? How else can I reliably test that the sensor components are actually working and not full of paint /spiders? I bet that they test smoke detectors at the factory in a real /well simulated fire environment...

 

[rootone]

I have found that some of those so called smoke detectors are pretty useless, I don't know how much veracity is to be expected of the 'Kitemark'.
A while ago all of the 'smoke detectors' in my flat were replaced by my landlord as 'routine maintenance'.
Every one of them eventually detected 'smoke' when nothing even vaguely like smoke was around.
One of them, (the first one). started to raise an alarm just because of bright sunlight, closing the window curtain would shut it up.

 

[A.T.]

You can see that I'm not supposed to test it with flame, smoke nor heat, and the results might be misleading! How can this be?

I guess, if done too often, it could cover the sensors with soot.