The extractor fan in the downstairs toilet next to the utility room has failed - yet again. The damned thing has only been in existence for about 9 years and it's already killed 3 fans.
Firstly I installed a (British) Ventaxia brand fan which lasted perhaps 3 years, by which time it was past its official 2 year warranty. Quite how watertight that claimed immunity beyond 2 years would be if I tested it legally would remain to be seen but I had a life to live, so I thought I'd simply replace it and move on.
So I then bought a different (British) brand (I forget the name), only to find that when I removed the cover to connect it up, it had the exact same electronics. I simply swapped the boards over, to minimise the effort and damage that otherwise occur when replacing the entire assembly.
Of course, that failed after another 3 years, so I bought a different (British) brand again - Xpelair this time. And of course, 3 years later - yep, it's failed. This time it started to cut in and out and lost the ability to do a timed run-on.
I ordered yet another (British?) brand - this time it claimed a 5 year warranty and 40,000h life. This is the Tornado ST100PR with PIR sensor and timer. at £102 a pop. 4th time lucky??
By way of a sanity check, 24/7 operation for 5 years would suggest 43,800 hours, although it's not going to be running flat out 24/7. That arrived this afternoon.
But what about the old POS?
What was wrong with it? There's not a lot on these things, so let's take a closer look. This model has a run-out timer but no humidity sensor, so the PCB only occupies about half of the available space. Today's versions are different again - although they look superficially the same on the outside, they are "continuous running" ie they tick over constantly and ramp up when triggered by the light input, a humidity sensor or (in some brands) a PIR sensor. That's annoying when fitted in your bedroom en suite. I assume this is due to some change in the planning regulations.
There's a large dropper resistor that appears to be 110R - and has failed open circuit. It's in series with the motor, so is one obvious reason the thing doesn't work.
There's some evidence of overheating on the top side of the board and underneath that discolouration are a load of 68k resistors:
Turns out that there is a pin header with a jumper that allows the motor to be run either directly across the mains or through that resistor, ie a high / low speed selector. Indeed the PCB ident says so, although the handbook is so utterly crap it doesn't mention that at all. There's talk of a setting for "bathroom or kitchen" which we can only guess is referring to this option - but why the actual fuck would you want to say so in the user instructions??? The dickhead who wrote the instructions knows what he meant but as the rest of us aren't psychic yet, it's it chocolate teapot territory.
Rant over. But what it means is that you can move the jumper to the "high speed" position and the motor runs again, albeit at the higher (noisier) setting.
And of course, that isn't the end of the story. There's no time delay at initial turn-on and the fan turns off almost immediately after turning the light switch off ie no run-on delay, which is supposed to be set at the default 15 minute mark.
And of course you can snip out the duff resistor, as it serves no purpose now.
There's some surface mount stuff on the bottom of the (single sided) PCB:
The circuit is fairly simple, as you might expect / hope.
- Half wave rectifier with resistive dropper and zener to regulate the rail for the IC.
- That IC looks like a double opamp or comparator (same pinouts are generally used for both).
- Tiny triac (that SOT-223 device).
- Voltage sense input (when the light is turned on).
- That large 2-3W resistor in series with the motor for 2-speed control (shorted out by the jumper link for high speed).
There aren't many components that can be responsible for setting delays other than capacitors - and there are only 2 on there whole board that could be fulfilling that role. It's these shitty "Chang" branded things and although they seem to measure as having the right capacitance, I assume their ESR has gone through the roof.
I simply can't be arsed to measure them. Instead I simply replaced them with some similarly crap Chinesium caps I got from Aliexpress. "Swpeet" indeed!
There.
And sure enough it works ok-ish now. It's still destined to be replaced by the Tornado when I get round to it but at least the reason for its failure is fairly clear. To summarise - seems these boys employ interns who have all the design expertise of a sixth former. What an utter shit show.
