OT-ish: using negative rail for fan in amp?

[MrStab]

hi,

i started building a stereo LM3886 amp ages ago, which initially had a pre-amp, but it was such a colossal pain in the arse in the enclosure size that all kinds of stuff kept going wrong and i just abandoned it. i've recently revived the project as just a power amp, and after another silicon massacre or two, it's working pretty well. despite a large heatsink, i'd like some airflow on the transformer and power filter/rectifier, so i want to put a 12V fan in there.

there's now a regulated +/-9V power supply going unused (a bit lower than the +/-15V or so you usually see in a pre-amp, as i used some Sanyo EQ chips with a lower max Vcc rating). i used the top half of that as a +9V output for pedals, because why not.

in order to give the +9V output as much available current as possible, for digital pedals or electrical fires, with the -9V regulator going to waste, would there be any major downside to hooking up the positive wire on the fan to 0V and the negative to -9V? noise or death or something? i would probably current-limit the fan further with a resistor, as well.

[bluebunny]

Noise, perhaps (tack it in and have a listen).  Death, no.   

[antonis]

Place a 100nF ceramic cap between 0V & -9V (just because..) 

[MrStab]

i have a 4,700uF electro and 100nF straight outta the rectifier, with just a 100uF on the regulated output. the standard 7809 crap. should that suffice?

my first instinct is that dumping the negative from the fan onto -9V would be a bit like contaminating the drinking water, but the audio path is all pre-regulator at -28V, so maybe it'd be fine? i have no clue about any inherent noise characteristics from fans. i just don't know enough, really.

centre-tap isn't connected to enclosure (which IS safety-earthed), FWIW. i heard that can cause a bit of buzz.

[anotherjim]

I think fan motors can be some kind of brushless type? Ordinarily, DC motors, because they usually have commutator/brush gear, put out lots of crackling interference noise unless fitted with a ceramic cap close to the motor. It would be safe to fit a 100nF ceramic anyway, as close to the motor as possible.

[MrStab]

ah! okay, will def put a cap across it then. only ever used em in computers and a pedal PSU. as it's not a mission-critical thing, i ordered a total piece of crap (Evercool EC8010M12E) so i should expect the worst.

i'm used to the working approach that you dump unwanted signal stuff to ground, but i have a feeling i need to take conventional current vs. electron flow into account here. which way would any noise come out of the fan?

cheers for the info!

[anotherjim]

Ye of little faith. How could a cooling fan called "Evercool" possibly fail?

[MrStab]

[antonis]

If you have a little space availiable, you can make your fan "temperature sensitive" with only a general purpose op-amp, a driver BJT, some resistors, a trimmer and a couple of diodes (or BJTs wired as diodes)..

[MrStab]

i have to be totally honest, Antonis - even though i logically know everything is fine, i'm a bit phobic of adding too much more stuff. it just increases the odds of me overlooking something. i wouldn't care so much if it wasn't £5 a chip!! lol

maybe in future a control circuit like you described could be essential. i'll need to let the amp run at high volumes. the fan just arrived, it's a 3-wire one. no power or current rating, only voltage. i must be dealing with some high-end stuff here. i'm gonna try 2x 1W 33R resistors in series at first, to see how it goes. assuming fans are often 80R or so. 

it looks like i can stick a 47p cap under the sticker. hmm... i'll report back tonight

[MrStab]

so i got rid of the blue wire and managed to get a handy little space for a 47n ceramic by bending the red wire sideways a bit. 100nF was just a bit too big. this fan's on the baseplate, with a metal grill on the outside and a plastic one on the inside, so i need it to be flush on either side. covered it up with some electrical tape.



fan itself seems to work fine between 0 and -9V, no added noise as far as i can tell. it's nearly 2am so i can't test it at a decent level yet. at low volumes, the heatsink is a lot cooler to the touch despite the fan being slowed a fair bit.

here she is. lol



for anyone wondering, this is how i did the stereo/mono thing (the 0.47R resistors are 10W):