DC fan on audio rails?

[Dreams]

I've got a PSU for some channel strips. +/- 15v, 48v. The audio rails are powered by an off-the-shelf Condor 3 Amp supply. The strips collectively pull about 3/4-1 amp per rail.

The thing gets hot. Too hot, I'm thinking.

So I want to add a fan without adding another transformer, ideally. I have a bunch of 12v DC fans, in the range of 1/4 amp. Would it be ok to string some diodes off the +15v rail to knock it down to the 12v that the fan needs? Would that throw junk onto the audio? Is there a good way to decouple the fan?

Should I just take AC off of one of the PT secondary windings and regulate it down?

Do I HAVE to add a transformer for the fan?

Just kinda looking for some outside opinions or pros and cons of each way; I can't really decide which way I want to go....

 

[JohnRoberts]

Coincidentally just this afternoon I hooked up a 12v fan to a 15v rail. It turns out for my particular fan it needed around a 55 Ohm resistor in series to drop 3V across the R and deliver approx 12v to the fan... I actually ended up using two 100 ohm 1/4w Rs. Putting a decent cap across the fan at the junction of the fan and dropping resistor should help keep any noise away from the PS.

JR

 

[PRR]

Agree: resistors are much nicer droppers than diodes for DC fans.

And traditional. Take your car heater apart. There's a resistor to give your Med/Low blow speeds. (Usually hidden inside the fan duct, because a 10 Amp fan's resistor can get HOT.)

AND.... hang a cap from fan + to -, that will filter-off much of the heavy crap that can come from fans. 100uFd at least, more is gooder.

 

[Dreams]

....and now I know.

Ohm's law strikes again.

I didn't even mention using a resistor to drop the volts because I was leaning away from adding something that generated MORE heat, but hey. It works.

And it works well! Everything is nice and cool and the audio is still as quiet as it was. If only I could make the air blow quieter....

Thank you!

 

[Dylan W]

Stupid question time: What's the advantage of a resistor over diodes? Can a diode string (not) act as the resistive component in an RC filter?

 

[joaquins]

The equivalent resistance would be much lower for the same voltage loss, so your filter will filter less noise than with a resistor, plus the resistor is cheaper, and doing a better job is the load is known as yours, so I guess the decision is easy.

JS

 

[PRR]

> something that generated MORE heat

The diode(s) will get hot too.

Same hot as the resistor.

You want to lose 3V. Say a small fan is 0.1 Amps. 3V*0.1A is 0.3 Watts. Diodes don't have any magic transporter to throw this heat elsewhere. That's why hi-current diodes are big, and HIGH-current diodes bolt to heatsinks.

You "could" use switching technology to take 15V only 12/15th of the time and arrive at 12V "without losses". Except switchers have losses in switch, coil, catch-diode, even in their caps. They also buzz (modernly supersonic, but that artifacts into the audio band). And designing a switcher is not as trivial as computing a resistor, even with the "Simple Switcher" tables the chip-makers promote. And it's just a third of a Watt.

 

[tubestation]

Hello Dreams,
This is a very good and quiet Fan  : 8 dBA 1200 RpM  80* 80 mm  12V ,15€. http://www.blacknoise.com/de/products/it/6/Noiseblocker-NB_Multiframe_S_Series_80mm
The M8 S1 Type.
Greetings
Lothar

 

[Dreams]

Thank you Lothar. I already have a fan (and a deadline) for this project, but I'll keep those in mind for the future, or when the owner of this power supply complains about the fan noise.

I'm a little unclear on something. JR said (edit: JS, not JR) that the equivalent resistance would be lower for the same voltage loss using diodes. Resistance is lower for the same voltage drop...current goes...up? Not exactly sure how to phrase this... Probably missing something simple (again).

 

[JohnRoberts]

I'm a little unclear on something. JR said that the equivalent resistance would be lower for the same voltage loss using diodes. Resistance is lower for the same voltage drop...current goes...up? Not exactly sure how to phrase this... Probably missing something simple (again).

I guess you are confused because JR never said anything about using diodes.

As is often the case on the WWW, people don't stop posting advice after you already got a good answer.

JS shared how the apparent impedance of diodes would be lower than a resistor for same voltage drop, so it wouldn't make a very good RxC filter but it's even worse than that, it's not an R at all so will never make a RxC filter but that is TMI.

The heat generated is a function of voltage times current and will be identical. 

JR

 

[Dreams]

Whoops, sorry about that.