heatsink fans

[w00t1985]

I dunno if this has been suggested before, but I was just thinking today...

How feasible would it be to have fans on the chip heatsinks, but have the fan's power controled by an extra pot? It would have to be ganged with others for the two (or more) channels, but that way as you turn up the volume, the fan speed increases.

That way, at low volume levels, little heat is produced, and the fan runs slowly and therefore quietly. But at high volume levels, the fan runs faster and stays effective, but the volume of the fan is surpassed by the audio.

Would this be useful? Or would it be cost prohibitive due to the extra pot required? Also, are fans really that needed for the amp chips?

w00t!

[DigitalJunkie]

Hmm,It sounds like a decent idea to me,but there might be one minor problem.
Brushless fans tend to be electrically noisy,and I'd be kind of leary of running the fan power through/near the same ganged-pot as the input signal,You might end up with some hum/noise/hash in the signal from the fan(s).
If you could connect two pots with a rod,perhaps the volume pots in the front panel,and the fan pots mounted back a few inches on a PC board or something,connected with a shaft between the two.That way you could have some space between them,and not get induced noise from the fan pot(s).

But then,with adequate heatsinking,it's sort of unnecessary.

[Geek]

I use brushless computer fans on the heatsinks for the filament regulator in my tube designs. Never had a problem. Mind you, I do keep a good 6" between the PS and tubes.....

My point is that if the fan doesn't bother a very high-Z circuit, it might not touch a solid-state one, unless the PS rails are badly decoupled.

[ofb]

w00t, it's feasible but not exactly useful. dig a bit and you'll see you can get adequate heatsinking with bar stock, or even the right chassis. did you have a specific condition in mind where a fan is needed?

cost of a fan and pot would be diddly; you can likely get what you need out of the scrap bin. including a fan already mounted on an obsolescent computer cpu heatsink. have a look at http://www.7volts.com/ for what the pc crowd does to quiet fans a bit from full speed howl.

can't hurt to play with it, but i've spent so much time trying to make my computers quieter than my late night listening levels that the idea of having a fan in the amp sounds nutters.

also note that computers are starting to catch up with noiseless amp construction. http://www.sharkacorp.com/products/tnn.html i'd love one of those. cat would too.

if you do play with this, note that computer fans need a little more kick to get them going when stopped than they need to maintain low rpm. you'll have to fine tune your volume adjusted control for that, and maybe have a "startup kick" circuit or crank the level yourself on each start.

you *may* not need an indicator for failed and stopped fans. with the lm3875 i think the spike protection kicks in on overheat and your sound will distort or cut out. but be ready to be wrong.

[w00t1985]

Eh, it wasn't for any specific purpose, I was just toying with the idea. Also, as I've just about saved enough to start my own gainclone ventures, I'm not too familiar with the amount of heat they give off. Sure you can read numbers and stuff from the spec sheets, but I learn best from working with stuff.

The only real reason I thought it might be a bit pricey, I saw a few places ganged pots can be pricey. Atleast for quality ones. And you'll want quality pots for volume control, so I didn't know if you could gang an el cheapo pot to a good quality setup.

Also, I'm thinking of making a wood case, so no metal chassis to use as a heatsink also. Guess I'll just play around, and trust when you say they don't give off that much heat

Cheers,
w00t!

[Vikash]

Originally posted by w00t1985
Eh, it wasn't for any specific purpose, I was just toying with the idea. Guess I'll just play around, and trust when you say they don't give off that much heat

The only reason I can see to incorporate a fan into your design would be for cool points. You don't need it with a nice chunk of metal for a heatsink, and you're introducing more issues if you do.

I use a 4mm copper sheet attached to the the aluminium base (also 4mm) and it is adequate. It hasn't been more than warm all the times I have touched it during play.

If you're worried, just use a larger heatsink.

[w00t1985]

Hmm...cool points you say?

I guess I'll just make sure to get a good enough size heatsink, I have a few old PC ones around, maybe I could find a way to use those. But that's for the opinions and such...

w00t!

[Ed LaFontaine]

How about controlling the fan speed with an NTC thermister. They have decreasing resistance with increasing temperature. The value of resistance could be chosen so the fan (powered by a wall-pack) is off at room temperature. The thermister would serve as a temperature sensor and be mounted on the heatsink near the chip. As temperature rises, resistance decreases so fan speed increases. This is just a mental stretching excercise, I haven't put this into use.

[psychosteve]

all sounds good, i thought of using a temp depndant resistor attatched to the sink, to feed the base of a transistor, with the collector and emmitor shorting the fan supply through a shunt resistor to decrese amp through fan. if you pick corect trany and temp. dep. res. you get same effect, variable speed with the temp, plus it will start the fan and then reduce it, or stop it when not needed, I THINK? again though no specific application for this but just playing with the grey cells!! steve.. ..

[BlackDog]

Originally posted by psychosteve
all sounds good, i thought of using a temp depndant resistor attatched to the sink, to feed the base of a transistor, with the collector and emmitor shorting the fan supply through a shunt resistor to decrese amp through fan. if you pick corect trany and temp. dep. res. you get same effect, variable speed with the temp, plus it will start the fan and then reduce it, or stop it when not needed, I THINK? again though no specific application for this but just playing with the grey cells!! steve.. ..

You're right, I've used this kind of circuit in the past and it will work, you'll just need to play with the resistor values a bit. I think the transistor is essential, as most Temp. Dependant Resistor are of too high a value to be able to drive the fan directly.

From experiment I have done at home with heatsinking GC with fans, I found that even a tiny amount of air movement (like the minimimum steady speed of the fan) makes a HUGE difference to the heatsink temperature. -- If it's a finned heatsink, not just a metal bar. --- If you were planning to use a fan for each channel, I would expect that mimimum fan speed would be adequate for all but the tiniest of heatsinks.

Steve