[star882]

Has anyone thought about that? Semiconductors exhibit lower noise levels at low temperatures. (Which has been proven by numerous overclockers.) Sure, it's expensive and requires special tools and skills, but aren't many people here already willing to spend a lot to gain a little?
http://www.xtremesystems.org/forums/...splay.php?f=80

Or to be less extreme, peltiers can work almost as well for small amps and is much cheaper. Even watercooling can be a step up.

 

[cire]

haah...maybe i should try sticking my AC chiller's CPU block on some opamps....

 

[strogg]

seems like just a tad overkill for me. only a tad...

but while we're on that topic, why not try some liquid N2? sure it'll shorten the lifespan of the IC to 6 hours, but it'll be worth it right? LN2 is cheaper than milk, and a couple opa627's cost only $32 each, right?

:d

seriously, though, it doesn't seem like that bad of an idea, but you also have to account for the condensation that's going to occur around the chip. it's probably not worth it at all. but hey, this is the guy who still has a thermaltake volcano 7+ heatsink lying around waiting to be fitted for a MOSFET. I'd rather use the 120mm thermalright, but those cost money

although now peltiers seem like an interesting idea. they do make peltiers small enough for mosfets, but removing that heat + mosfet heat will be rather difficult. watercooling will probably work the best. get some small ID tubing, then get a mini fountain pump (one of those $15 ones that have a 24" head), and fashion your own waterblock/radiator with a couple copper blocks or one copper block and the smallest radiator in the market. it should be quiet enough that no one would notice.

 

[dsavitsk]

I don't know about you, but my amps sound better when they are hot.

 

[Emon]

Quote:

Originally Posted by star882 Semiconductors exhibit lower noise levels at low temperatures.

I do believe that depends on the type of semiconductor being used.

 

[kramer5150]

I would think you would have to worry about moisture condensation on some of the metal parts... if you chill / cool them too much.

Garrett

 

[cire]

Quote:

Originally Posted by kramer5150 I would think you would have to worry about moisture condensation on some of the metal parts....

thats what seal string, dielectric grease and RTV/conformal coating are for

 

[markot86]

I feel like the added noise of the TEC system would kill any benifit of a low noise amp; Why not just stick a small heatsink on there with some as5 and try that out? no added noise and maybe a lower noise level =P

 

[JahJahBinks]

Most ICs will stop working once the temperature reaches -40C, military grade can do -55C. So LN2 is out. If you ever want to know how to make a chip that can survive even lower temperature, get a job at NASA and you might learn what they use on the space probes and mars explorers. I looked at the data sheet of OPA627, one of the most used opamps in DIY, the only thing I can think of that might affect an music signal is GBP and slew rate, but the change in temperature between -50C and 25C is equivalent to a change from 20MHz to 16MHz in terms of GBP, that's insignificant considering the low audio frequency, and the slew rate remained a flat line from -50C to 25C. Therefore cooling this opamp in particular has no effect on the sound that a human ear can pick up, (I doubt you can with any commericial available scopes either)

 

[star882]

Quote:

Originally Posted by JahJahBinks Most ICs will stop working once the temperature reaches -40C, military grade can do -55C. So LN2 is out.

Nonsense. It's out of official specs, but it will probably still work (except in the case of the "cold bug"). Some cascades can reach temperatures of -100 degrees or even lower.
Quote:

I would think you would have to worry about moisture condensation on some of the metal parts... if you chill / cool them too much.

Try telling that to the engineers in my area who built a quad Britney "server" with phase change technology. The controller circuits have been programmed to keep the Britney CPUs at about -20C. They simply insulated the cold parts so well that even moisture is kept out. The system has been running for over two months and it still works.

And yes, I know that the difference it makes would probably be too low to notice or maybe even measure. But then again, don't a lot of people here spend a lot on high quality wiring in the belief that it will increase quality? Same here.

 

[JahJahBinks]

Yes I agree with you on the cold part, the IC will work but be out of specs.

However there's more I want to say on this idea. You mentioned about cooling CPU (yes, I am a regular at xtremesystem too), first of all, CPU is a digital IC which is immune to low level noise. However in an audio amplifier, almost everything is analog. In other words, you can keep the opamps cool but the other part of the circuits will be either generating noise (like Johnson noise from resistors) or picking up noise. I don't know if you can insulate and cool an entire amplifier like Headroom Blockhead (like putting the amplifier in a freeze then suck the air out) but otherwise noise generated from opamp isn't the weakest link of the whole system, let alone some amplifiers use discrete transistor design. that's why you have people swapping existing decoupling caps with Blackgate ones, and use silver solder/wiring, higher quality resistors, etc., some have worked for some people, but some will remain for debate. But your idea is a good one.

 

[ATAT]

If we want to be fanatical, johnson noise would be reduced at lower temps .. so cooling packs on resistor arrays would work. Then the same could be done to ICs..
See here for amplifier noise ..http://www.edn.com/article-partner/CA313918.html
but then again, this is just fanatical. Johnson noise is nearly impossible to hear.

 

[blip]

I don't know about phase change cooling... But I've thought for quite a while that water cooling could reduce the price of the average 25-50 watt class A amp. As for headamps... I don't know of too many (other than the dynahi) that vent enough heat to make it worth while.

Anyway, phase change cooling on an amp would be pretty amazing...

 

[[AK]Zip]

My previous computer ran on a custom phase change system I built, but I probably would never do it to any amps that I build. Too much money goes into building the phase change system and i'm not even going to go into the electrical bill.

-Alex-

 

[taylor]

A CPU has millions of transistors... even a complex amp only has about a dozen. I think phase changing is overkill, but watercooling would work...

BTW, I think I know you from bit-tech. It's always funny knowing people from multiple forums. I think there are a few other bit-techers here too (or a few head-fiers at bit-tech, depending on how you look at it)