Heat and Amps

[The Monkey]

At what point should you start to worry about how hot your amps (and other components) get? For example, my HR Balanced desktop runs hot, as expected, and so does the Desktop Power Supply. Further the DAC1 also runs pretty warm, which is also expected. Regardless of the fact that it is expected, it still concerns me. Should it? At what point do you consider the heat as a sign of a problem, or the imminent cause of one. Er, aside from fire that is.

[luvdunhill]

my rule of thumb is nothing (except perhaps tube envelopes) should ever be 15-20C above ambient. If you live in the frozen north, maybe higher, but I live in Texas

[n_maher]

Vincent, the amount of heat that would cause a problem is going to be dictated, in large measure, by the temp rating of the components used inside whatever it is you're worried about. In the case of the HR Balanced Desktop, I can't imagine it being an issue since HR tends to be pretty conservative with what they'll tolerate for operating temps. I don't remember that amp getting all that hot (relatively speaking), I mean it was never "I can't touch it hot" like the Wheatfield. If it really is worrisome for you get a IR thermometer and monitor the temps periodically. I think you'll find that what's warm/hot to the touch isn't all that hot really. For example, caps like the Panny FC's are rated to operate up to 105C, which is way hotter than any of your components are running (at least I hope so!).

[spritzer]

my rule of thumb is nothing (except perhaps tube envelopes) should ever be 15-20C above ambient. If you live in the frozen north, maybe higher, but I live in Texas

Frozen north FTW!!!!

[ojnihs]

my lunchbox runs hot, but then again, tubes run hot

[The Monkey]

Thanks Nate, I will get the IR thermometer.

[Duggeh]

I can cook eggs on the Pass Labs. It probably gets to 20 above ambient. But Scotland is pretty cold too.

[Chekhonte]

My tube/mosfet hybrid amp gets really hot as well. Not hot enough to be painful to the touch, but very close. It freaked me out quite a bit before I knew it was normal.

[recstar24]

There are 3 levels of hot that I have experienced...

(1) Zana Deux hot - the chassis gets pretty warm due to the chassis acting as a heatsink for the 6c33c's. It is not too hot where you can't touch the chassis, it is warm to the point where you can touch the chassis at its hottest point and keep it there for an extended period of time without getting uncomfortable. Works great in the winter especially in Chicago. Most class A solid state headphone amps and tube amps fall under this category, which is perfectly normal and usually well below maximum operational points. Obviously the more biased into class A, the hotter it will get typically.

(2) Pass Labs/First Watt hot - the hotness takes a whole other level. You can touch the chassis but after awhile it gets a bit too much and you have to take off your hand. However still perfectly operational and nothing to worry about with internals and safety.

(3) Prototype Raptor at Chicago Meet hot - this is the highest level of hot that I have experienced. The Raptor was debuted at the Chicago meet at Ray's January 2005 and the chassis was so hot you could not touch it, if you did it you would have to lift your finger up immediately. The production was much better in that regard, and when the power supply went to the switching supply it was fine.

Levels 1 and 2 are fine...level 3....ummmm....someone else can comment.

[aardvark baguette]

I never realized the raptor used a switching supply. Boooooo.

My Ear HP4 gets what I would call warm, but never hot. If my window and door are closed, with the computer running and the fan off, then it probably would get significantly hotter; my computer is a freaking space heater.

In purely scientific terms, I estimate the average heat after 5 hours of use to be equal to that of a plate holding food that has been in a microwave for 30 seconds to one minute.

I've never opened the case, so I am only going by the metal top (which is also vented btw).

[Dusty Chalk]

It depends what. I wouldn't worry about the Wheatfield until it actually burst into flame (or singed my cat's hair).

[luvdunhill]

Vincent, the amount of heat that would cause a problem is going to be dictated, in large measure, by the temp rating of the components used inside whatever it is you're worried about.

yes, the temperature ratings are one important thing, but I worry more about things like output transistors. Actual die temperatures aren't something that you can easily measure, nor simply read off a component. If you're oversubscribing parts, then you're in uncharted territory. Whether the devices last a day, week, month, years is something that is unknown and is probably a function of how sane the designer was.

[n_maher]

yes, the temperature ratings are one important thing, but I worry more about things like output transistors. Actual die temperatures aren't something that you can easily measure, nor simply read off a component. If you're oversubscribing parts, then you're in uncharted territory. Whether the devices last a day, week, month, years is something that is unknown and is probably a function of how sane the designer was.

Yes, it's very true that even reading the temp off a heatsink will yield very little information about the junction temperature. Based on what I've read (pretty little, actually) if you're operating a heatsink at more than half the part's rated junction temperature you're probably in trouble. But taking chassis measurements is at least a place to start and is near the temperature that the other components (like caps) are being exposed to.

[Smeggy]

My old Hafler runs hot as a girl gone wild in Florida! A really, really hot girl.

[justin]

Yes, it's very true that even reading the temp off a heatsink will yield very little information about the junction temperature. Based on what I've read (pretty little, actually) if you're operating a heatsink at more than half the part's rated junction temperature you're probably in trouble. But taking chassis measurements is at least a place to start and is near the temperature that the other components (like caps) are being exposed to.

You can get the junction to case C/W in the datasheet, then add in the C/W for your mounting components. I try to keep all measured heatsink temps at 60C or below, 50C or below if it's exposed

[swt61]

Frozen north FTW!!!!

That's even funnier because Frozen North is the name of a great electronics supply house in my former hometown of Anchorage, Alaska.

[spritzer]

That's even funnier because Frozen North is the name of a great electronics supply house in my former hometown of Anchorage, Alaska.

Hehe that is a good one. It isn't really that cold up here and -15

[luvdunhill]

That's even funnier because Frozen North is the name of a great electronics supply house in my former hometown of Anchorage, Alaska.

(unfortunately) I got that phrase from a cartoon my son watches about a blue octopus named Oswald, whose voice is none other than Fred Savage. He has a friend who lives next door named Henry who is a penguin (whose voice is David Landers aka "Squiggy"). One episode, Henry's cousin visits from "The Frozen North". Cousin Louie is played by Michael McKean (who played Lenny). Pretty funny stuff...