|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
68.6.139.182
In Reply to: RE: TO-220 package wattage rating ? posted by Ugly on April 25, 2016 at 20:09:58
regardless of the amount of HEAT, Silicon devices can only withstand so high a temperature before malfunction. For even higher powers, you might be advised to go to an ACTIVE cooling solution.
Fans are good and don't forget what the Computer Overclockers are up to with various cooling solutions.
Some active solutions are clearly too noisy for stereo use.
Too much is never enough
Follow Ups:
"regardless of the amount of HEAT, Silicon devices can only withstand so high a temperature before malfunction."
Oh. You mean like I said when I said:
"The key parameter here is die temperature."
and
"As long as you keep die temperature in specified range you are all good. If you take the die temperature above what it is rated by the manufacturer to withstand then there are no guarantees about it's reliability and it's probably going to fail."
???
I guess it IS worth repeating!
Yes, sorry to simply 'repeat'.
Die cooling is somewhat of an art.
These days, the Drain can be the TOP of the die and is secured with a copper bus which is very heat and current conductive.
I was in semiconductor manufacture for 3 decades (more?) and saw a lot of stuff come and go. the main point with computer CPU is that Geomitries shrink and heat density goes way UP.
Thus the water cooler over clockers at the front of the pack.
Too much is never enough
Indeed the newer dual sided pqfn packages are really neat, both from a thermal perspective but also electrcally. The lead inductance on those is quite low and so also very exciting.
Though in this case the op is using to220s. I've never seen to220s with the dual sided cooling .
However it gets accomplshed, more cooling is the way to lower temps.
In many cases this could be something as simple as adding a better thermal interface material between the device and heatsink.in other cases the only way to get there is through more drastic measures such as adding more heat spreading, thermal mass or making what is already there work more efficiently.
Is it OK at this point to introduce 'safe operating area' and the idea of 'derate'?
The ESP article is as good a place to start as any.
Too much is never enough
It may indeed help illustrate the complexity of the situation at hand.
It's unfortunate no info about application, assumptions or goals were ever given by OP. It would be much easier to be helpful if there were more pointed details available.
I'm afraid we may only be sending deeper hole digging support down into the rabbit hole.
It's all good. Hopefully, though I'm a complete douche to put up with sometimes, my limited experience helps turn on an occasional light bulb. I know I have learned a great deal and very thankful for the cranky SOB's I've come across on this forum.
You make an interesting point.
When asked a question about device ratings and usage and such, RARELY is enough information given.
And the OP can or COULD read some data sheets so like was compared TO like.
I'm fairly certain that buried in all the documentation is some reasonable answer.
Too much is never enough
" To much is never enough" I like it. What I am thinking about doing is using a device way higher rated than the one used now because I think it has a good chance of sounding better, right or wrong. The question is still why would a 6oov 10a device be rated higher wattage than an very similar device rated at 650v and 18a. Amps cause heat, don't they? There should be thicker dies, wire or more surface area in the 18A device, which should allow it to take more heat. Something is missing. It seems to have something to do with the insulated package. When you look at the non insulated package wattage numbers things make a lot more sense. There is a greater correlation between current rating and wattage rating with the same or very similar voltage rating. Most of what I care about is that the higher current device has bigger stuff in it, lead wires...
I don't know ALL the measurables of a semiconductor. I never worked the probe area.
However, one big measurable we were concerned with was called RdOn. Resistance of the device in the ON state. Lower is better. Higher produces more heat.
;You are right about one thing. Surface area is a big hitter. I'm looking at a 6" silicon wafer which has about 138 opportunities for 'good' die @100% yield. They are VERY large die, in this case called IGBT. Isolated Gate Bipolar Transistor. Of course, you don't get near that many, and you may automatically EXCLUDE the edge die from even being tested.
Other devices will have Thousands of die for the same wafer size.
You'll have to do some research on what is called 'packaging'. I know the stuff I worked on had the DRAIN as the bottom of the die which was bonded to the heat sink portion of the finished product. The conductive layer was a sandwich of metals applied to exacting specifications for purrity and thickness. The 'Gate' was of course a tiny conductor, and the SOURCE was distributed over the wafer surface in a proprietary manner.
I link a drawing from Digikey, of all places, showing SOURCE and DRAIN. GATE is another contact, but doesn't carry much current.
Modern device heat management ALSO addresses the TOP of the die, not just the bottom.
Too much is never enough
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: