Path: ...!eternal-september.org!feeder3.eternal-september.org!news.eternal-september.org!.POSTED!not-for-mail From: Cursitor Doom Newsgroups: sci.electronics.design Subject: Re: Operating temperature derating Date: Wed, 5 Jun 2024 16:57:51 -0000 (UTC) Organization: A noiseless patient Spider Lines: 71 Message-ID: References: <17fs5jhcf319ce35l8tt92iokkdbvn2bbc@4ax.com> MIME-Version: 1.0 Content-Type: text/plain; charset=UTF-8 Content-Transfer-Encoding: 8bit Injection-Date: Wed, 05 Jun 2024 18:57:51 +0200 (CEST) Injection-Info: dont-email.me; posting-host="fe7321ab6985b475819f3cff75d69cc7"; logging-data="1112915"; mail-complaints-to="abuse@eternal-september.org"; posting-account="U2FsdGVkX1+spOD9Tq8jGUa4CQNKV9ECqgL2yFcWp+4=" User-Agent: Pan/0.149 (Bellevue; 4c157ba) Cancel-Lock: sha1:sY+gkfulvK+zpq8g9Ag3KZtGo+c= Bytes: 3822 On Mon, 03 Jun 2024 15:05:58 -0700, john larkin wrote: > On Mon, 3 Jun 2024 21:41:59 -0000 (UTC), Cursitor Doom > wrote: > >>On Mon, 03 Jun 2024 09:03:10 -0700, john larkin wrote: >> >>> On Mon, 3 Jun 2024 07:52:58 -0700, Don Y >>> wrote: >>> >>>>Presumably, one should feel comfortable using a device at the >>>>published operating temperature extremes "forever". >>>> >>>>But, what sort of derating likely went into that specification in the >>>>first place? Sad another way, how much *beyond* those limits might >>>>want suspect you could operate the device? >>> >>> I don't think that there are any standards that manufacturers use to >>> make datasheet abs max specs. >>> >>> I like to test parts to destruction and estimate what they can >>> actually do. For example, max voltage ratings can sometimes be safely >>> exceeded, or a surface-mount resistor can dump a lot of power if it's >>> well heat sunk with copper pours. >> >>Some fellow on Youtube a while back managed to get a kilowatt out of a >>TO-220 packaged device rated at 100W by blasting it with plumber's >>Arctic Spray. >> >>> Thermal imaging can quantify hot-spot temperatures. >>> >>> Pushing parts past specs is only a good idea if there is a serious >>> performance upside. >>> >>> RF parts are especially likely to be under-rated. Like a 2 volt rated >>> schottly detector diode that starts to leak at 7. Or RF amp parts that >>> actually run at 2x rated max voltage in normal use. >> >>Probably because it's common in RF to end up with the part being >>subjected to voltages double its normal supply rail figure. > > When I use RF parts in time domain/pulse applications, I have to make up > my own rules. > > It's rare to get even basic DC curves on RF parts. The data sheets say, > literally, adjust the bias until it works. Or AC couple in and out and > don't ask questions about "volts". Which opens up the possibility of > novel biasing schemes. Yeah, RF biasing - well, everything to do with RF in fact is a blast, isn't it? No wonder the guys at the cutting edge of it have always coined the BIG BUCKS. >>> GaN fets are interesting. As are mosfets that avalanche at some >>> unspecified voltage. >>> >>> A bit of air flow can really change things. I assume that most specs >>> are based on still air. >>> >>> I've seen a few parts that were unreliable within their abs max specs >>> and needed to be derated from there. >>> >>> Blow up some parts. It's fun. >> >>Best fun you can have with your clothes on as they say. > > Clothes on and safety goggles. Indeed. It's amazing how many decibels you can get out of even a 10uF electrolytic.