An important additional aspect - you didn't ask this but it's vital. Electrolytic Capacitor lifetime drops by approx a factor of 2 times (pessimistic rule of thumb) for every 10 degree Celsius temperature rise. If you are consistently running at 80C you MUST use 105 degree C capacitors or your lifetime will be minimal. Specs vary but typically caps are specified for eg 1000 hours at rated temperature on the assumption that you will run them MUCH lower than this. eg an 85C part run at 45C is 40 degrees lower so lives 2^4 = 16 times longer = 16000 hours. Actually it would be somewhat better than this. Pseudo Randomly opening my Philips cap book (well thumbed page) gives specs for their 037 series which are common. 85C 2000 hours 40C 70000 hours . Another CRITICAL fact is that if the capacitor has no voltage applied or low voltage compared to its rating it will die much quicker at elevated temperature. The above 037 caps are rated for 500 hours at 85C when not powered !!!!. this may not be intuitive! -----Original Message----- From: Lawrence Lile To: PICLIST@MITVMA.MIT.EDU Date: Wednesday, September 16, 1998 5:20 AM Subject: Electrolytic Capacitors and Heat [OT] >Ok, so this time it doesn't have a PIC in it. But my prototype IS on >the same table as my ICEPIC, so it's remotely related..... > > >I'm having trouble with an electrolytic capacitor (100uF, 85 C, >50WVDC) that drops in value as it gets hot. It is in an RC >timing circuit, so a change in capacitance is a problem. The circuit >has to work in an environment that starts at room temperature, and >climbs to 80C. I don't know how mush the capacitor drops, but the >slope of the voltage on the capacitor gets steeper the hotter it >gets. > >The question is: Do all electrolytic capacitors change capacitance >as they get hot? If so, how much and what direction? Is there a >type that is more stable at higher temperatures? >-- Lawrence Lile > > "The ideal design has zero parts" - > (attributed to Harold Hallikainen) > >Download AutoCad blocks for electrical drafting at: >http://home1.gte.net/llile/index.htm >