Very durable electrolytics or similar?

If the ESR of the electrolytic capacitor is high or the ripple frequency is high, you might parallel a film cap to eat some of the ripple current to reduce ESR induced heating

Regards

Klaus

The ripple frequency will be tens of kHz and I am planning on providing a few uF in ceramics in parallel. The main problem is externally generated heat. It'll be like inside a car sitting in a parking lot on a hot sunny day.

--
Regards, Joerg 

http://www.analogconsultants.com/

As electros age and go out of spec (which is counted as a failure for MTBF) they loose capacitance - you could combat that by using larger than necessary capacitors.

"David Eather = pig ignorant troll "

** Nonsense.

When electros lose enough electrolyte, their ESR goes high and only much later the capacitance falls.

That is why ESR meters are such a great tool for service work and capacitance meters are almost useless.

.... Phil

"Joerk"

** This makes no sense - cos as usual Joerk make no sense.

Why does the electronics have to sit in DIRECT sun ???

Air temps are nothing like 70-80C anywhere on earth and it is not sunny all day, every day nor is it windless all day every day.

FFS - organise some shade for the electronics module and regular electros will do fine.

I constantly see filter electros that have worked in valve amplifiers for over 40 years with local ambients of 80C.

.... Phil

"David Eather = pig ignorant troll "

** That IS the failure mode of most electros. ** Then he's an utter idiot just like you. ** Large electros have longer lives than small ones - whatever their value in uF.

It's the ratio of surface area to volume that helps.

.... Phil

Maybe the CY series @105c would be better, its rated for 10000 to

12000 hrs.

Cheers

Where can you find ceramics that are actually single digit uF *at* that kind of voltage?

Also, how much ripple is expected of the electrolytic -- give or take what ESR it might have, of course?

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 

Sounds like something suitable for an automobile or a pizza oven.

Since cost is not a consideration, I suggest tantalum caps:

100uf @ 200V is going to be expen$ive if you can find a vendor. I couldn't, which suggests you'll need several caps in parallel.

You might also look into using a PCB stuffed full of tantalum polymer caps to achieve the necessary capacitance and voltage.

--
Jeff Liebermann     jeffl@cruzio.com 
150 Felker St #D    http://www.LearnByDestroying.com 

Nichicon BX series? 150c

Cheers

"Phil Allison"

** Really - why ?

The ambient air temp in the shade is not 80C.

.... Phil

No kidding. Mouser's picker leads to Vishay/Roederstein MKP1848C71050JY5 , which meets a lot of your specs, but is 2.3" x 2" x

1.4", and costs $28. (It only meets the life spec at less than its 105 C max temperature; they derate life sharply with heat.)

Maybe tantalums, but solid tantalums don't seem to go that high in voltage. Wet tantalums get closer, but the ESR seems high, and they are only about 20 dB over budget.

Can you cool it down in any way? $5 buys you a 38 CFM fan that claims to last 70,000 hours.

The film caps that come up often have "DC link" on the sheet, like for an inverter or VFD. Maybe look inside an industrial VFD to see what it uses? (On the other hand, that application has the luxury of more room and maybe a blower.)

Matt Roberds

Did you read the bit where he said he wanted more than the usual

105C/12000h?

I have had excellent experience with the Sprague TE series in the past (from 50 years ago to ~30 years ago). Perhaps the quality is still there in present production.

Check on the tube electrolytics; back in the daze of 4,5, 6, and

7-pin tubes and WET/LIQUID electolytics, those lasted in the region of 20 years, and (most of the time) could be "rescued" by re-filling of the electrolyte.

Nowhere, you have to parallel many. That is one of the problems I am facing several times a year these days, poor datasheets. Very few companies offer truthful capacitance versus applied voltage data and sometimes you have to cajole it out of them. When I can't I walk.

About 500mA.

--
Regards, Joerg 

http://www.analogconsultants.com/

Thanks, a good hint. I had looked at them but the 1000h an 2000h ratings had turned me off. However, assuming the lifetime doubling for every 10C lower this could indeed work. Their seated height ain't all that bad either, about an inch. These could work.

The ripple current is lowish but I can live with it.

--
Regards, Joerg 

http://www.analogconsultants.com/

Well, the first ones Martin suggested contain some 150C/2000h, they should be able to get there. The CY series probably won't.

--
Regards, Joerg 

http://www.analogconsultants.com/

Tantalum is blacklisted here :-)

--
Regards, Joerg 

http://www.analogconsultants.com/