I would never place tantalums in power apps. Seen too much grief (with other guys' designs).
Can't do that. In some environments any moving parts and especially fans will cause reliability to drop way down.
Many use electrolytics and that's pretty much what goes bad when they croak. From what I've heard that happens a lot.
-- Regards, Joerg http://www.analogconsultants.com/
It's Vishay now. I only found the (very expensive) 600D series but it's
125C/2000h.-- Regards, Joerg http://www.analogconsultants.com/
How much do you want to pay? I use 50V 1210s. 22uF (33uF, too, AFAIK) is available in X7Rs at 50V. Even derating 70% for 50% of rated voltage, that's still "single digit uF".
Don't understand the question. Ripple currents into the amps aren't uncommon. I have one application where aluminums won't cut it so we're going to multiple polymers.
We are talking 200VDC here.
Amps? My circuit will be quite merciless when it comes to bombarding the caps with ripple current.
-- Regards, Joerg http://www.analogconsultants.com/
Why not visit those web sites about tesla coils and induction heating? especially induction heating caps take a brutal beating and regularly are run hot. Just a thought that they may be onto what you're looking for.
Joerg Look at your thermal situation: the T is >> ambient, so see if you can dissipate it. Please describe that in more detail.
As far as the caps, here is one solution: Use two banks and provide a switch for the user to flip if failure occurs - provide an led for that eventuality. Hey, it's not high-tech, but CYA.
j
You'd be surprised. On commercial gear, also VFD and such, the warranty is often limited for the electrolytics. They are considered wear and tear parts. Much of that is three-phase anyhow and then there isn't much need for capacitors.
-- Regards, Joerg http://www.analogconsultants.com/
It's going to be a fairly closed cavity and there can't be any forced venting. Similar to a car, just smaller. In summer this sits in the blistering sun all day long. In some locations the ambient can be 45C in the shade for weeks and there is no shade.
:-)
It likely won't help much if the caps simply dry out, then the spare bank is going to be almost as sick.
-- Regards, Joerg http://www.analogconsultants.com/
How much difference does it make to use excess capacitors to reduce ripple, self heating and improve ESR tolerance?
I wondered about jigging a circuit so it doesnt rely so much on storage caps, but its gotta cost more.
NT
Last I checked, 0.01uF 630V X7Rs were 1812 case or so. But then, I was using them at rated voltage. Which is as far away from the present case as that example, so double-nevermind. :)
Ripple current in the amps is just fine, if it runs at 40C. ;-) "Sunshine in Arizona" temperatures, plus even relatively little ripple (heck, none at all?) is how capacitors are normally rated. Example, a stock 100uF
200V aluminum electrolytic might be, oh Idunno, 0.5 to 2A (depending on grade) ripple, at 105C, for 3khr life. Even the 10khr ones (which are usually rated for a bit less ripple, I think.. so they're a little wussier, but live much longer?).Stacking polymers might not be too horrible (might even do away with the ceramics entirely?), but that cost figure rides up quickly too. Actually, if the ceramics could go away, does that spare a few more bucks for the rest?
Those types sure are ridiculously robust, but they also require heatsinking (or water cooling!) in that duty, and the energy density is total junk. But hey, when you need to resonate a kiloampere, you can spare a few soda cans' worth of space.
Alternative thoughts: can you make it a Cuk so it has less supply/load ripple, or something like that? ;-)
Tim
-- Seven Transistor Labs Electrical Engineering Consultation Website: http://seventransistorlabs.com
1206s, now. .1uFs are available in 1812. 85C. You have to watch the ripple frequency with aluminums, too. We're running 500-800kHz. According the usual suppliers, hybrid polymers will take it. Aluminums won't.
I was looking at the OP's need (28V), not the hundreds Joerg was discussing.
Aluminum foil and glass plates. ;-)
That's what I'll do anyhow but eventually there comes a space and cost limit.
This time I need the storage. I could relegate that to a lower voltage rail and convert back and forth but things aren't much better down there.
-- Regards, Joerg http://www.analogconsultants.com/
Hi Joerg
It seems you already found something suitable from Nichicon, so here's a not-so-serious brute force "solution".
While it's hardly low-cost, one could in theory "synthesize" something with a film cap, lots of polymer electrolytics and some active circuit:
With dry polymer electrolytics you get temperature endurance, but only at voltages up to ca. 35V, higher rated ones seem not to exist. On the other hand, they are available in thousands of uF and a small size, so they can be paralleled in quantities. Add 2 or 3 extra for fault tolerance if needed.
If you run the converter unregulated with one fixed duty cycle, it will run in both quadrants and essentially transform a low-voltage high-capacitance cap into an equivalent high-voltage low-capacitance one with the same stored energy. Unfortunately the inductor is going to wreck the ESL (and ESR), but that can be "fixed" with one or two "DC link" film caps with a dozen uF and high ripple current ratings.
Alternatively, put the polymer caps in series and trade the drawbacks of the converter against the drawbacks of voltage balancing :)
Regards Dimitrij
Thinking "out of the box": Is there any way you could re-design it with an inductor to do the energy storage part of the job and then have an oil-filled paper capacitor of a few microfarads to by-pass any H.F. signals?
(Is there any restriction on size and weight?)
-- ~ Adrian Tuddenham ~ (Remove the ".invalid"s and add ".co.uk" to reply) www.poppyrecords.co.uk
Forget the constant duty cycle. Adjust it so that the usable voltage range of the high side converts to a full voltage range on the low side. You'll have virtually a very large capacitor for the voltages you're interested in.
If, for example, your high side working voltage is 170-200 V and the low side is rated for 65V, adjust the duty cycle so that L=(H-170)*2.16. Synchronous buck regulators can do this if you modify their feedback. The trick is that they need to have a continuous mode and you must not allow the feedback line to stray into UVLO or OVLO regions.
snip
This hasn't been mentioned, but you can cool it with a TE. or a fan? You haven't explained where this temperature comes from... j
I presume Joerg has ruled out all the other electric energy storage options, eg inductor, supercap, battery et al. NiCd, NiMH or even primary cells should manage several 10s of khours.
ps Oiled paper is one of the least reliable cap types.
NT
I was thinking of the hermetically-sealed metal cans with ceramic terminals, the better brands seemed to last for many decades as power-factor correctors across the mains or as HT smoothers in big valve amplifiers and transmitters. If modern materials last even longer, that would be better still.
-- ~ Adrian Tuddenham ~ (Remove the ".invalid"s and add ".co.uk" to reply) www.poppyrecords.co.uk
ove
r.
ith
ry
All cap types routinely last decades, even lytics. From repairing pre-war r adios I've found lytics are often still good, but in a not especially big s ample I dont think I've found a single paper cap still working. They're kno wn for their reliability issue. That includes both rolled paper jacketed an d sealed metal can types.
NT
The OP and Joerg are one in the same. :)
Tim
-- Seven Transistor Labs Electrical Engineering Consultation Website: http://seventransistorlabs.com
ElectronDepot website is not affiliated with any of the manufacturers or service providers discussed here. All logos and trade names are the property of their respective owners.