Very durable electrolytics or similar?

Electrolytic next to a rectifier tube is "low stress"?

With my grandparents I know for sure that their tube radio saw tons of accumulated hours. The radio was turned on when they got up in the morning and off when they went to bed. For decades.

--
Regards, Joerg 

http://www.analogconsultants.com/

Unfortunately that's just a drop in the bucket, even 1uF/250V is.

Why not? Or did you want to push them to the limits?

Me is the OP :-)

There is also a 28V rail that needs electolytics but that's easier.

[...]

--
Regards, Joerg 

http://www.analogconsultants.com/

Use whatever type of capacitor has the best life at high temperature in the biggest value that is convenient, then take care of the rest of the energy storage requirements with an inductor. i.e. design backwards from the available type of capacitor and trade off capacitance against inductance. In general, the reliability of inductors can be made much better than capacitors at high temperatures and you have more control over the design and materials of an inductor.

If this is a filter, could the specification be met with two or three smaller sections, rather than one big one?

--
~ Adrian Tuddenham ~ 
(Remove the ".invalid"s and add ".co.uk" to reply) 
www.poppyrecords.co.uk

It's not a filter, it is a PWM stage driving a highly capacitive load that sometimes needs to be muscled around. Since there also is a low ripple requirement that result in a large output cap on the power supply. One option is to provide the reservoir at a lower voltage and run a converter at high frequency but that comes with increased losses. Which we'd have a hard time to stomach. The clean solution would be to have a well-performing elecrolytic up there on the HV rail and the Nichicon caps Martin suggested could be the ticket.

--
Regards, Joerg 

http://www.analogconsultants.com/

The peak current could come out of a boost (not a buck) if strong enough, that's not a problem. But filter ringing must be muffled and that would then have to be "bucked" down to where the bulk capacitance is, in order to use the newer polymer caps that are only available at lower voltage. However, if the Nichicon caps perform that is clearly the straightforward method because no buck is then required.

--
Regards, Joerg 

http://www.analogconsultants.com/

Wouldn't a buck limiter for the PWM output work better?

The buck could engage if the peak current hits the wall during a on cycle of the PWM.

That would at-least be more efficient over a linear method of current limiting and would most likely make the caps happier, too!

If you make the buck switch much faster than the PWM base line, that should offer a nice roll over when needed.

Jamie

Gee, can you at least get the mechanical guys to paint it something other than matte black? ;-)

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com

How do you know it didn't fail by way of greatly reduced capacitance or excessive leakage, back in those days they could have used a slice of bread as a filter cap...

I have a box of oldies laying around somewhere. But I doubt the ESR would be good enough, you see. ;-)

...Which is another interesting point...if manufacturers needed to make a long life, poor performance component...could they? Presumably so. Don't really know how it would be different. And as high ESR and capacity goes, I haven't read any articles about supercaps' lifetime.. how do they compare? Not that that matters to the present thread.

Tim

--
Seven Transistor Labs 
Electrical Engineering Consultation 
Website: http://seventransistorlabs.com

Yes, and I have seen capacitor heatsinks as well recently.

Of course, depending on the thermal resistance between that heat sink or the outside of the cap and the inside of the cap that is getting hottest, it may or may not help a lot. Maybe a lot of smaller caps rather than one large cap would be better with such a cooling technique ?

I'd like to get an electrolytic cap with an internal thermistor to give an idea of what is actually going on inside.

boB

I'm pretty sure you could just make it with these Extended life jobs:

...no one ever knows with your projects since you never give any circumstantial information.

Then the lifetime of nearly anything can be extended indefinitely with suitably arranged redundancy. Just because you can't think of anything doesn't mean it's impossible.

Just making the point that these things are getting smaller. Perhaps you'll just have to wait a hundred years, or so. ;-)

To? Past, is the problem.

Yes, that's what you said. That's what stuck in my mind from an earlier read-though.

** No way.

** Fahrenheit - not Celsius.

Even a perfect black body, in direct sun with no cooling air, only reaches

69C.

Cos at that temp it radiates the same energy it is absorbing.

... Phil

If you study reaction rate theory, it's just saying that the decay process has a particular activation energy, around 50 kjoule/mole.

As far as I know it's just as good for electrolytics as it is for semiconductors. Quite why that activation energy is magical escapes me.

--
Bill Sloman, Sydney

That seems counterintuitive: I recall hearing and seeing steam come out of a solar collector (flat plate with no concentrating optics) when the fluid was first poured into it, and that was mid-morning. This was a long time ago so probably before they came up with the coatings that are simultaneously high-emissivity in the wavelengths of sunlight and low-emissivity in the long infrared.

Using the formula from here:

I get that a black body at 91.27 degrees C would radiate 1000 W/m2, which is a rough approximation for the power density of sunlight, (though it often exceeds this intensity in Australia).

I think that would be the temperature attained on earth by a black object in a vacuum chamber, which could only see the sun and received no other thermal radiation, e.g. if the object was surrounded by a black box cooled colder than liquid helium, but with a window so it could see the sunlight. In practice, most objects can also see the earth, which is quite a bit warmer than absolute zero, so they receive thermal radiation from that direction also.

Chris

** The sun only shines on one side of a ( black) surface.

The same surface radiates from both.

Might account for the difference.

In any case, the original claim of 115C is rubbish.

.... Phil

Silicon rectifiers are more stressful than tube rectifier. The peak rectifier tube current was quite low and thus you have to increase the conduction angle by low capacitance capacitors with pi type (CLC) filtering. When the set is turned on, the capacitors are charged _slowly_ when the rectifier filament is gradually heating up.

With silicon rectifier and only one big filter capacitor, the conduction angle can be very short, so up to 10 times average current may flow during voltage peaks. The initial turn on current peak can be huge, limited only by the network and transformer impedance.

[...]

I don't like to push these kinds of part even close to the limits WRT ripple and such because that negatively affects the MTBF of the whole system.

[...]

--
Regards, Joerg 

http://www.analogconsultants.com/