Somewhat OT: Long term design

Do that on DimBulb and it'll come back brown.

Well, I meant the real serious wars where vast stretches of land go up in flames or wild hordes destroy everything in their path. They always had the little spats between individual fiefdoms. For example, Germany wasn't really a country until "recently".

That's a problem, in as much as the TV show had the computer (which wasn't actually mentioned, or was referred to as a "core drive"), housing a human personality. I wouldn't know where to start in terms of defining its capabilities in a way that would make engineering sense. Indeed, it may not be possible, even in principle.

Sylvia.

If it's hidden away, and only exposed to the sun after the time is up, then we need a mechanism to do the exposing, and that mechanism will require a power source.

Sylvia.

This is where I wonder about difusion. Will those doping atoms stay put over those sorts of time scales, or would we end up with a piece of silicon pretty much equally doped throughout with both doping materials?

Sylvia.

Sine i am a baer, may i growl and say "Gurr-avity"?

CK722s were made with relatively impure chemicals and processed WRT what is available today, and i suspect the base junctions were not all that swift - but there are some CK722s that have reasonable characteristics today. Tough to extend that "decay" to silicon solar cells, but may give a crude guesstimate as to worst case expectations..

Back to needing a mechanical mechanism that will work.

Sylvia.

will

they

(how?)

Indeed, consider the traps in ancient tombs.

The original 'lint' traps? ;-)

If it needs to be reliable solder is out of the question. Just look at automotive electronics. The really reliable stuff is (sort of) spot welded and covered with some sort of goo. Other ways are wire-wrap, press-fit and crimping.

Tantalums are very prone to failure. I avoid them if I can. Even electrolytics are better because they don't cause a short. Nowadays I use the MLCC capacitors where I can.

how

the

will

difusion

they

Well an 8 inch 58 military cannon accelerates an approx 100 lb=20 projectile to about 3000 ft/s in some 22 feet. The acceleration=20 curve looks kind of like a truncated "F" distribution.

On a sunny day (Sat, 16 Jan 2010 13:30:54 GMT) it happened snipped-for-privacy@puntnl.niks (Nico Coesel) wrote in :

Not if you use them right

I have some CK722's and also CK760/61 dating from ~1956.

When I get a moment I'll test them. ...Jim Thompson

Tantalums are fine as long as their peak current (ie, dV/dT) is limited. Of course, that makes them useless for most places you'd like to use them.

I sometimes use them on linear and switching regulator outputs. I make sure that dV/dT is controlled and derate them 2:1 or better 3:1 on voltage. They do have nice ESR numbers, ideal for a lot of regulators that don't tolerate all-ceramic loads.

John

Wot mostly consists of leaving them safely on the reel.

Best regards, Spehro Pefhany

On a sunny day (Sat, 16 Jan 2010 12:02:26 -0500) it happened Spehro Pefhany wrote in :

Well, then they still may fail ;-) Just that you won't know about it. But on the more serious side, I have never had one go kaput, I like them because of low ESR and small size. They do not dry out, some have been on for 20 years... I have seen exploded ones cause damage in equipment, most likely because they were put in in reverse... I put one in (actually a whole series) in reverse myself one day, because the + was marked with a big -- That came out when the first one was tested, and the resistor in series with it burned a hole in the PCB. That was supposed to be a fusible resistor, I have now learned that metal film resistors do not fuse very well.

;-)

s/know/care/

I certainly have. We have a huge fallout on 220uF 35V tantalums; right off the reel. They're better than aluminums, though. The aluminums didn't make it through the RoHS process. :-(

Some light bulbs have made it 50 years, others have cause somewhat more problems.

were put in in reverse...

was marked with a big --

burned a hole in the PCB.

resistors do not fuse very well.

That too. Even ones inserted the right direction tend to short, taking out the series resistor. We're seeing a lot of charred resistors next to those 220uF caps.

Many years ago, it we found that no matter what we did, 1% got inserted backwards. ...even if the manufacturer had to put them in the tube backwards to meet the target. ;-) The solution was a fuse in the cap, non-polarized cases (+-+ or +--+ pinouts), or big/little pin, though the latter still could be force fit.