high voltage charge pump

Piglet wro

If I wanted 500V->2V step-down at 2uA and reasonable efficiency (say,

50+%), how would you approach that requirements? The Oxford bell? :-)

I am thinking about a ~5uW supercap replacement for backup purposes, which can't dry out -- a 10uF polypropylene cap at 1kV is equivalent to

10F@1V in 1/2*C*V^2 terms. According to the specs, supercaps are expected to last some mediocre thousands of hours, e.g. 4kHr is just 166 days at the max ratings.

Why there are no glass-embedded supercaps, or 'lytics in general, looking like vacuum tubes? What did they use in Voyager?

Best regards, Piotr

Or UCC28701, I use it in a HV PSU together with a 1.7kV SiC MOSFET.

Best regards, Piotr

But they are super-hard to design properly, a lot of prototyping is required. An UCC7870x controller will work reliably in the first iteration and requires no optocoupler. A very robust little thingy.

Best regards, Piotr

Huh? If it's not to late to ask, (being next year and all*) What was the 48 V 10 uA source? George H.

*Happy New Year :^)

I think 500V to 2V at 2uA is a very tough challenge, John Larkin said to use a regular inductive converter in burst mode, but I am unconvinced.

Voyager used off-the-shelf capacitors e.g. ceramic, mica, film-foil at smaller values and wet tantalum at higher values. Wet tantalum used within ratings has no known wearout mechanism but the prices will amaze.

piglet

Happy New Year George.

Power source was plain old telephone system, in the on-hook idle state telcos had limits on loop current that could be drawn in some cases 50 or 100uA and multiple units of my item could have been present hence the low allowance.

piglet

By "off-the-shelf" I did not mean commercial grade. Clearly aerospace/military spec parts were subject to extra screening and qualification processes. Voyager used capacitors of similar technology that we can still access.

piglet

2uA I_out, I should have added. Nanoamps at the HV side. Indeed, very though, but there it is. The Oxford bell electromechanical converter might not turn out to be that silly an idea, methinks.

The price of a service visit can be even more amazing, so sometimes it pays off overall. But why is there no wearout? A rubber seal is just a rubber seal, i.e. leaky.

Best regards, Piotr

Well, the 500V side will average 4 nA or so, so a switchmode solution has to have leakage well under that. Maybe you want an acoustic delay line, triggering (and excited by) a UV flash tube, feeding a long-delay phosphor and snazzy photocells. The phosphor is your filter, the UV tube can be made self-starting with small amounts of radiioactives, and the acoustic delay will look like a slinky. To quench the arc, though, you might want to have some more recognizable electronic components clustered around the slinky coupling transformer.

That presumes you can get by with PWM and the transformation ratio of the photonic conversions, but no feedback. Its tricky to get feedback when the permissible leakages are down under the level of mass-production testing.

Perhaps an electroluminescent panel and silicon cell is a more sane alternative, but a lithium cell can give your 2 uA for a decade, and that's how I'd usually go.

I think that means you have not seen a wet tantalum? They are hermetic, glass-metal sealed. Plenty to be seen at Mouser and Digikey - admire the prices.

piglet

Correction: cheaper ones can be elastomer sealed, the really high spec units are hermetic. I don't know the details but there must be tricks to keep hot pressurized sulfuric acid inside!

piglet

You are perfectly right, I have no experience with them. This one looks decent:

But the specification of this kilobuck part says: "Life Test: capacitors are capable of withstanding a 2000 h life test at a temperature of +85

Pardon me, how many hours? Three months of uninterrupted operation? A hamster lives for 2 years despite not being exactly hermetic...

This unreliable electrochemical technology should have long been forgotten.

Best regards, Piotr

Even the primary side capacitor would need to have leakage well under that, which makes this entire approach a brain fart in practice. But it can have some value as a food for speculation. Certainly doable, this unit has been working since 1840:

A ball on a thread turns out to be sufficient to get a 2Hz fixed duty cycle switched capacitor converter. :-)

If a decade is long enough, then no question about it. But how about a century? Betavoltaics?

On a similar note: why there are no small NiFe batteries with catalytic hydrogen burners on the market, say 1Ah? They would be eternal, have a look at that:

A guy revived a number of 85yo Edison cells and brought them close to their peak capacity.

Best regards, Piotr