Photovoltaic isolator

It seems like a good idea overall, if your 1% figure is correct. 1% sounds like a good minimum, don't fall over if you do better. Note, however, that anything I say here could be wronger than wrong, so take them all as suggestions for things to test.

I'm pretty sure that the physics of photovoltaics are such that once the photon smacking into a silicon atom has more energy than the material's band gap, you get one and only one electron bumped up to the conduction band for each photon. Assuming that LED energy efficiencies are roughly similar across the board, you may be better with IR or deep red LEDs than with blue ones because each photon will come at a lower energy cost.

You probably already know this, but shop around for your emitters and photovoltaic arrays -- efficiencies vary wildly, and you definitely want the most bang for your buck here.

I think I'd start with a photovoltaic array that can deliver 100mA at 3V in full sun -- expecting your artificial light source to have about

There are high isolation DC-DC converters out there, not cheap, but ultimately superior to a photvoltaic system.

A home-made inductive coupling thing wouldn't be difficult. It could work through a piece of plexiglas for visible isolation, or just be traces on opposite sides of a PC board. That could be really cheap in production.

Or wind a ferrite toroid to make a transformer, and use really well insulated wire.

But as you say, it's a lot easier to buy an appropriate dc/dc converter.

A home-made transformer is completely practical. I made a 0 to 250V supply, floating on top of 15kV, using a simple TL494 driving a 1:1:10 transformer at 50kHz. I used a 0.47-inch dia core and machined a special two-section bobbin, with 1-inch inner dia, giving 0.25-in insulation. The two sections were separated by 0.20 in, overall bobbin length 0.8 in. This performed flawlessly. (The feedback scheme I chose for the supply was a bit dicey, requiring two 30kV 150M resistors, with considerable care for the mounting and insulating.) Reference RIS-623.

For uA of current using a stack of photodiode optocouplers could work, but for mA inductive coupling is likely the way to go.

The Starbucks near me has those "Powermat" chargers where you stick an inductor ring to your phone's charging port and lay it on the inductor ring on the table; I guess they chop the AC up to a high enough freq that air makes a decent enough core, and a loop only slightly larger than a quarter can transfer enough energy to charge up a smartphone in under an hour

It's a few hundred kHz, AFAIK.

And yes, it may work better to deliver 10mA at 3V. But it's boring -- I think Clive should go with the glowing things.

If you do it all on one circuit board (how much isolation is 1/16" FR-4 worth?) then you can vastly increase efficiency by using stick-on ferrite plates, available from DigiKey as EMI reduction thingies.

SparkFun has some itty-bitty solar panels for about a buck fifty each, but their maximum power point looks like about 40uA at 250mV, so it'll take quite a stack to hit 10mA.

Since he's got kV available, why not wrap a gas discharge/neon tube in some strips of PV sheet?

Let's spread the rumor that the field will cause warts or insanity or something.

Den torsdag den 5. januar 2017 kl. 22.42.58 UTC+1 skrev John Larkin:

the promise of wireless power seems to be able to cause some forms of insanity

Whut? Cell phone charging stations cause warts that lead to insanity?

OMG! Get 'em out of the Bay area!!!

Sno-o-o-ort >:-} ...Jim Thompson

Thanks, Win, but Tim is right, it should be fun and a bit different.

I've used packaged photovoltaic isolators before for micropower things - a few uA at a few V is quite usable and of course it's DC-DC with no switching so really simple to do.

In the present case, I rather think the customer would like the reassurance of a visible air gap, and if achievable, a low efficiency of

Cheers

Not blue;; go for infrared. The sensitivity peak for silicon is 900-1000 nm, and blue (420 nm) loses by a factor of two or three.

The mass-produced remote control LEDs are possibly your best bet; perhaps Vishay TSAL6100. They will probably need some heatsinking, if the rated 100 mA is applied.

My last couple of phones have had this built in. It's know as Xi (pronounced "chee"). I "hang up" my cell phone on one when I get home.

Can you suck power off the HV and ouput/input via an optoisolator? Use a battery? (You have to remember to turn off the HV before changing the battery... speaking from experience.) As Tim and others said I've gotten more photons from IR leds, you are going to have to stack several photo cells in series to get to 3V. 10mA each. LED's are not all that efficient in turning electrons into light. A diode laser might be better. We use these ~70mW ones that would give ~

(OK this raises an interesting question, why not blue lasers for lighting? (blast 'em into a phosphor prism thing-a-ma-bob.)

George H.

If you're the type of plug-in hybrid driver who finds it just too burdensome to plug in a cable, you can even have one for your car. ;-)

It's not that it's a burden. Micro-Bs have a limitied life. I'd rather charge it, every day, with the Xi xharger than cycle the connector another time. Sure, I'd like to have one in the car, too. You'll see more of them in cars, soon.

It's amazing that Apple didn't pick up the technogy. I probably would have gone with an iPhone when I got my current phone if they had Xi.

They don't talk about efficiency, but the charging rate is lowered by