Theoretical efficiency of opto-coupling for power applications

Thread Starter

xox

Joined Sep 8, 2017
391
So this is more of just a little thought experiment really, but just out of curiousity I wonder if it is currently (or nearly so) possible to say take a high-quality solar panel and then mount a sufficently large and dense panel of high-powered LED's to it (may of course need some distance and a lens to avoid damage due to direct heating) in such a way that the transfer of energy could be achieved at reasonably efficiently rates (like > 50%)?

Seems like an interesting way to isolate a circuit...but then again, coils are probably WAY more efficient! Maybe just for situations where electromagnetic interactions are for some reason not desired? I don't know.
 

Thread Starter

xox

Joined Sep 8, 2017
391
X -> Y ~~~> Y-> Z

For power transmission it's a dog walking on hind legs.

DC to photonic: https://www.sciencedirect.com/topics/engineering/plug-efficiency
resulting in an overall wall plug efficiency of typically 40%–50%

Photonic to Photonic: Lets just give it 99% if it's a short distance.

Photonic to DC: https://www.nrel.gov/pv/cell-efficiency.html
Expect realistic efficiencies in the range of 15-20%.
Okay so roughly 6%-10%. In other words...abysmal! You'd have to be working on some pretty exotic problem to justify that kind of energy waste. Unless you're talking like really low voltages and currents. But then why would anyone need isolation for a 5 or 12 volt circuit in the first place?

I still need to finish that first article anyway, very interesting read so far. Thanks for the awesome breakdown by the way.
 

crutschow

Joined Mar 14, 2008
24,318
why would anyone need isolation for a 5 or 12 volt circuit in the first place?
Some applications require that input and output grounds be isolated (such as if the input is derived form the mains voltage without transformer isolation).
 

Thread Starter

xox

Joined Sep 8, 2017
391
Some applications require that input and output grounds be isolated (such as if the input is derived form the mains voltage without transformer isolation).
I'm still wondering how it could possibly be more appropriate than choosing a simple transfromer though. Magnetic sensing equipment maybe, or perhaps more deterministic control over voltage spikes and such?
 

MrAl

Joined Jun 17, 2014
7,120
So this is more of just a little thought experiment really, but just out of curiousity I wonder if it is currently (or nearly so) possible to say take a high-quality solar panel and then mount a sufficently large and dense panel of high-powered LED's to it (may of course need some distance and a lens to avoid damage due to direct heating) in such a way that the transfer of energy could be achieved at reasonably efficiently rates (like > 50%)?

Seems like an interesting way to isolate a circuit...but then again, coils are probably WAY more efficient! Maybe just for situations where electromagnetic interactions are for some reason not desired? I don't know.
Hello there,

I actually suggested this idea a long time ago when it came to biasing an op amp with a small negative supply voltage so the output could get all the way to a perfect 0.000000 volts.
A photocell and small LED is good enough as long as the op amp current is not too high.

As far as power transfer goes though it's quite inefficient. It would take probably 5 times as much power to put in as to get out. But there's another stumbling block in here and that is the available surface area. I would say you'd have to have a reasonably large surface area to transfer a decent amount of power like say to charge a cell phone at a reasonable rate. You'd also have the light spillover to think about which could be objectional.
Now for low current isolation applications such as biasing a chip with both LED and photocell would be nice as that would achieve the end result in one package. In fact an glass body diode acts as a photocell too although the surface area is very small.

So there are some ideas for this but i dont think power transfer is on the list unless it is very low power.
 

MrAl

Joined Jun 17, 2014
7,120
The photovoltaic mosfet driver is an example of using photonics for power transfer. Tiny power and not efficient but another example of using light for this purpose. See:

http://www.vishay.com/solid-state-relays/list/product-83469/
Hi,

You are right in that opto couplers that use that method or any method for that matter do transfer power, but the power level is so small that it is more like signal processing than real power transfer. When we talk about power transfer we usually are thinking of much higher levels where the efficiency becomes a real issue. A wireless cell phone charger would be a good example.
 
Last edited:

crutschow

Joined Mar 14, 2008
24,318
Silicon solar cells are sensitive to near IR light, which takes less energy to generate than visible light (IR LEDs have a lower forward voltage drop than visible LEDs), so I would expect for best efficiency you would use high efficiency IR LEDs to transfer energy to the solar cell.
 

Thread Starter

xox

Joined Sep 8, 2017
391
You are right in that opto couplers that use that method or any method for that matter do transfer power, but the power level is so small that it is more like signal processing than real power transfer. When we talk about power transfer we usually are thinking of much higher levels where the efficiency becomes a real issue. A wireless cell phone charger would be a good example.
Well exactly. It has to be producing power for some circuit obviously. But if you did have say a low-powered CPU or something you could probably provide enough energy this way. Like a laser to a lens mounted in front of a solar panel such that the beam spreads out across it fairly completely. You might get a lot of distance out of a setup like that actually.
 

Thread Starter

xox

Joined Sep 8, 2017
391
Silicon solar cells are sensitive to near IR light, which takes less energy to generate than visible light (IR LEDs have a lower forward voltage drop than visible LEDs), so I would expect for best efficiency you would use high efficiency IR LEDs to transfer energy to the solar cell.
I learned an interesting thing about carbon too. Just out of curiosity, a friend and I once decided to hook up his wedding ring to a little lithium battery (via some aluminum foil and pencil leads) and in fact we were able to get a remarkable amount of power out of it. And the response seemed a lot faster too. The moment you close the switch the meter just immediately jumps! So yeah, too bad the stuff is so overvalued, it seems like a great material for making solar panels.
 

MrAl

Joined Jun 17, 2014
7,120
Well exactly. It has to be producing power for some circuit obviously. But if you did have say a low-powered CPU or something you could probably provide enough energy this way. Like a laser to a lens mounted in front of a solar panel such that the beam spreads out across it fairly completely. You might get a lot of distance out of a setup like that actually.
Hi,

Im sure there are extreme examples out there.
 
Top