General > General Technical Chat
Idea: MEMS electromechanical boost DC DC converter - works from very low voltage
(1/2) > >>
daqq:
Hi,

Thought experiment:

I've been looking into DC DC converters that can start up from a low voltage. There's not that many of them and the lowest I've seen is able to start from 13mV ( https://www.mdpi.com/1996-1073/13/20/5501/pdf ). Now that's impressive as hell, but I was thinking if it would be possible to achieve an even lower startup voltage using a MEMS solution to provide the initial kick.

Say you had a MEMS version of a conductive bimetalic strip that would act as a thermal cutoff switch - above a certain temperature it would disconnect, below it it would stay connected. This would be connected in series with a transformer (or inductor - boost topology type thingie). "Schematic" included. Once voltage is applied to the device, the switch starts to heat up - 1mV might not seem like a lot, but at microscopic scales it should be able to heat a very small volume by a LOT.

Once the MEMS switch disconnects, the transformer or an inductor will get a current change and a jolt of much higher voltage will appear on the secondary, allowing the DC DC converter to start.

Because this is a "mechanical" solution, it doesn't have to deal with low voltage semiconductor physics at all, so if you can create a MEMS thermal cutoff switch that disconnects itself when you apply 1mV to it, you can get a converter that starts at 1mV.

Now obviously there's a lot of drawbacks, but MEMS relays exist, as do all sorts of weird and wonderful structures. So...

Do you think it would work? It's just a thought experiment.
daqq:
Without transformer.
Kleinstein:
The mems relais are more like very high performance / low leakage switches, not high power. Not sure the contact rating is that great and there may be some wear from very frequent switching.

At small scale transformers / inductors get worse and would need increasing freququency.
daqq:
Kleinstein: Thanks for the feedback. But there would only have to be a few switchings really - just to start up the converter, say, 10 per startup? Afterwards the whole MEMS block gets disconnected through the bottom bit - say a depletion mode MOSFET.

And the power would not need to be high. And none of the other parts need to be MEMS at all.
T3sl4co1l:
You may be better off looking at JFETs or DMOS.  The trick is to set up a zero-bias [voltage] amplifier, with adequate transconductance to overcome load resistance and start oscillating.

The available power is still extraordinarily small, at best usable to bootstrap a conventional converter, which again will still only deliver tiny outputs (fractional watt).

Seebeck modules are made of numerous junctions for a reason; I'd recommend sticking with that, rather than trying to force-fit single junctions or something like that.

Tim
Navigation
Message Index
Next page
There was an error while thanking
Thanking...

Go to full version
Powered by SMFPacks Advanced Attachments Uploader Mod