T
Tony
Guest
Hi,
Hi all,
I'd like a simple but efficient non-isolated buck regulator to produce as much
current as possible at 3.3V from an input of 8V/10mA to 30V/5mA (a loaded-down
full-spec 48V phantom power supply can deliver about 24V/6mA). The ability to
handle 50V inputs would be nice, as it would then only draw the current actually
needed (many "phantom" sources aren't really up to spec). But it's not
essential, as I can clamp the input if necessary. The more efficiency I can get,
the more smarts and performance I can put into the product, but of course
cost/complexity/space are always tradeoffs.
The problem is all the chip makers seem to aim towards much higher currents, and
efficiency is always pathetic at such low input current. And I wouldn't trust a
pulse-skipping type near an audio circuit. I've lashed up a few simple discrete
designs also, but startup is messy, I've barely cracked 50%, and they tend to be
bulky.
I'm not expecting a ready-made solution here, but any new ideas would be
welcome.
Going one step further, down the track I'd also like to do something similar for
a new smart industrial sensor compatible with 4-20mA, HART, Profibus PA,
Foundation Fieldbus H1 and a 2 wire Modbus variant. In that case the typical
available input might range from 8mA/9V to 4mA/18V (not much, but presumably the
lower input/output voltage ratio might help efficiency a little).
Tony (remove the "_" to reply by email)
Hi all,
I'd like a simple but efficient non-isolated buck regulator to produce as much
current as possible at 3.3V from an input of 8V/10mA to 30V/5mA (a loaded-down
full-spec 48V phantom power supply can deliver about 24V/6mA). The ability to
handle 50V inputs would be nice, as it would then only draw the current actually
needed (many "phantom" sources aren't really up to spec). But it's not
essential, as I can clamp the input if necessary. The more efficiency I can get,
the more smarts and performance I can put into the product, but of course
cost/complexity/space are always tradeoffs.
The problem is all the chip makers seem to aim towards much higher currents, and
efficiency is always pathetic at such low input current. And I wouldn't trust a
pulse-skipping type near an audio circuit. I've lashed up a few simple discrete
designs also, but startup is messy, I've barely cracked 50%, and they tend to be
bulky.
I'm not expecting a ready-made solution here, but any new ideas would be
welcome.
Going one step further, down the track I'd also like to do something similar for
a new smart industrial sensor compatible with 4-20mA, HART, Profibus PA,
Foundation Fieldbus H1 and a 2 wire Modbus variant. In that case the typical
available input might range from 8mA/9V to 4mA/18V (not much, but presumably the
lower input/output voltage ratio might help efficiency a little).
Tony (remove the "_" to reply by email)