> If you can switch from 5V logic to 3.3V or 3V logic you only need an LDO =
with
> low shut down current.

Or, if he can tolerate the losses he can use just a boost converter
when Vin falls too low.
For Vin =3D 4-6V, Vo =3D 5V, you can use a linear regulator when Vin >=3D a=
bout 5.4V.
That's 0.6/2 =3D 30% of the range.
>From 5.4V down you need a boost converter, but efficiency will be high
when Vin is close to Vout. Even at 4V in the efficiency will be OK.
eg if actual conversion efficiency is 50%, at Vin=3D4V, you need a 1V
supply "standing on" the 4V pedestal.
Overall efficiency is 4V at 100% and 1V at 50% or 5/6 =3D 83%.
At voltage above that it gets better.
At the top end with Vin =3D 6V you are again back to 85% with a linear regu=
lator.

If this is a battery supply then Vin mean is perhaps about 4.5V giving
a mean conversion efficiency of perhaps 90%+.
YMMV depending on supply time/voltage characteristics.

___________________

Ugly truck ...


Long ago I (and Olin did the same subequently and independently)
produced a smps to LDO solution where the SMPS target voltage was ~=3D a
Vbe above Vout. He and I placed a bipolar be junction across the LDO
and drove the smps to hold the transistor just at turn on, giving
about 0.4VF of headroomn to the LDO.

In that case the smps used a CD40106 hex Schmitt inverter package
driving a bipolar - but a FET would do as well or better.
Circuit (probably posted to PICList in June 205 and probably since)
"R100 Ugly SMPS 70.jpg" attached.
[[Cut and paste extract from larger circuit, please excuse quality etc
etc ]]. Operation is more subtle than may be obvious - actually
providing turn on pulse width modulation under light load.

In this case the aim was to preserve a 5 Volt logic supply in a system
operate from 4 C primary cells in the presence of a high current drain
motor system. [[Don't even start to ask why the original designer
thought he could run 5V logic in such a way]]. End circuit idles at
well under 100 uA unloaded (the LDO quiescent current is the real
limit). This does NOT of course turn off when not operating, but an
extra MOSFET would allow that.




 Russell
--=20
http://www.piclist.com PIC/SX FAQ & list archive
View/change your membership options at
http://mailman.mit.edu/mailman/listinfo/piclist
.