Searching for a 3.3V Buck-Boost converter

[PeterFW]

Hello!
My humble dream of a DIY handheld µC thingy has only one big hurdle left to take, the voltage regulator. Running straight from the battery is not possible.
What i would not like to use is a buck, boost or LDO regulator.
But i got a few LDO selected and board routed to test them.

Vin:          3-4.2
Vout:        3.3
Iout:         ~400 to 500mA
Iq:            <=50µA @ 500µA load
The catch: Package should have legs to solder by hand.

Right now the only one i managed to find is the LTC3440.
But the power safe mode has to externally controlled by the MCU, that is not that big of a problem but maybe there is a "common" alternative out there that has a build in power safe feature.

Greetings,
Peter

[Paul Price]

HIGH EFFICIENCY SINGLE INDUCTOR BUCK-BOOST CONVERTER WITH 1-A SWITCHES     TPS63030, TPS630311

FEATURES APPLICATIONS
2• Up to 96% Efficiency • All Two-Cell and Three-Cell Alkaline, NiCd or
• 800-mA Output Current at 3.3V in Step Down NiMH or Single-Cell Li Battery Powered
Mode (VIN = 3.6V to 5.5V) Products
• Up to 500-mA Output Current at 3.3V in Boost • Portable Audio Players
Mode (VIN > 2.4V) • PDAs
• Automatic Transition Between Step Down and • Cellular Phones
Boost Mode • Personal Medical Products
• Device Quiescent Current less than 50?A • White LEDs
• Input Voltage Range: 1.8V to 5.5V
• Fixed and Adjustable Output Voltage Options DESCRIPTION
from 1.2V to 5.5V The TPS6303x devices provide a power supply
• Power Save Mode for Improved Efficiency at solution for products powered by either a two-cell or
Low Output Power three-cell alkaline, NiCd or NiMH battery, or a one- cell Li-Ion or Li-polymer battery. Output currents can
• Forced Fixed Frequency Operation and go as high as 600 mA while using a single-cell Li-Ion
Synchronization Possible or Li-Polymer Battery, and discharge it down to 2.5V
• Load Disconnect During Shutdown or lower. The buck-boost converter is based on a
• Over-Temperature Protection fixed frequency, pulse-width-modulation (PWM) controller using synchronous rectification to obtain
• Available in Small 2.5mm × 2.5mm, QFN-10 maximum efficiency.

[T3sl4co1l]

A "flying inductor" buck-boost, or coupled inductor SEPIC, would do nicely here.  The disadvantage to the SEPIC is it's always switching; but the logic inside a typical buck-boost is always "twiddling its thumbs" even when Vin ~= Vout.  Also needs twice the switches (and that counts synchronous rectification on the SEPIC).

In reality, I expect the difference in efficiency is a wash, and limited more by quality of the inductor and limitations of the controller (how fast is it, what's minimum/maximum duty cycle on the switches, how well timed are the switching events, etc.) than the topology itself.

Don't have any recommendations offhand though.  I'd search TI's listings, they have good stuff for the most part.  Without looking, I'd bet Paul's recommendation is good.

FYI, you should probably relax your package requirement; although, you don't specify how long you want the legs, so I would just as well assume you mean SOIC and TSSOP, because they have legs, which *can* be soldered individually by hand.  I've done it before...not that you necessarily *want* to.   If you literally just meant "no DFN/QFNs", I agree with that.

Tim

[PeterFW]

Don't have any recommendations offhand though.  I'd search TI's listings, they have good stuff for the most part.  Without looking, I'd bet Paul's recommendation is good.

TI is allways my first stop when searching for a nice chip, but...

FYI, you should probably relax your package requirement; although, you don't specify how long you want the legs, so I would just as well assume you mean SOIC and TSSOP, because they have legs, which *can* be soldered individually by hand.

Yes, that is what i wanted to say. I do not care how small, as long as it has itty bitty tiny little legs.
I can solder SSOP, MSOP, TQFP, nearly annthing like that and will do so happily.
My first attempt at SMD soldering was a SSOP28 deadbug, wich still runs fine to this day, many years later.

But QFN, SON, BGA and so forth are the devils work, i can solder them on a board with dirty tricks (big via, many small via, long pads) but i do not want to.

And that is the problem, neither TI, Maxim or Analog has something "with legs".

The only one i found was the LTC3440 that meets my criteria.
The LTC3530 has a automatic PFM (low load efficiency) mode but will not be that efficient at 25mA load.

The thingy id like to build will do one of four things, worst case calculations:
<500µA standby, display off.
~1mA standby, display on.
~25mA active, idle.
~150-300mA active, writing to SD card, transmitting wireless data.

Since this is only for myself, a single unit, i do not want to wander into the realm of reflow and solder paste.

Greetings,
Peter