Yet Another Power Switchover Circuit - Single Cell AAA and 5V USB

[smaroukis]

Circuit layout is a 1.2V NiMH battery is boosted to 3.3V to power an MCU, and there is also a 5V USB input with a LDO.
When the battery is not present the boost converter should be disabled and the battery should be disconnected from the circuit. The MCU should also be able to disconnect the battery from the circuit upon low voltage condition.

Here's what I have so far, it is the standard series PMOS that is turned off when the USB voltage is present, since USB voltage will always exceed battery voltage.  An issue I see is finding a suitable PMOS with low Vgsth and low Rdson (the max current draw will be 100mA). I'm looking at the 1.8V logic level FETs that look like they will start conducting below -1Vgs but I don't have any on hand to test.

BOOST_EN is from the MCU which when low should disable the Boost regulator. The SW1 allows the user to enable the Boost regulator manually which is required on start up for the MCU to power up and hold its output high (probably need an RC circuit here to hold it high for some time).

On the Boost output since I am also connected the LDO 3v3 output to the 3v3 rail as shown I added a nBOOST_CTRL signal to disconnect the boost load side from the circuit.

Any advice is appreciated!

[Ian.M]

Cant work with a single NiMH cell.  Look at the TPS61023 datasheet: VIN_UVLO Under-voltage lockout threshold VIN_rising typ. 1.7V max. 1.8 V.   It will never start as Vin <=Vbatt and will never exceeded 1.5V, even just before end of charge.

[smaroukis]

the marketing prose on these data sheets ...

Yes I went through it again and startup conditions are >1.8V and Enable pin > 1.2V.

So I'll look for another boost or use a charge pump as input to this one.

[Ian.M]

or use a two cell series NiMH battery.   That will also give you much better efficiency (due to lower I²R losses).   The shutdown current of the TPS61023 is only 0.2uA max, and it shuts off its synchronous rectifier internal MOSFET during shutdown, + has output overvoltage protection so you don't actually need any power switching MOSFETs in the +batt to +3.3V path.   Simply pulling its EN pin low is sufficient.       

[smaroukis]

I see why every electronics product has at least two batteries now, but I did find a Microchip app note on single cell boost converters, as well as the MCP1640 which claims starting down to 0.65-0.8V

http://ww1.microchip.com/downloads/en/Appnotes/AN3506-Application-Note-DS00003506A.pdf?t

https://ww1.microchip.com/downloads/en/DeviceDoc/20002234D.pdf

[Buriedcode]

The MCP1640 is a great little chip, but its quoted efficiencies are overly optimistic (like all datasheets!).  Note there are 4 versions, with combinations of PWM/PFM and output disconnect on shutdown, and output bypass (pass through).

I don't think you'll get 100mA 3.3V output of it from a single alkaline or NiCD/NiMH. I haven't managed to get that from the SOT23-6 version, but the DFN version surprisingly did better, despite no indiation in the datasheet of differences.

From memory, the lowest startup voltage I tested it with was 0.95V, driving a single white LED @ 3.2V output.  I never did fully test one over the whole Vin range though.

There are many single cell boost converters, but not all are synchronous and as such, can't disconnect the output from the input during shutdown - it's much mroe common these days, but not a guaranteed feature.