0
\$\begingroup\$

I have a battery powered system that currently supports up to 5.5V input voltage (which amounts to 2-3 AA batteries in series.)

I'd like it to support up to 4 batteries in series (so up to 7.2V input so have some margin.)

What kind of circuit would you recommend to put between my input voltage and my system power input to allow my system to consume higher than 5.5V input without exposing my inputs to greater than 5.5V?

It's obviously very important that this circuit not put more than a negligible marginal quiescent load on the batteries (less than 100nA ideally.)

What circuits / strategies do you have up your sleeve to help me out here? Also of consequence, after functionality, is certainly marginal cost of the added circuit.

Update In response to several of the comments, allow me to clarify.

What current does your system take on full load and standby and at any other relevant time that you feel may be appropriate to mention? What is the minimum voltage your system needs and, is 5.5 volts an absolute maximum limit?

During standby the load on the battery is about 1 microamp. During full load, the peak current is around 150mA, but my 3.3V regulator (TPS63050) is capable of delivering up to 500mA. In principle it can accept an input voltage anywhere from 2.5V to 5.5V, and those limits are consistent with other components that are at the periphery of my system that directly interface with the unconditioned battery voltage.

Are you serious about the 100nA quiescent current limit? How do you define quiescent...unchanging inputs? clock stopped? system went to sleep?

What I meant by quiescent in this case is that when the system is in its low-power mode, consuming nominally 1uA, I don't want to increase this figure by more than 10%, ergo 100nA, if possible. My survey of available switching buck converters suggests that's not a realistic expectation, the lowest quiescent current I've identified is the TPS62745 at 400nA, and I could live with that.

However, is there a much simpler approach, like a few diodes in series with the battery voltage? Even at 1uA current draw, could they drop the voltage into range during sleep mode? Talk me out of that?

\$\endgroup\$
11
  • \$\begingroup\$ Two questions that may or may not be relevant: 1. Do you care about the conversion efficiency (linear vs. switchmode) 2. What's your lower voltage limit? \$\endgroup\$
    – pipe
    May 16, 2020 at 16:45
  • \$\begingroup\$ What current does your system take on full load and standby and at any other relevant time that you feel may be appropriate to mention? What is the minimum voltage your system needs and, is 5.5 volts an absolute maximum limit? \$\endgroup\$
    – Andy aka
    May 16, 2020 at 16:49
  • 1
    \$\begingroup\$ Are you serious about the 100nA quiescent current limit? How do you define quiescent...unchanging inputs? clock stopped? system went to sleep? \$\endgroup\$ May 16, 2020 at 16:49
  • 1
    \$\begingroup\$ What are the 1.8 volt batteries you are currently using? \$\endgroup\$
    – Andy aka
    May 16, 2020 at 16:53
  • \$\begingroup\$ What's the reason why you need 7.2V? Instead of two sets of 5.5V serie batteries? Anyway a mini switchmode converter seems to be the obvious solution. Costs about $1.5/pcs. \$\endgroup\$
    – Fredled
    May 16, 2020 at 20:49

1 Answer 1

0
\$\begingroup\$

I'd simply replace your TPS63050 with TPS62745 or TPS62056 which support your desired input voltage range.

An alkaline battery has a best-case self-discharge rate of about 2% per year. On a 2000 mAh battery this represents 40 mAh/year or a continuous equivalent discharge current of 4µA.

Therefore having a standby current as low as 100nA is unnecessary. The batteries will be dead and corroded long before a difference between 100nA and a few µA is noticed.

If this is about extending battery life, then a linear regulator would be useless, and optimizing the DC-DC efficiency when the circuit is drawing 150mA might be a better option, although that depends on the duty cycle.

For example if it draws 150mA at a 0.1% duty cycle, this corresponds to 150µA average. An efficiency gain of 1% on this due to a better DC-DC or component choice will be 1.5µA which would swamp the DC-DC's idle current. So you have some math to do. Depending on the duty cycle, the DC-DC with longest battery life may not be the one with lowest idle current, efficiency is important too.

\$\endgroup\$
1
  • \$\begingroup\$ Unfortunately there is other circuitry on the current battery voltage node that is only 5.5V tolerant, and for which it is essential that it not be "behind" the regulator. \$\endgroup\$
    – vicatcu
    May 19, 2020 at 0:28

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.