0
\$\begingroup\$

I am designing a power supply circuit for a battery-operated device. My list of requirements:

  • Very low quiescent current (ideally under 1 uA), where by "quiescent" operation I mean no or very light load (< 10 uA)
  • High efficiency (> 90 % for >= ~2 mA load)
  • Input voltage range from 0.8 V to 3.8 V (guaranteed start-up at 0.8 V)
  • Output voltage of 3.3 V
  • Roughly 125 mA peak current (at 3.3 V at the load side)
  • Few external components

My current solution utilizes MAX17227J. This ticks almost all the boxes, except that it's a boost-only device: if fed with input above 3.3 V, it switches to a "pass through" mode, and the output voltage is equal to the input voltage. Unfortunately, the rest of the circuit requires ~3.3 V (with a 5 % tolerance).

Motivation: I want to be able to operate the device with both Li/FeS2 (1.8 V full charge, 0.8 V cut-off) and Li/SOCl3 (3.7 V full charge, ~2 V cut-off) batteries. Update: And I also want to be able to tell the customer that "any AA-sized battery will work". Some battery chemistries, such as NiMH, discharge very low.

I've been looking very hard for a suitable part, but it does not seem to exist: every buck/boost wired converter I have found so far is either:

  • Unable to start below 1.8 V, or
  • Unable to accept more than 1.8 V.

For example, the EFP01 family from Silabs is great, and they even have a "Single cell boost" device, but it suffers from this exact problem.

Two solutions come to mind:

  • Give up on supporting both battery chemistries. Although this technically solves the problem, I really don't want to.

  • Include two devices: a buck converter and a boost converter (or a boost converter and a buck/boost converter) and depending on the battery voltage, decide which one to run. This seems quite complicated and will increase the cost of the device.

What doesn't work:

  • Putting a buck after a boost: the compound efficiency wouldn't be great, and efficiency is the most important factor here.

  • Perhaps putting a buck converter after a boost converter could work, providing that: the boost converter would boost to 3.3 - e V and pass through anything above; the buck converter would buck to 3.3 + e V and pass through anything below; e would be small enough. I don't think there's an appropriate combination of devices to achieve this: either both would be running, resulting in poor efficiency, or they would interact in possibly weird ways, or 2e is above the 5 % tolerance.

Am I missing something?

\$\endgroup\$
11
  • \$\begingroup\$ hm, When I click through to TI.com's DC/DC switching reg -> buck-boost&inverting regs- > buck-boost… (integrated switch) and enter the requirement that V_in (min) <= 1.8, I get several recopmmendations. You sure you can't just solve this by buying one of these? For example, the TPS631000 seems both to cover all your input voltage as well as your efficiency requirements. \$\endgroup\$ Apr 3 at 13:52
  • \$\begingroup\$ I get several hits at Analog too. Particularly from 1.8V and up but also lower. Something along the lines of this? analog.com/en/products/ltc3106.html Oddball part but the dual input could as well use different battery chemistries yeah? \$\endgroup\$
    – Lundin
    Apr 3 at 14:01
  • \$\begingroup\$ @MarcusMüller The TPS631000 doesn't appear to be able to start up at 0.8 V input. In fact, I don't even see any indication that it can run off voltages less than 1.6 V, even after startup. \$\endgroup\$
    – Hearth
    Apr 3 at 14:09
  • \$\begingroup\$ @Hearth huh, you're right, top of the question asks for 0.8 V startup, middle of the question says they haven't been able to find anything for 1.8 V. Considering the requirements with 0.8 V lower-end battery voltages, I have latched onto the wrong number. \$\endgroup\$ Apr 3 at 14:11
  • 1
    \$\begingroup\$ @Marcus Müller That's a good point! However, I have simplified things a little. It's the Li-FeS2/Li-SOCl3 dichotomy that's causing the most trouble, but my ultimate goal is to be able to say, "just stick any AA-sized battery into it and it will work, albeit for best results, use either Li-FeS2 or Li-SOCl3". Most of the NiMH datahseets I checked, for example, show that these batteries operate around 1 V when half-discharged. Sorry about that, I will update the question. \$\endgroup\$
    – Foo
    Apr 3 at 17:37

1 Answer 1

0
\$\begingroup\$

Such low input voltages make things a bit difficult, but there are a few options out there.

The MAX20310 is an inverting buck-boost that requires a negative input voltage (it's designed to work off a battery with its positive terminal grounded). It can operate down to -0.7 V input, and is guaranteed to start up from a -0.8 V input. It's designed to provide multiple output voltages, and requires programming to set its outputs--it's not practical in a system that doesn't involve a microcontroller.

The LTC3106 in conjunction with a backup battery (which it can also recharge, if a rechargeable cell is used) can start up off of inputs as low as 0.3 V, or without a backup battery it can start off of 0.85 V inputs. As a four-switch converter, its efficiency is pretty poor at low input voltages, but it does do better as input voltage increases.

I've checked TI as well, but they don't seem to have any buck-boosts that can start off of such low voltage, as you said. The lowest I can find requires 1.3 V (the TPS6380x series).

\$\endgroup\$
2
  • \$\begingroup\$ I did find the LTC3106 in Analog's portfolio but ruled it out (mostly on the grounds that it's geared towards energy harvesting, so I figured I must be doing something wrong). But the MAX20310 I missed, I'll check it out. Thank you! \$\endgroup\$
    – Foo
    Apr 3 at 17:29
  • \$\begingroup\$ I might be wrong, but it seems that this part only accepts input voltage under 2 V (above -2 V)? If so, that's unfortunately one of the problems listed above - when it can start low, it usually doesn't handle above 1.8 V (or in this case, 2 V). \$\endgroup\$
    – Foo
    Apr 3 at 17:46

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.