Boost converter, but sometimes with input = output

[Alex Eisenhut]

If I have a marginal 5V power supply system, that may vary from 4.8V at full load and 5.2V at no load, and I want to boost anything below 5 to 5 and anything over 5 just passes through, what kind of boost topology am I looking for?
1A load.

[rteodor]

I just fried one module with FP5139 and had the datasheet open. I'm only posting it as an example:

[PGPG]

SEPIC is what comes to mind for the first. It will give you always 5V at output.
But I think that standard step-up can also be taken into account but it will have to work with extreme fill factors and I don't know what problems it can generate.

[jwet]

A regular boost converter will do what you want.  When the input is 4.8, it will do a very short boost to 5- low duty cycle and could be low peak currents.  When Vin is >5v + diode drop- won't run.  If you use a stout schottky for the boost rectifier, V drop will be about .2v so will be boosting ocassianally.  A PFM or similar will work best.  A constant frequency PWM will try to boost with diminishing duty cycle near zero.  Sepics will work but you'll have to spend some time getting them stable, not a great solution if you're operating right at the crossover.  The other option might be to do nothing- if full load is unusuall and high line doesn't matter, let it roll.

[Benta]

Buck-boost and SEPIC come to mind. Check here:
https://www.ti.com/power-management/buckboost-sepic.html

[Phil1977]

Pure boost will be much more economic and efficient in that case. As already said, it passes the voltage through its forward diode and gets only active if the output voltage gets below its set voltage.

Cuk and SEPICS will have to convert all DC to AC and back - that´s always lossy.

There is the 4-switch buck boost which could theoretically be set to a more or less passive pass through:


(Source: https://en.wikipedia.org/wiki/Buck%E2%80%93boost_converter#/media/File:Basics_of_the_4-switch_buck-boost_converter.png)

I don't know if there is any controller IC or off-the-shelf module that supports this mode. Probably a simple boost-converter will do best.

[Siwastaja]

+1 to bog standard boost topology. If high efficiency is important, then it needs to be a synchronous boost, but probably a 0.3V diode drop won't be the end of the day - it's like 6% efficiency loss.

[Alex Eisenhut]

Thanks everyone, efficiency is not terribly important. I will read datasheets more closely and probably review my switch-mode power supply notes from school... decades ago...

[mariush]

You could get a buck-boost regulator, they're cheap. 

ex  TPS631010 https://www.digikey.com/en/products/detail/texas-instruments/TPS631010YBGR/17748415


If you don't care about efficiency... add one or two 1n400x (1n4004 to 1n4007) diodes in series with your input voltage, to be sure your input voltage will be less than 5v at all times, then use a boost only 5v regulator. Even at the lowest current, a 1n400x will drop around 0.5v  ... a 1n5817 will give you around 0.3v drop at the lowest current.

[Siwastaja]

If you don't care about efficiency... add one or two 1n400x (1n4004 to 1n4007) diodes in series with your input voltage, to be sure your input voltage will be less than 5v at all times, then use a boost only 5v regulator.

Why this extra complication? OP specifically said that input exceeding 5V can pass through. Although, if the actual input maximum is just 5.2V and the boost output is set to regulate at 5.0V, then due to the Vf of the boost diode OP is never going to see over 5V output anyway.