Buck-Boost DC/DC Converter question

I'm happily soldering around right now, exploring the wide world of crafting my own regulated LED driver. I've toyed around with a couple of the staple drivers from Maxim and TI, but I've ended up with a question for our resident electronics experts:

The majority of small form factor DC to DC converters tend to have what appears to be a fairly narrow input range (for flashlight friendly applications, at any rate). 2.4V to 5.5V seems to be standard fare. This hasn't been a problem in my initial few soldering birds nests because I've been driving my circuit via a single 18650, but I have been trying to figure out the best option for driving a light from a wider variety of input voltages. I'm trying to make myself something that could be driven from a single 3.7 as well as from several in series (using a FM 3x18650 holster, for a lark), but that would leave me outside the input range of the handful of converters I have been using thus far.

On that note, I have no idea what happens when you overload a converter. This may be a task for me and my fluke and some safety goggles this evening. It IS Friday, after all.

At any rate, does anyone have a recommendation for a buck-boost converter that could run from, oh, 1.5 to 15v or so? Google has not been my friend in this regard, but I could simply be searching for the wrong things. Additionally, I may well have not given enough information to garner a useful response, but hey, gotta start somewhere.

I'm driving one of these beauties, and eventually I'd like to see about cramming it into something that could fit in a 6P.

Thanks for giving this a quick look. I've been a lurker for far too long, figure I may as well poke my head around.

Good question ... would like to hear an answer from an expert point of view too. Best examples at the moment are the driver from the Jetbeam Jet III IBS (only 1x 18650 allowed) and the the GD series from the Sandwiche Shoppe (up to 5.5V) ... all these (fine) buck-boost drivers are limited to the use with one LiIon cell. You can call it almost a logic conclusion, that all current lights which are supporting 18650 and 2x (R)CR123, use "buck only"-drivers and are going to direct drive with single LiIons. Well, this concept has some advanteges too, but why is there no driver that combines the advantages of current buck-boost and buck drivers ... a constant flat regulation both with one or two cells. There must be some kind of technical barrier I guess ...

Hi there,


The basic question seems to be is it possible to build a converter that can handle
1.5v to 15v input range.

One way to tackle this is to build a pre-converter that converts 1.5v to say 5v to
power a circuit that can convert over that range but needs 5v to operate properly.
The circuit gets 5v from this pre-converter to power the control circuit and drive
circuit, while the main power transistor still gets its power from the input source.
A MOSFET output transistor would be best.

It would be easier if the lower range was at 2.5v, where you could use a uC ic
to start with.

Another idea might be to use a regular circuit that goes from 1.5 to say 5v, and
if the input exceeds 5v switch to a buck converter front end. Not as efficient
but it should work.

So what you are looking for is the all season of batteries, or as the experts call it, the wrong tire for every season. I could build a converter that worked over this range, but it would be compromises at ever step...and you would not like it most of the time.

I think you are better off taking a step back and thinking about how you would really use any given light and design the converter that matches that. Other than using one battery (small) or two (runtime/output) I have never felt a need to have that much range on input. I can always put extra batteries somewhere (where they are not heavy in my hand).

On the 2.4 - 5.5V range that is a combo of what is needed for LiIon batteries plus the limits of low voltage CMOS processes for making the ICs.

Is it possible to build a converter that really consists of several converters optimized for their own narrow range, and then have some switching logic to pick between which one to use depending on input voltage?

SemiMan said:

So what you are looking for is the all season of batteries, or as the experts call it, the wrong tire for every season. I could build a converter that worked over this range, but it would be compromises at ever step...and you would not like it most of the time.

I think you are better off taking a step back and thinking about how you would really use any given light and design the converter that matches that. Other than using one battery (small) or two (runtime/output) I have never felt a need to have that much range on input. I can always put extra batteries somewhere (where they are not heavy in my hand).

Click to expand...

The wrong tire for every season comes into play quite a bit during my day job as a software engineer, so I know exactly what you're getting at. I think I am having a bit of capability envy reading the spec sheets of a few Chinese manufacture lights that I have never had a chance to look at first hand, rather than an air tight design like the natulus. There really should be no reason to need so much of a range in my input voltage for most practical purposes.

I think it mostly started from wanting to make a 6P dropin that would run on either 2 x CR123 or 1 x 18650 or similar. I can easily enough work around some of these perceived limitations for the time being and try and make something that works for my current setup (18650), then move forward from there.

Hell, it's a learning experience at this point, so I may as well learn. Thanks for the input. It's good to know it CAN be done, even if it potentially SHOULDN'T.

A 15v input seems highly unusual, and seems that if you have enough cells to get 15v you would rather have one or more of those in parallel for longer runtime. At any rate, I was looking around at maxim ic's and found this circuit: http://www.maxim-ic.com/appnotes.cfm/appnote_number/1021/
I'm not sure if this helps you or anyone else out there.

I've developed a converter a few years back that could both buck and boost. Minimum input is 1.2 volts and maximum input is 9 volts. Output current is only 400ma. I've recently combined a microprocessor with this converter and now has 3 levels of light with strobe and beacon features. Board width is 14mm and thickness is 1.6mm. Compatible in size with most of the converters sold at the shop.

ARC mania

My understanding is that, to do this, you would combine a voltage buck regulator, with a filtered output, with a boosting current regulator. The buck would be designed to always buck to ~2.5-3V (depending on what you want the lowest setting to be on the output), and the boost regulator would take it from there. The Voltage buck regulator would need to go "passive" when the Vin drops below the Vout, and be able to handle the higher current that the boost regulator would draw when the voltage dropped down low. It's very do-able in my understanding, but like others have said, the efficiency is not so great as you have more components in the system wasting energy as heat. In theory, a voltage bucking circuit could be designed to be ~95% efficient or better, but most boost regulators are around 80-90%, final efficiency would be around 85% or less most likely...

This may seem fine, but considering that a buck-only driver, can achieve 95+% efficiency, you would be sacrificing 10% or more efficiency when you have a high supply voltage. The "wrong tire for all seasons" analogy is very appropriate here.

Eric

Hi again,


These days in electrical engineering it seems that nothing is impossible. It just
depends on how much time you want to spend on it and how much money you
want to spend and what size constraints influence the design. Building flashlight
circuits is definitely on the low end of the spectrum, except sometimes when
the size is important.

For a special purpose buck boost, i think the main
issue would be the cost because if it works that good a little bigger flashlight
probably wont matter much, and a specialized circuit is going to cost more.
The preregulator alone would have it's own chip, inductor, cap, etc., although
it would be designed for low current.

There are a number of topologies that could achieve the buck boost goal,
and with a livable efficiency too. The question is which to choose...
buck boost, flyback, sepic... cascade isnt the only choice.

When i was talking about a preregulator previously in this thread i wasnt talking
only about cascade, i was talking about a preconverter that would only have
to supply the rest of the circuit with its bias, not with it's main power. The
main power still comes from the battery. The only catch here is you have to
start with chips that will allow the independent supply of power to the last
drive stage. Some chips allow this via a separate pin, some dont.

I dont think it is out of the question to create a design like this. For example,
one simple way to get buck boost with very good efficiency is to build two
circuits...one buck, one boost. When the battery is high, switch on the
buck circuit, and when low switch on the boost. Use diodes to logically OR
the two outputs or even low sat transistors. If you dont like that, build
two circuits AND use two LEDs, where one LED is driven when the buck
is on and the other LED when the boost is on. I would probably be happy
with the diode OR technique, or transistor version.

Anyway, the way to start this would be to hunt around for some chips
that have a separate pin for the output stage drive, or else settle for
a general purpose chip that is only a controller chip with no output
transistors. This way the preregulator i talked about could power the
chip alone so it always functions normally, while the output transistor(s)
are powered from the battery. A tiny Zetex boost might work for
the preregulator.