Constant output voltage buck-boost converter

[CallumA]

Hi! This is my first post on this forum (although I'm not completely new, I've just been reading without membership).

Anyway, I've been looking for a method of getting a constant output voltage from a buck-boost but am finding it quite tricky (Google is being difficult, only promising thing is an IEEE page but I can't actually find anything but a piece of text saying it's somewhere else).

The idea is I can have 4 AA batteries with a maximum total output of, say, 6.6v and as they discharge a minimum of 4v (looks like the discharge curve just cuts off around there) and have the output be constantly around 5v for some logic ICs which have analogue measurements quite sensitive to voltage so a voltage around 4.75 to 5.25v should be the largest range permissible (although of course the closer to 5v the better).

I would also like to use this for USB voltage output (completely different project though) and that same voltage range would be perfect for that.

Any ideas to get the voltage constant? Full schematics would help (if you don't mind the effort that is, otherwise I could probably work it out)

Thanks in advance!

 

[(*steve*)]

Yeah, some type of buck-boost power supply would do it for you.

There are two types. The first is to have a boost regulator followed by a buck regulator. Te boost regulator will be possibly in the 80-85% efficiency to boost the input up to (say) 7V. Then a buck regulator can drop it to 5V at around 90% efficiency. So your overall efficiency will be about 75% (ish).

The other option is a topology which can either boost or buck. This seems better since there is only a single conversion, but these topologies tend to be far less efficient overall.

In both cases you can purchase modules very cheaply, although there is no reason why you couldn't build your own if you've had some experience with construction.

Here is an example of the buck-boost in the second topology. Note that it has a single regulator chip and a single control to set the voltage. It also has 2 inductors on the board.

Here is an example of the former. It has two chips on board (one for boost, the other for buck) and it will frequently have more controls (separate voltage control for boost and buck stages. This one may also have current limiting as well.

You will note that one design is significantly cheaper than the other. Both are probably cheaper than you could build them yourself.

 

[KrisBlueNZ]

Hi Callum and welcome to the Electronics Point forums

My first suggestion would be to use more cells! If you use six cells instead of four, your minimum voltage will be 6V and you can use a buck converter.

If that's not feasible for some reason, another option might be an inverting converter. These can buck and boost, and they only use one inductor and one diode. The output voltage is the opposite polarity to the input voltage, but this isn't an issue provided that your input source is fully floating with respect to the output.

The converter is non-isolated, and the negative terminal of the input is the positive terminal of the output. This should be OK if your input is just a battery and there is no chance of an external connection being made (e.g. a battery charger with an earthed negative rail plugged in to charge the battery when the regulator's negative output is also earthed).

Several manufacturers make suitable devices: Linear Technology (http://www.linear.com), Maxim (http://www.maxim-ic.com), Texas Instruments (http://www.ti.com) and others. Personally I recommend Linear Technology.

 

[CallumA]

Thanks for your replies.

I like the ideas of the independent buck and boost as well as the cell number change.

With the cell change I have an issue with upping the size so I'd have to see what I can do but I was wondering if maybe I use 3 or 2 cells for a maximum of 4.8 or 3.2v and used a boost converter would that be too inefficient? It would be nice to have a good efficiency so if boost converters are too inefficient I will see what I can do with the 6 cells.

With both methods, the problem I'd have is that I can't work out a way to adjust the converter to output a constant voltage. I couldn't come up with anything so I had a go thinking about using a microcontroller (I would like to avoid that as an option, bit overkill) but even then I would need to get an acceptable voltage reference source which I have no clue about. Also, I would have to make a divisor for the output voltage to make sure it's maximum is in the minimum for the microcontroller and then I could control something like a digipot.

Any ideas on that one?
Thanks

 

[(*steve*)]

Getting a constant voltage is what regulators do for a living.

If at effective end of charge the voltage of your batteries exceeds 5V (plus whatever regulator margin is required) then a regulator will work.

For greatest efficiency a switchmode regulator is generally the best choice.

The adjustment is simply set once and forget.

 

[KrisBlueNZ]

I think generally a buck converter will be more efficient than a boost converter. If you're trying to get every little percentage point out of the converter, you could look into synchronous rectification. This can be used with buck converters; I'm not sure whether it's practical with boost converters though.

That's why I suggested six cells. If you can make room, I think that would definitely be the best solution.

I don't understand what you're asking about a constant output voltage. These converters are always regulated; the converter IC itself includes a voltage reference and the output voltage is set by comparing that reference to the feedback voltage, which is taken from the output - either directly (and fed through a voltage divider that is within the controller IC), or through an external voltage divider. The controller varies the switch duty cycle so the feedback voltage, after division, is equal to its reference voltage.

If you want to vary the output voltage, you can do this in various ways. A digital potentiometer in the feedback loop is one way, but you can also vary the output voltage over a moderate range by modifying the feedback loop and using a variable voltage, which can be generated from a PWM signal from a microcontroller. This is only required if you need an adjustable output voltage.

 

[CallumA]

Ah, I see now, I always thought buck converter outputted a voltage that was proportional to the input voltage but now I realise that doesn't make sense as you could just due a divider.

Anyway, I'll try using 6 cells, I came up with the idea of using AAAs to save space so that should be fine.

Thanks for all your help!