Can I get 5V 2.5A from a 12V 1.5A power brick?

[technix]

The power brick is rated for 12V 1.5A, or 18W. The 5V 2.5A (12.5W) is intended for a NeoPixel LED strip, in its a worst case scenario (maximum brightness on all LEDs.)

Can I get 5V 2.5A from that power brick using a single-phase 3A-rated buck converter (e.g. TPS563201 and an appropriately-sized inductor,) or can I use a two-phase converter with two 2A output phases (e.g. MP2122 with its two output phases in parallel,) or do I need a bigger power brick?

[metrologist]

You would look at the conversion efficiency and power capability of the buck converter.

You need 70% conversion efficiency with 2.5A continuous duty, minimum.

[technix]

You would look at the conversion efficiency and power capability of the buck converter.

You need 70% conversion efficiency with 2.5A continuous duty, minimum.

Both chips are no less than 80% efficient at all loads. With that in mind, will a single-phase buck converter work, or do I need a two-phase one?

[metrologist]

Either type should work as long as it can handle the load.

One of your chips is rated 2A, so not enough. Also:
Continuous Power Dissipation (TA = +25°C) (2) ................................................................. 1.25W

If you are running 12.5W out at 90% efficiency, that is about 14W in and 1.4W dissipated.

That chip will not last.


Here is some info about multiple phase converters:

http://www.ti.com/lit/an/slyt449/slyt449.pdf

https://www.electronicdesign.com/power/supply-meets-demand-use-multiphase-converters-solve-tough-power-design-challenges

[james_s]

I would expect this to be fine. It should tolerate being run full brightness continuously even, but in practice you will likely have plenty of margin.

[technix]

Either type should work as long as it can handle the load.

One of your chips is rated 2A, so not enough. Also:
Continuous Power Dissipation (TA = +25°C) (2) ................................................................. 1.25W

If you are running 12.5W out at 90% efficiency, that is about 14W in and 1.4W dissipated.

That chip will not last.


Here is some info about multiple phase converters:

http://www.ti.com/lit/an/slyt449/slyt449.pdf

https://www.electronicdesign.com/power/supply-meets-demand-use-multiphase-converters-solve-tough-power-design-challenges

The 2A chip is rated 2A each phase. Should I be using that chip it would be two phases in parallel for a possible 4A maximum current.

[metrologist]

I would expect this to be fine. It should tolerate being run full brightness continuously even, but in practice you will likely have plenty of margin.

Well, I didn't really look at everything, but he wants to power from a 12V brick? Will that still work?

Wide 2.7V-to-6V Operating Input Range

If one channel is 2A and he wants 2.5A, can he use those channels in parallel?

[ogden]

Multiphase converter would be overkill for 5V 2.5A. Just properly size your synchronous converter IC. Obviously as already noted, sot23 is way too small

TPS54427 seems to be matching your requirements. Do not forget to check board layout & docs of evaluation module, SLVU579.

[james_s]

I bought a bag of those KIS3R33 modules some years back that I use for this sort of thing. I don't know what they came from but there seemed to be huge quantities of them surplus in China, pulls from something or other. There are loads of other buck regulator modules out there that will do the same thing if you don't feel like designing something yourself.

[Zero999]

Why not simply use a 5V 2.5A brick? That's got to be the easiest solution.

Is there a requirement for 12V? If so, what current? Then the question is whether it's better to use a 5V brick and boost to 12V.

If only 5V 2.5A is required, then use a 5V brick. Converting 12V to 5V, when it's not necessary is bad design.

[mikeselectricstuff]

If only 5V 2.5A is required, then use a 5V brick. Converting 12V to 5V, when it's not necessary is bad design.

Unless you have any significant cable length. Local conversion at the load end both reduces losses and gives you tolerance of voltage drop.

[Zero999]

If only 5V 2.5A is required, then use a 5V brick. Converting 12V to 5V, when it's not necessary is bad design.

Unless you have any significant cable length. Local conversion at the load end both reduces losses and gives you tolerance of voltage drop.

That's true, but wouldn't it make sense to use an even higher voltage in that case?

I suppose 12V might be more optimal, because a simple buck converter can convert 12V to 5V more efficiently, than 24V to 5V.

[james_s]

It would also make sense if you happen to have some nice 12V power supplies already, or if something else you also want to power requires 12V.

[Zero999]

It would also make sense if you happen to have some nice 12V power supplies already

That's what I was thinking, but would disagree with it being a good reason for this. It will probably not cost much to buy the correct power supply and save the 12V ones for another project. In the past I've made some poor design choices because I resorted to using what's available, rather than being more patient and ordering the correct part. Fair enough, the part is expensive or takes too long to get, then it makes sense, otherwise it's generally better to wait and spend a little.

or if something else you also want to power requires 12V.

That makes more sense, but if the 12V device is small, compared to the 5V device, it begins to make more sense to use a 5V PSU and a boost converter. Of course this is assuming the cable length is short, otherwise it's better to transmit the power at a higher voltage, though a long cable, as previously mentioned.

[technix]

Why not simply use a 5V 2.5A brick? That's got to be the easiest solution.

Is there a requirement for 12V? If so, what current? Then the question is whether it's better to use a 5V brick and boost to 12V.

If only 5V 2.5A is required, then use a 5V brick. Converting 12V to 5V, when it's not necessary is bad design.

It would also make sense if you happen to have some nice 12V power supplies already, ...

I just happen to have a bunch of 12V 1.5A and 12V 2A power bricks in spare.

Here in Shanghai those are handed out as part of standard telecommunication hardware from China Telecom, but those field service engineers just never bothered to collect them when returning devices for upgrades or repairs, allowing them to accumulate here.

... or if something else you also want to power requires 12V.

There is nothing that requires 12V in this LED strip.

Unless you have any significant cable length. Local conversion at the load end both reduces losses and gives you tolerance of voltage drop.

Is 1.5 meters to 2 meters a significant cable length? The copper might be thin though, being Chinese-made el cheapos.

[Mr. Scram]

You could measure the wire resistance and calculate the expected drop.

[digsys]

Unless you want to go to the trouble of rolling your own, something like the TI PTN78060W is perfect. 3A, high efficiency.
Actually designing them in right now :-) into a few projects. Just add a couple caps and a Vset resistor.

[Zero999]

You'll probably find building a 12V to 5V converter will cost as much or more, than the correct power supply in the first place. You might be able to save a tiny amount by purchasing a 12V to 5V ready made module, but that's just penny pinching.

[technix]

Unless you want to go to the trouble of rolling your own, something like the TI PTN78060W is perfect. 3A, high efficiency.
Actually designing them in right now :-) into a few projects. Just add a couple caps and a Vset resistor.

Currently it is choosing between a single-phase SGM6130 based design or go with a M1VRv2-2DX2 software-defined multi-phase converter design.

[technix]

You'll probably find building a 12V to 5V converter will cost as much or more, than the correct power supply in the first place. You might be able to save a tiny amount by purchasing a 12V to 5V ready made module, but that's just penny pinching.

I need to build a MCU board for the NeoPixel strip anyway, so this power module is simply piggybacking. No significant cost occurred with or without that power module.

[mikeselectricstuff]

If only 5V 2.5A is required, then use a 5V brick. Converting 12V to 5V, when it's not necessary is bad design.

Unless you have any significant cable length. Local conversion at the load end both reduces losses and gives you tolerance of voltage drop.

That's true, but wouldn't it make sense to use an even higher voltage in that case?

I suppose 12V might be more optimal, because a simple buck converter can convert 12V to 5V more efficiently, than 24V to 5V.

Yes - 24V would be my default choice, and there are plenty of converters that would do it. However for intermediate cable lengths, it may be that local regulation is the primary concern rather than efficiency, and 12V may be easier to get - e.g. plug-in and desktop type PSUs are less common at 24v, though 19v laptop supplies are another common option.
Once you go over 24v, thinks like circuit protection and PSU availability become more of an issue, so you'd only go that way if you had a particular need for it.