Switching power supply with small to high loads

[Yaroooo]

I want to realize a circuit that allow to have a regulated voltage of 5v and be able to supply up to 1.3A .

I can't use a linear LDO cause a huge voltage drop on it, input may vary between 24 to 40v and with this current load isn't suitable.

On this board I've also a Micro that have a very small load current. Since my circuit may have only micro ON as well as all load ON (some LEDs in example), it should have an operating range between 1mA to 1.3A.

Looking at some datasheets, common switching regulators that allow to supply such current has good efficency above 300mA. Where i.e. LT8608 is too expensive.

Do you have advices or components usable to this porpuses? My idea was LM2501x series and TPS5414 , but they have efficency near to 0 at 1mA.

[madires]

There are tons of buck converter ICs to choose from, but you won't find any with high efficiency at very low and high loads. Anyhow, a low efficiency at a load of few mA doesn't matter much because the losses are small too.

[Yaroooo]

Anyhow, a low efficiency at a load of few mA doesn't matter much because the losses are small too.

Thanks for your answer.

Are there alternatives in circuits methods to avoid this? Double power supply or double regulators? I'm asking piratical examples in common products.

[SeanB]

If you atre doing a lot of LED's and they are going to draw a lot of power, ranging from all on to all off, then 2 separate power converters will work. Low power one to drive the microcontroller, and a separate one with an enable to do the LED drive voltage. When they are all off the power supply is switched off with the enable, otherwise you enable it. As an advantage you can choose LED drive voltage to have multiple series strings, if the LED's are always going to be on like that, and if so then you can also use LED driver converters to do both the drive and current control, and save on power converter and switches, using LED drivers with dimming and enable as needed.

[David Hess]

Switching regulators which support burst mode operation can have good efficiency at high and low loads.

I would consider using separate converters for the control and power circuits.

24 to 40 volts down to 5 volts is quite a step.  Consider using a boosted current buck converter with a tapped inductor to get a more reasonable duty cycle.

[Yaroooo]

What happen if I use a switched regulator at low efficiency (if efficiency is not a requirement)? In example, with a 10 mA load I get 20% of efficiency.

What this will mean? This 80% of loss will go to heat dissipation like a linear regulator?
Or back to power supply getting 100% power and generating 20% of it? So no heat on regulator?

Seems an expensive solution, for cheap boards, use 2 regulators one for a small micro (i.e. microchip) and one for power some LEDs. In example cheap light signals.

[madires]

Yes, the losses are converted into heat. In your example the output is 10mA * 5V = 50mW. If those 50mW are 20% then the 80% losses are 200mW. Let's assume the efficiency is 90% at 1A. So the output is 1A * 5V = 5W and the losses are 555mW. Unless your design goal is to minimize power consumption the low efficiency at low loads doesn't matter much.

[David Hess]

The other problem is that at low load currents, the converter will almost certainly transition from continuous conduction mode to discontinuous conduction mode which has different duty cycle and frequency compensation requirements.

[Yaroooo]

Thanks for your answers. In this application I can use a constant load since some LEDs are always on.

I've calculated about 160mA Minimum, 400mA Medium, 560-700mA Maximum. I'll choose a regulator that have at least 50% at 160mA and max output at 1A. So there'll never be a micro powered alone.

But I still looking for a solution on how to cheaply power a micro and a load with a voltage regulation from 24v to 5v. Any suggestion?