Bucks in series or parallel

doronraz

Joined Jun 7, 2022
8
Hi Friends,

I have theoretical question about step-down convertor.
If I have input voltage of 12V, on my board I need 2 types of voltages: 5V & 3V, each cycle will demand 0.5mA.
I have 2 options to generate them, in series 12V to 5V and then 5V to 3V.
The second option, in parallel: 12V to 5V and 12V to 3V.
Which is preferred and why?

If there are articles about it, please share with me, I am pretty new in this field.
Thank you.

MrAl

Joined Jun 17, 2014
9,549
Hi Friends,

I have theoretical question about step-down convertor.
If I have input voltage of 12V, on my board I need 2 types of voltages: 5V & 3V, each cycle will demand 0.5mA.
I have 2 options to generate them, in series 12V to 5V and then 5V to 3V.
The second option, in parallel: 12V to 5V and 12V to 3V.
Which is preferred and why?

If there are articles about it, please share with me, I am pretty new in this field.
Thank you.
That small current?

Ian0

Joined Aug 7, 2020
6,315
At 500uA a linear regulator will probably be more efficient.

BobTPH

Joined Jun 5, 2013
5,785
And for linear regulators, the power loss would be the same either way.

dl324

Joined Mar 30, 2015
15,117

By putting them in series, you force the 5V regulator to handle the load from the 3V regulator. If they're in parallel, you can size each regulator without regard to the load on the other.
If there are articles about it, please share with me, I am pretty new in this field.

Ian0

Joined Aug 7, 2020
6,315
It might just be that the output from the 5V regulator is quieter than the 12V supply, which might just improve the noise on the 3V output.

doronraz

Joined Jun 7, 2022
8
That small current?
At 500uA a linear regulator will probably be more efficient.
And for linear regulators, the power loss would be the same either way.
My mistake I meant 0.5A.

By putting them in series, you force the 5V regulator to handle the load from the 3V regulator. If they're in parallel, you can size each regulator without regard to the load on the other.
What about the switching losses of the controller? or other losses because the voltages dropping?
If I understand correctly, its wiser to design them in parallel...

dl324

Joined Mar 30, 2015
15,117
What about the switching losses of the controller? or other losses because the voltages dropping?
You're going to have switching losses in either case.

In switching regulators (which is what I assumed you meant when you said buck), you still have to be concerned about the maximum current allowed.

If, for example, you were using regulators that had a 1A current limitation and a 0.5A load on the the 5V and 3V regulators, putting the 3V regulator on the output of the 5V regulator wouldn't make sense because the 5V regulator would now be operating at its maximum current.

If you were using linear regulators, it could make sense to put them in series if power dissipation in the 3V regulator became an issue.

MrAl

Joined Jun 17, 2014
9,549
My mistake I meant 0.5A.

What about the switching losses of the controller? or other losses because the voltages dropping?
If I understand correctly, its wiser to design them in parallel...
Hi,

This depends highly on what bucks you are using. That's because efficiency varies widely with load and the ratio of input to output voltage.
The only way to get maximum efficiency is to try both ways and measure the efficiency of both and add them together.
If you can not do that, then you have to have the curves for efficiency vs load current and efficiency vs input output voltage ratio. Efficiency vs load current is the most significant usually. What that means is the higher voltage buck may have better efficiency when loaded more so connecting them in series would most likely result in higher efficiency.

BobTPH

Joined Jun 5, 2013
5,785
I cannot see how series could be more efficient for switching supplies.

Say you have two buck can converters that are 90% efficient. If in series m, the lower voltage supply would lose 10% from each converter, resulting in 81% efficiency. In parallel, it would be 90% efficient.

Ian0

Joined Aug 7, 2020
6,315
If you were smart about it, and only one needed accurate regulation, it could be done by one converter with two windings.

MrAl

Joined Jun 17, 2014
9,549
I cannot see how series could be more efficient for switching supplies.

Say you have two buck can converters that are 90% efficient. If in series m, the lower voltage supply would lose 10% from each converter, resulting in 81% efficiency. In parallel, it would be 90% efficient.

Hello there,

You would be right if converters had one set efficiency, but the efficiency varies with load.
Also, i did not say they HAD to be wired in series, ONLY if the connection results in better efficiency.
For example, say the 12v unit had 90 percent efficiency at 2 amps. It may only be 70 percent efficient at 0.5 amps. Now if the lower voltage buck connected to that output and drew 1.5 amps, the 12v unit would be back up to 90 percent efficient.

But this is highly dependent on the exact nature of the bucks. If the top one has 90 percent at 0.5 amps and the bottom one has 90 percent at 0.5 amps and the load for both is 0.5 amps each, then they are better in parallel.

I thought i made it clear it depends on the bucks but i guess not. That's why i suggested testing as the best way to be sure.