12V, 100A SMPS with ten interleaved 12V, 10A buck converters.

Status
Not open for further replies.
T

treez

Guest
Hello,

Do you think that a 12V, 100A SMPS is best done with ten 12V, 10A bucks (50KHz Fsw each) or twenty 12V, 5A bucks (25KHz Fsw each)?

Input voltage = 24V.

I wish to do phase interleaved switching so each buck is phased equally apart from the next one, so to speak.
 
Last edited by a moderator:

In principle there's no limit to how few or many phases you can put in parallel, as long as you force current sharing properly. Your choice should take into account the required size, ripple, efficiency, and cost of the design. More phases can lead to higher efficiency, smaller size, and lower ripple, but at greater cost.
 
Reactions: KonstanU and treez

    T

    Points: 2
    Helpful Answer Positive Rating

    KonstanU

    Points: 2
    Helpful Answer Positive Rating
Interleaving helps reduce the input ripple (since in a buck is discontinuous current) and reduce the output filters requirements.

But the key is close current sharing, and you need a load controller such as the UC2907 to achieve that.

You also need to have the means to isolate (disconnect) each individual unit, such that if one fails, it won't take down the whole system.
Therefore......
I would guess that a smaller number of individual SMPS units would be preferred from the cost and size stand point.
 
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
But the key is close current sharing, and you need a load controller such as the UC2907 to achieve that.

Load share controllers are surely not necessary for interleaved SMPS's?

Load share controllers are for when you have multiple separate smps's each in parallel, supplying the same load?
 

This is a recent thread about interleaved buck converters.

https://www.edaboard.com/threads/285741/

It shows my simulation of 4 interleaved buck converters. Equally spaced control signals. Equal duty cycles.
I made the last coil a much smaller value than the others.
Result: In CCM, current does not differ by very much all round.
However in DCM the smaller Henry value ramps up current more quickly, which can create problems with saturation and/or overheating.
 
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Hi Treez
In addition to the Bradtherad's statement : your power rating is not so high it is just 1.2Kva . it can easily be harvested by a single stage too . and also it can be obtained by an H bridge or push pull too .
But if you like to use some power supplies ( some buck convertors ) in parallel together then you need to know that you must take an specific feedback which has been called master/slave feedback . it's duty is try to keep all of the modules exactly the same together . if voltage of one of the PSU s be lower than the other one then the power supply module which has lower voltage in compare with the other ones , will be considered as the load for the other ones ! t is one of the biggest problems that many of the designers don't pay enough attention to it ( unfortunately of course ! )

Best Wishes
Goldsmith
 
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
You should have at least two phases because input current ripple ideally cancels out for 2:1 voltage ratio. Dual phase controllers are pretty standard.
 
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
Is the attached schematic OK?

It shows three bucks interleaved and assumes that the same error signal can be shipped to each controller (LT1243, which is pin for pin compatible with UC3843)
 

Attachments

  • Schematic ..Interleaved Bucks.pdf
    42.2 KB · Views: 213

Is the attached schematic OK?

It shows three bucks interleaved and assumes that the same error signal can be shipped to each controller (LT1243, which is pin for pin compatible with UC3843)

Is there a way to stagger the on-off sequences of the converters? So they can never get into sync and cause large current spikes to (a) be drawn from your power source, and (b) to go back and forth through the smoothing capacitor(s)?
 
Reactions: treez

    T

    Points: 2
    Helpful Answer Positive Rating
they are all switching 120 degs apart...as you can see from the phased voltage pulse sources at their RT pins
 

Status
Not open for further replies.

Similar threads

Cookies are required to use this site. You must accept them to continue using the site. Learn more…