Power Tips: How to be discrete


One of the most common complaints of a power supply designer is that power is always an after-thought during the design process.  It seems that most of a system’s design schedule is vanished when someone finally says, “Oh, and we also need a power supply.” 

For telecom applications, power supply brick modules are great in these situations.  You can pretty much just plop them down on your board and they work.  Not much design effort is required.

Many times, the first pass design of a product will have a budget large enough to absorb the cost of a module.  However, after the product gets into production, the volume starts to increase.  Next, the corporate bean counters will start to look at the profit margin and ask for a new cost reduced version of the product.   One of the first places to look is at the power supply.

A discrete design can easily reduce the component costs of the power supply by well over 50%.  Of course, you also need to consider the non-recurring development effort that is required and the assembly costs of many components versus just one module.  In large volume products with an adequate amount of development time, it can make a lot of sense to use a power supply made from discrete components.

Related article: Telecom Power Systems: Why be discrete?

This is where you can really leverage the resources in PowerLab.  A moderate-level knowledge of power supply design is required, but the bulk of the design work is already complete.  This significantly lowers the development costs and risk. 

Some of the PowerLab designs are even designed to be footprint compatible with the industry standard brick footprints.  This makes it easy to drop a prototype of the power supply right into an existing system and compare the performance versus the module.  Most of the module replacement designs in PowerLab were designed to meet or exceed the performance specifications of the modules.

PMP7451 discrete power supply

Example one discrete design: PMP7451

Of course, there are reasons other than time constraints to use modules.  Modules are still lower risk and have a size advantage.  Modules makers invest a lot of time and money into packaging the most power into the smallest size and you see that in the price.  That being said, many systems don’t require the absolute smallest size, or maybe the space available for the power supply is not a perfect 0.9” by 2.3” rectangle.  

So, for those of you that are under pressure to cost reduce your telecom power supply and for those of you that are just curious, I would like to share just a few of my favorite brick replacement designs...

  • PMP8861 – 18-72V Input, 12V/10A Quarter Brick
  • PMP7451 – 36-75V Input, 12V/10A 8th Brick
  • PMP6886 – 36-72V Input, 2.5V/5A 16th Brick
  • PMP5364 – 36-72V Input, 5V/10A Quarter Brick

In summary: Use modules when time and development resources are limited; go discrete when cost is king.

Do you prefer to be discrete?

Related blogs:

  • Brian:

    I cannot tell you how true your comments are. When I worked for HP back in the day, the brand new engineer right out of college got the job of designing the power supply for the project. The guy (back in those days) had just finished some great cpu design and was gung ho to continue this. You can imagine the enthusiasm for power supply design.

    This was pretty much standard though out the company until a friend of mine working in Colorado back in the day grew sick of the power supply disasters that happened all to frequently and simply took a power supply from HP's power supply division and stuck it in a piece of equipment and said "There, job done".

    From then on the job of power supply design went to HP's power supply division who made fabulous power supplies and even enjoyed doing it.

    Power supplies aren't always sexy, but when they crap out then the whole product looks bad.

    Cheers.

    Glenn.

  • Thanks for your comments, Glenn.  It seems that everyone at some point has had similar experiences with power.  Power is like that inevitable task that everyone knows needs to be done, but often gets put off until the end of a project under the assumption that it is quick and easy.  One of the ideas behind our PowerLab library is that it might help non-power experts out of a jam in which they might find themselves, or better yet, it might help them plan their power strategy earlier in the project.  

    And designing power supplies actually can be fun... everything is not 1's and 0's.

    thanks,

    Brian