230501 TI - Engineer It - How to test power supplies

Hi, I'm Bob Hanrahan, Application Engineering at Texas Instruments.
This is the first on a series on testing power supplies.
We'll be talking about some of the basic tests, not all, but the basic tests needed to ensure a reliable power supply.
Now, testing a power supply is critical.
Today, many of us use simulation tools, such as TI's WEBENCH, to go from a power supply requirement
all the way through even to a working prototype.
But a working prototype doesn't mean it's a reliable, fully operable system.
Only real bench testing can verify that.
So problems such as noise issues or stability issues,
As an example, might not be severe enough to cause a failure on your prototype, but could cause problems later on.
So we'll be delving into subjects such as just that, the stability,
we'll be talking about noise, and we'll be measuring efficiency of a power supply throughout the series.
Before you start any testing, you want to start with a test specification.
So that would be all the different minimums and maximums that your system can tolerate.
With that, create a test plan and execute to that plan.
Also, you'll be needing equipment.
And let's refer to the graphic here to show you the type of equipment you'll be needing.
Of course, on the input, you'll need a DC power supply that can provide the voltage and current necessary for your specific system.
And on the output, you'll be needing some type of a load,
preferably an electronic load or dynamic load to provide the ability to perform transient load testing.
You'll be needing at least two volt meters, one for the input and one for the output.
You'll be needing two current meters as well, one in series with the input and one on the output.
If your electronic load has the ability to measure, accurately measure the current,
then you could use your load box.
You also need an oscilloscope for noise measurements, capable of 500 megahertz or above,
and a network analyzer, a low frequency network analyzer,or a frequency response analyzer,
as they're often called, is necessary for your traditional stability measurements.
Now also, you will be needing to set up your system to allow access for power testing.
So I'm going to refer down to one of our evaluation boards shown down here.
This is our LM5117 evaluation board.
That's a buck converter, wide input range buck converter, that was just appropriate for this demonstration.
You'll notice what I did.
I put very short but high gauge wires on the input,
appropriately marked with the colors so you don't mix them up, and on the output.
So if your power supply is directly connected to your system,
you'll want to break that connection and provide wires so you can get in there with your dynamic load.
We'll connect them later during that test sequence.
Also within this design is a resistor that's added in for the stability measurement.
And I'll show on a graphic where that resistor goes.
So shown here is your typical power supply regulator circuit.
Could be a linear or a switch mode power supply.
It monitors the output, brings it through.
Usually it's a voltage divider back into the input, which goes into the error amplifier within the regulator.
So you want to add an additional resistor above the feedback network in the order of 10 to 50 ohms.
And that will provide an injection point for the network analyzer for accurate stability measurement.
So in summary, simulation is great for expediting a power supply design,
but it's never going to give you all the results that you need to verify your design is running properly and reliably.
So for more information or videos on actual bench testing, go to the following web addresses.
Thanks for watching.

526互联

230501 TI - Engineer It - How to test power supplies - Overview