Need help measuring noise power

[Speakerguy]

Hi guys,

I have to measure a DC/DC converter at work and make sure it's noise power is under a certain spec. Bandwidth is 10Khz to 2Mhz, and it has to be something like -100dBm/Hz. Is there a simple way of doing this with a scope or a piece of test gear my company could rent to check this? Any help would be appreciated. Thanks!

 

[bountyhunter]

You need to find out what EMI spec they are trying to meet. The spec gives a plot of noise EMI versus frequency that the system has to stay below. The envelope of allowable noise gradually gets lower as frequency goes up. I have done EMI compliance a few times on our power supplies, we always hired specialists because the equipment is very specialized and expensive.

 

[Warpspeed]

Agree with Bounty Hunter.

You need to get a copy of the specification that needs to be met.
More often than not it will include details of exactly how the measurement is to be performed.

There are a variety of ways to define and measure noise depending upon the application, and applications vary hugely.

 

[bountyhunter]

There are specific specs called out which will define the test methods. We had specs like MIL-STD-883 and others, and they give the specific test setup to take the data. You always have to get it straight with the customer exactly what they are requiring.

 

[RCinFLA]

The piece of equipment you will need is a spectrum analyzer.

You will likely need at least a 1 GHz analyzer but depending on what spec you are required to meet it might extend higher since there are a lot of mobile phones operating in 2 Ghz frequency range.

There are radiated specs, requiring an antenna site to verify, and conducted specs where power cables and other cables are directly connected to spectrum analyzer.

There are service agencies that will make the measurements for you and certify the results.

 

[Warpspeed]

Maybe not if the noise measurement is conducted noise into the dc power source and the dc load, and not radiated (into the surroundings) radio frequency noise.

This is why the specification is so very important. Without the specification, there is no real way of knowing what is actually to be measured, or how to go about measuring it.