Actions
Forum Thread Details
  • State Not Answered
  • Replies 21 replies
  • Subscribers 2561 subscribers
  • Views 2669 views
  • Users 0 members are here
Related

Difficulty getting good ripple / noise measurements

Fred27

Im currently roadtesting the Rohde & Schwarz NGC103 and as part of my investigation I'm trying to measure the ripple / noise and compare it to other supplies. The problem I'm having is that I don't seem to be able to get a reasonable measurement. I get the feeling that isn't swamped by external interference.

I watched a helpful EEVBlog video and tried to optimise my setup but I don't seem to be able to get close to any meaningful measurement. In fact, there's not much different if I just connect the two banana plugs together and measure the noise on my 0V there!

Here's a photo of the best setup I've managed so far. I tried differential measurement with one probe on + and one on - and measuring the difference. Whilst I could see the common mode noise on both channes that were being cancelled out, the result was still noisier than this. I have AC coupling, a 20Mhz bandwidth limit and have tried via a 10:1 probe, a 1:1 probe, and probeless direct via banana to BNC as pictured.

Noise measurement setup

Noise measurement screenshot

Any suggestions? In fact, is measuring noise in a setup like this even useful? Would you just trust the datasheet on this? For reference, the datasheet says < 6mV peak-to-peak and < 1mV RMS. I'm getting over an order of magniture more.

Parents
  • 0

    R&S have a good reputation so I would expect their PSU to meet its spec.

    The way I measure noise on embedded PSUs is to solder an SMA connector to the PSU and use an SMA to BNC lead to connect to the scope.

    Set up your scope for 1M AC coupled.

    Step 1 is a basic sanity check, look at the noise with just the lead connected. Then make a short SMA to 50R 0805 or 0603 resistor lead.

    In some environments the open lead will be pretty quiet but your power supply is a low impedance source so the 50R load is fair. If you can't get  a good noise measurement with just the 50R terminated cable then your scope isn't up to the job.

    You will get less noise if you set the scope for 50R input impedance - but be very sure you are AC coupled and do not overdrive the scope input.

    If that's OK then keep the 50R set up but connect the earth side of the 50R to the DUT 0V. Nothing should change but often things will. If you get a lot of excess noise with this setup you can try filters in the mains leads of scope and/or DUT but it usually means that one or both devices are injecting noise into the mains. Historically it was common to remove the mains earth connection from the scope but this is very much not approved of now. You can try an isolation transformer but it may well make things worse.

    Assuming you get past this hurdle OK, remove the 50R.

    Since you can't solder and SMA connector to your PSU then make the shortest adapter lead that you can.

    Start off by measuring the noise with the PSU plugged in but turned off.

    Move on to, turned on with no load.

    Then finally add the load (use a fixed power resistor for noise measurements - electronic loads (even expensive ones) make their own noise.)

    Looking at the spectrum of the noise may help in down the source.

    In the past I have resorted to connecting the scope signal earth to earth on the DUT directly with copper braiding.

    In difficult cases there are lots of other things to consider and tune - I'll be interested to hear how you get on.

    MK

  • 0 in reply to Andrew J

    Some of it will be covered in my Road Test on the Vishay Sic967 which is due within a couple of weeks but I'll think about a dedicated blog for later in February if I get time, but I have a lot of stuff on the go right now.

    MK

  • 0 in reply to michaelkellett

    We used to use ferrite rings with several turns of the cable coiled through them to prevent external noise sources getting in the way. It has to be the right grade of ferrite though, and after all this time I can't remember what it was.

    I've just looked at the 'scope traces again. Is it possible that something is picking up the signals from a mobile phone? When heard on an AF amplifier they sound a bit like Morse code and are present when the phone is powered (charging?) and being used or not.

Reply
  • 0 in reply to michaelkellett

    We used to use ferrite rings with several turns of the cable coiled through them to prevent external noise sources getting in the way. It has to be the right grade of ferrite though, and after all this time I can't remember what it was.

    I've just looked at the 'scope traces again. Is it possible that something is picking up the signals from a mobile phone? When heard on an AF amplifier they sound a bit like Morse code and are present when the phone is powered (charging?) and being used or not.

Children
No Data