Dismiss Notice
Join Physics Forums Today!
The friendliest, high quality science and math community on the planet! Everyone who loves science is here!

Creating oscillations on the output of a DC power supply

  1. Jul 4, 2018 #1
    Is there some common parameter for DC power supplies that provides the maximum oscillation amplitude allowable at the output?

    For some context - I would like to generate an oscillating signal powered by a DC power supply, but to prevent most of the oscillations reaching the DC supply I am going to implement a radio frequency choke (RFC). I need to know how large the RFC must be to satisfy the power supplies maximum allowable oscillations at it's output.

    Thanks in advance!
     
  2. jcsd
  3. Jul 4, 2018 #2
    Usually it's better to view this the other way: your 'filter' (usually a well sized capacitor bank close to the power parts) is not there to protect your PSU but to provide local (low impedance) 'backup' for the circuit at high frequency.

    It is already bad enough if you have RF running on the power cords.
     
  4. Jul 4, 2018 #3

    anorlunda

    Staff: Mentor

    The answer depends on the details. Can you tell us more? What are you trying to accomplish? What voltages, powers, frequencies are you working with? Maximum instantaneous power or max average power?
     
  5. Jul 4, 2018 #4
    I'm unsure on the specifics but I was more just hoping someone knew the name of the parameter that details the maximum allowable oscillations at the output. I guess the answer to the following question will help me understand:

    why is it common to block AC frequencies from reaching the output of a DC power supply?
     
    Last edited: Jul 4, 2018
  6. Jul 4, 2018 #5
  7. Jul 5, 2018 #6
    It is common to 'block' that AC as close to its source as possible, so it won't affect the DC power rail across the whole circuit (the power rail is not only about connecting something to the PSU: it is also about connecting everything together). When there is RF on the PSU what requires external/additional choking you are most likely already failed with the design process: since most circuits are far more sensitive than the PSU, when you have to start protecting the PSU the main stuff is usually a complete mess.

    Ripple
    Decoupling
     
    Last edited: Jul 5, 2018
  8. Jul 5, 2018 #7

    berkeman

    User Avatar

    Staff: Mentor

    Glad you found that reference. I was just about to post something similar. Testing for the stability of switching power supplies when the load is varying is an important part of characterizing their regions of stability (and regions of potential instabilities).

    BTW, can you say more about the application? It's kind of unnatural to impose a signal on a low-impedance power supply rail. Is this some sort of "Link Power" communication network, like "Power Over Ethernet" (POE)? There are some tricks to such applications that can help to keep the costs down, compared to brute-force things like large blocking filters...
     
  9. Jul 6, 2018 #8

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    It's good practice to reduce RF Interference as much as possible. There may be other equipment nearby and also there are regulations about RFI levels.
    It's standard to have decoupling capacitors between the supply rails and ground in order to make sure that the circuit works 'as planned'. A signal path along the internal supply rail in a circuit can be a severe embarrassment, sometimes causing an Amplifier to Oscillate or an Oscillator just to Amplify.
     
  10. Jul 6, 2018 #9

    Tom.G

    User Avatar
    Science Advisor

    Or a digital logic circuit to operate illogically!
     
  11. Jul 6, 2018 #10

    berkeman

    User Avatar

    Staff: Mentor

  12. Jul 7, 2018 #11

    sophiecentaur

    User Avatar
    Science Advisor
    Gold Member
    2017 Award

    Yes.
    When you get down to it, every 'digital' circuit is analogue. It is only the information that's digital. It's a waste of capacity to have the waveforms (internal as well as external) looking like the box-cars that we are presented with when digits first come our way. Timing is what it's all about in high speed digital circuitry and a sniff of something on a supply rail can cause deadly intersymbol interference.
     
  13. Jul 12, 2018 #12
    You really haven't described the problem - and so I wouldn't presume you have a solution yet.
    Exactly why do you care about ripple on the power line? What other devices will be impacted by this ripple?
    The answers to those question will define you requirements - and from there you can develop a design.

    As @sophiecentaur noted, ripples on you power line can be a problem to devices that are not connected to the power line at all. You will have both "Conduction Emissions" and "Radiated Emissions". The Conducted Emissions could be a problem for susceptible devices sharing the power line. The Radiated Emissions could be a problem for nearby susceptible devices - WiFi, radio receivers, etc.
     
Share this great discussion with others via Reddit, Google+, Twitter, or Facebook

Have something to add?
Draft saved Draft deleted