Connect with us

Stopping RFI from a PC

Discussion in 'Electronic Basics' started by W. Watson, Oct 18, 2005.

Scroll to continue with content
  1. W. Watson

    W. Watson Guest

    I replaced a PSU in my PC, and the result is that the new PSU is more
    reliable; however, the RFI seems to have dramatically increased. The PC is
    in a building about 100' from the house and 75' from the garage. If I turn
    my car AM radio on in the garage, the interference is quite noticeable. Same
    with aa AM radio in the bedroom. Since the PC is already in a metal chassis,
    what can I do to reduce the interference? Of course, turning off the PC
    helps a good deal, but the PC needs to be in operation 7/24.
    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet

    "... astronomy is useful because it shows [not only] how small
    our bodies are, but how large our minds are." -- Henri Poincare

    Web Page: <home.earthlink.net/~mtnviews>
     
  2. ehsjr

    ehsjr Guest

    A couple of things come to mind that might help:
    Check to make sure the ground at the receptacle is good.
    Also, wrap the power cord through a ferrite choke,
    as many turns as you can fit. An old TV yoke might
    be a source for a large enough ferrite core.

    Ed
     
  3. If you mount a die cast box on the side of the computer and run the
    power cord through it, with a good EMI filter inside, you should see
    significant improvement, assuming that the noise is getting out
    through that path.

    A good filter might be something like:
    http://rocky.digikey.com/WebLib/CII-Corcom/Web Data/Q Series.pdf

    Pick one with a generous current rating (say, double the expected RMS
    current to the supply), to keep the inductors from saturating on the
    peak current pulses through the line rectifiers.
     
  4. W. Watson

    W. Watson Guest

    Are you thinking of the AC receptacle for the PC?

    Note that one of the radios could be considered a portable. The one in the
    car. The noise in the house radio is still there when I unplug the radio and
    let it operate from batteries. It seems odd that the PC would radiate
    through the metal chassis. Of course, it might cause radiation through the
    circuit from the outside building into the house. They are all connected.

    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet

    "... astronomy is useful because it shows [not only] how small
    our bodies are, but how large our minds are." -- Henri Poincare

    Web Page: <home.earthlink.net/~mtnviews>
     
  5. W. Watson

    W. Watson Guest

    Ah, I see now the respondent above you was driving at the PC and not the the
    locations where the radios are, since you clarified where to put the filter.

    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet

    "... astronomy is useful because it shows [not only] how small
    our bodies are, but how large our minds are." -- Henri Poincare

    Web Page: <home.earthlink.net/~mtnviews>
     
  6. w_tom

    w_tom Guest

    Sounds like you did not spend $65 retail for that supply.
    To discount inferior supplies, important functions get
    forgotten. Sounds like the AC line filter that must be in
    every minimally acceptable supply was forgotten to cut costs.
    You would be buying expensive line filters because the line
    filter did not come in that supply - as is required to meet
    FCC regulations.

    What other critical functions are also missing from that
    supply? Functions discovered later after damage has occurred.
     
  7. Jasen Betts

    Jasen Betts Guest

    As a rule PC motherboards radiate most at much higher frequencies than
    AM radio, and since the change made was the power supply it's most likely
    coming from that.
    the wiring could be acting as an antenna or a transission line...
    do all you can to keep that interferance inside the PC'S box.

    Bye.
    Jasen
     
  8. W. Watson

    W. Watson Guest

    Well, almost. It cost me $49. It's a CE 350 watt PSU, CODEGEN version 3.02.
    The previous PSU was a CODEGEN 350 watt.

    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet

    "... astronomy is useful because it shows [not only] how small
    our bodies are, but how large our minds are." -- Henri Poincare

    Web Page: <home.earthlink.net/~mtnviews>
     
  9. Guest

    Does it say if it got PFC? (Power Factor Correction)

    Make sure there is a line filter. Like those with two inductors in serial
    to the mains.

    Check the front of the PC box for RFI leaks. It's usually plastic fantasic
    all the way.. And above all make sure grounding is appropiate. Use a multimeter
    to verify it's working as it should.
    Should all other approches fail. Build/buy a sufficient steelbox. A cheap
    approach is a wellpap box with tinfoil on the inside.
     
  10. W. Watson

    W. Watson Guest

    I'm checking on the pfc. The version # above is incorrect. It's 2.03.

    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet

    "... astronomy is useful because it shows [not only] how small
    our bodies are, but how large our minds are." -- Henri Poincare

    Web Page: <home.earthlink.net/~mtnviews>
     
  11. Jamie

    Jamie Guest

    get an RFI line filter for the AC cord.
    when grounding the case, try to use shielded
    ground, that is a piece of coax using the
    center as the actual earth ground and connecting
    the shield to the ground also at the ground end only,
    but not at the computer end..simply pull it back a
    bit so that it does not touch and tape it up.
    get some Toriodal cores and wrap the cords through
    them also.
    all of this helps top keep the RF using the cords
    as antennas.
     
  12. W. Watson

    W. Watson Guest

    Could this difficulty be caused by damage to the chassis? When I bought the
    chassis, the owner sold it at a low price, since it had been dropped off a
    table. The frame was slightly bent but easily fixed. A year later I began to
    suspect that the PSU had also got damaged. Until I put in a new PSU, the RFI
    problem was not noticeable.

    Is it possible that in putting in the new unit that something didn't get
    grounded properly.

    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet

    "... astronomy is useful because it shows [not only] how small
    our bodies are, but how large our minds are." -- Henri Poincare

    Web Page: <home.earthlink.net/~mtnviews>
     
  13. w_tom

    w_tom Guest

    If RFI noise is that strong, then grounding is irrelevant.
    Follow ground wire from computer, through walls, to breaker
    box. That long wire an antenna for RFI. So what has grounding
    accomplished? Nothing. Your problem exists at the source.
    From your own provided information, the power supply was
    defective when purchased. Price alone suggests that.

    However, did power supply vendor provide a long list of
    numerical specs - in writing - including things such as:
    Specification compliance: ATX 2.03 & ATX12V v1.1
    Short circuit protection on all outputs
    Over voltage protection
    Over power protection
    PFC harmonics compliance: EN61000-3-2 + A1 + A2
    EMI/RFI compliance: CE, CISPR22 & FCC part 15 class B
    Safety compliance: VDE, TUV, D, N, S, Fi, UL, C-UL & CB
    Hold up time, full load: 16ms. typical
    Dielectric withstand, input to frame/ground: 1800VAC, 1sec.
    Dielectric withstand, input to output: 1800VAC, 1sec.
    Ripple/noise: 1%
    MTBF, full load @ 25°C amb.: >100k hrs
    If not, then again, you all but know that power supply does
    not contain many 'essential' functions. Does it even claim -
    in writing - to meet those FCC requirements? Quote that text
    exactly.

    To solve this, then you did everything that pbdelete
    suggested. You never post again until you have opened up that
    power supply and verified those line filter coils exist. This
    is not a game of selecting only what is easiest or
    convenient. This is a game where you do everything and then
    post back. He did not waste time posting those actions. You
    want to solve the problem. Then do everything he has
    recommended.

    The power supply did not cost $65 full retail. Therefore it
    is most likely defective when designed. Again, this is not a
    number you can fudge as you 'feel'. That was a blunt hard
    number. What is the full retail price. Less than $65? Then
    is it likely missing essential functions. What a
    coincidence. Your symptoms are exactly what would happen if
    that line filter (as made necessary by FCC Part 15) did not
    exist. Common for power supplies marketed to bean counter
    types who are experts because they look at two numbers -
    dollars and watts. Such supplies are not marketed to those
    who deal in reality - demand those numerical specs in writing.

    Well you have this supply. Does it have the line filter as
    pbdelete said to inspect for? If not, then every other post -
    especially those grounding suggestions - is time wasted.
     
  14. W. Watson

    W. Watson Guest

    As I think I mentioned above, in another part of this thread, the vendor has
    not responded yet. There is no list. I'm waiting for a reply.

    If you are convinced that the power supply is at fault, then I can simplify
    matters by taking it back. However, you have not explained why the original
    power supply, which is virtually the same as the one that's giving me the
    problem, did not have this problem. It certainly had a problem, but not this
    one. It functioned quite well for many months, and then became erratic. It
    never caused this much RFI.
    Ah, I missed part of his message. I saw the part about PFC and put that to
    the venodor. As said above, no answer yet. I called the store where I bought
    it and they couldn't tell me.

    If you don't mind, I'll just wait for the vendor to chime in before
    responding to pbdelete's:

    "Check the front of the PC box for RFI leaks. It's usually plastic fantasic
    all the way.. And above all make sure grounding is appropiate. Use a
    multimeter to verify it's working as it should. Should all other approches
    fail. Build/buy a sufficient steelbox. A cheap approach is a wellpap box
    with tinfoil on the inside."

    While we are waiting for that magic moment, perhaps you, or someone, can
    tell me: 1. how I'm going to check for RFI leaks 2. make sure grounding is
    appropriate, 3. verify it's working as it should? A wellpap box? Huh? Build
    or buy a steel box? No thanks. I either spend money for a better PSU or
    exchange this for one that does work. I have no intention of turning this
    into an electronic project. If you plan to respond to these questions,
    because I really do not intend to follow them. I would like to know though
    what a wellpap box is?

    Continued below.
    No comment.

    Well, while I'm waiting to get the answers above, perhaps you can give me a
    clue about the following. I plugged the previous PSU into the wall with the
    computer turned off. No RFI. I presume that's because the PSU needs a source
    to actually power up, right?


    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet

    "... astronomy is useful because it shows [not only] how small
    our bodies are, but how large our minds are." -- Henri Poincare

    Web Page: <home.earthlink.net/~mtnviews>
     
  15. w_tom

    w_tom Guest

    If manufacturer and vendor did not provide numeric specs up
    front and in writing when you bought the supply, then neither
    will. Appreciate why so many power supplies are selling at
    less retail price and at greater profit. They are marketing to
    MBA types who somehow are experts and therefore do not ask for
    numerical specs BEFORE making a purchase. As long as those
    functions are not in writing, then problems are all upon you.

    Why then would they even want to waste time providing
    specs? They have your money which is the only purpose of
    selling power supplies missing essential functions. Should
    you return it, then you may only get credit towards further
    purchases (maybe after a restocking charge) assuming they are
    willing to accept returns on electronics. Either way, they
    still have your money. Next year they may be selling this
    supply under a different name.

    Let's apply your technical reasoning to another homeowner.
    His lights worked just fine. Therefore house electricity was
    completely OK. Except that his house electricity was wired
    completely wrong. Fortunately the resulting gas meter
    explosion occurred when he was not home. Just because a
    computer appears to work fine means the new power supply is
    OK? Nonsense. The new supply was creating RFI from day one.
    Therefore it was always defective. You are using the same
    flawed logic of that homeowner. You assume one can 'test' for
    quality which even students of W E Deming generations ago
    understood to be false.

    Why did the previous supply become erratic? We don't even
    know that supply was problematic. For all we know, system was
    fixed by breaking and remaking connections which cleans
    contact corrosion. Why, with numbers, was the old supply not
    working? Shotgun solutions only provide enough information to
    *speculate* that the old supply was defective. Meanwhile RFI
    function (another power supply function) was not defective in
    that old supply - according to what you have posted.
    Apparently an RFI filter function is missing in the new supply
    - as even suggested by its retail price.

    A grounding solution is bogus and an obvious waste of time
    if power supply even interferes with a car's radio reception.
    I had provided one good reason why.

    No RFI leak will be that powerful. Power supply would be a
    perfect transmitter if missing that essential function.
    Shielding or grounding will not accomplish anything sufficient
    - would only be time wasted. Power wires would be a perfect
    transmitting antenna for a 'defective by design' power
    supply. How will you identify AC power wires as a leak? They
    are suppose to 'leak' if the internal line filter does not
    exist - as pbdelete's inspection would reveal immediately.
    But again, pbdelete told you how to address the problem
    quickly. (I have suggested how to avoid the problem next
    time).

    ATX power supply is already inside a metal box. Therefore
    supply already has shielding and grounding. Just another
    reason why grounding will not be a solution. If you have no
    intention of turning this into an electronics project, then do
    the filter inspection that pbdelete requested ... now. Or buy
    a minimally sufficient supply from a responsible vendor.
    Waiting for a manufacturer to respond is a fool's errand -
    also called trying to take the easy and convenient way out.

    Power supply is a big square wave oscillator typically
    running at a base frequency of 20+ KHz. Any wire connected to
    that power supply without filters is an antenna. When power
    supply is off, then oscillator is not working - therefore no
    DC power to computer. DC power to computer comes from that 20+
    KHz oscillator. A defective and powered on power supply
    becomes an RF transmitting station.

    If power supply does not state up front that it meets FCC
    Part 15 requirements and European equivalents, then you can
    bet the farm. It will not meet specs if it was designed for
    bean counter customers. Was FCC requirement and other
    critical functions even printed on the power supply's label?
    If those functions exist, then they would be printed on that
    label.

    How would I solve the problem at less cost and time? Go buy
    a responsoble supply from a responsible vendor. Throw the RFI
    supply in the trash. Chaulk the experience as money spent to
    learn the difference between product people verses a bean
    counter mentality.
     
  16. W. Watson

    W. Watson Guest

    First, let me note that I am in a community where electronic supplies are
    not readily available--6K people in the Sierra foothills miles and miles
    from Sacramento. I'm 60 miles from the nearest really full scale vendor, and
    45 miles from a good used electronic parts store. While you wer complaining
    that I wasn't following pdelete's dictum, I was trying to find a reasonable
    coil to put around the the AC cord per other suggestions. None were readily
    availble. Hence, it the intervening time I certainly felt that is worth
    pursuing other avenues in this thread.

    Second, pdelete asked if the unit had PFC, so not knowing I tried to contact
    the vendor. The answer is now known as "No". Only their units distributed to
    Europe have them I am told. Further, I talked to the vendor on the phone
    finally, they have no spec sheets or circuit diagrams to distribute. What
    one sees on the box is it.

    Third, I decided to pull the new supply and put in the old one to find out
    if my observation was correct. Yes, the old unit did not produce any RFI. I
    then decided to turn on another computer in the room. It did not produce any
    noise in the AM radio I was using to test interference. I then plugged that
    computer into the outlet for the other 'noisy' computer. Voilla. Noise galore.

    Finally, this prompted to look for other noise sources in the three rooms of
    the building and examine the ground on outlet just mentioned. Since I had no
    clue as to how the four circuits in the small building worked, I mapped all
    the outlets, switches, and lights to the circuit breaker panel. Then I began
    experimeinting by turning parts of the circuit on and off, etc. The best I
    could do using this approach (so far) is to note that the one circuit that
    has the most devices on it is fairly noisy. I found one device, a network
    hub, that was a bit noisier than anything else.

    Where I plan to go from here is likely to see if I can just plain quell the
    noise at the questionable outlet and devices plugged into it. Beyond that,
    I'll live with the ruckus caused on the AM portion of the dial that distrubs
    me the most. In the meantime, I'm close to being off this little project,
    and getting onto other things of higher priority. The obvious disturbance in
    the AM signal attracted me enough to put this much time in on it, but not
    enough to keep me going for days trying to squelch it. Other things first.
    When they are out of the way, I'll look into this further. Summing it up I
    seem to have an RFI problem of unknown origin.

    As far as the new PSU unit is concerned, I'm keeping it. At least, it
    doesn't just belly up every 3-4 days.

    Since you are curious about this, I'll try to do you the favor of
    photographing the inside of both PSUs and posting it. That's not likely to
    happen today.

    If nothing else, at least I now have the circuit panel mapped. Thanks for
    your interest.


    --
    Wayne T. Watson (Watson Adventures, Prop., Nevada City, CA)
    (121.015 Deg. W, 39.262 Deg. N) GMT-8 hr std. time)
    Obz Site: 39° 15' 7" N, 121° 2' 32" W, 2700 feet

    "... astronomy is useful because it shows [not only] how small
    our bodies are, but how large our minds are." -- Henri Poincare

    Web Page: <home.earthlink.net/~mtnviews>
     
  17. w_tom

    w_tom Guest

    Noise is 'routinely' blocked inside the metal (power supply)
    box by the line filter - if filter exists. You are fooling
    yourself if you think some coil on the power cord or grounding
    is going to solve the problem. The power supply appears to be
    defective when designed. It appears to be another classic
    example of power supplies designed for MBAs types who call
    themselves 'computer literate'. Overseas manufacturers have
    discovered a lucrative market of technically naive computer
    'experts' AND have dumped product into that market for higher
    profits.

    You don't need a local electronics store. The internet is
    chock full of responsible vendors who sell minimally
    acceptable supplies. Supplies that meet Intel ATX power
    supply specs - and say so. Supplies that meet FCC Part 15 and
    the even more stringent European standards. Supplies that
    must also meet ISO, CE, UL, and other standards - all in the
    same box. Where are those approval labels on your power
    supply? What other problems will those missing standard
    (functions) create?

    A previous post listed other functions that the power supply
    must provide. For example, does that supply have over voltage
    protection (OVP)? If not, then a power supply problem could
    destroy every other component inside a computer - later. OVP
    was standard even 30+ years ago so that a power supply failure
    to 'never' damage other computer parts. And yet power
    supplies missing essential functions such as AC line filter
    would also 'forget' to include other important functions.

    How would you know if that supply has OVP? As pbdelete
    noted - open the power supply to identify an AC line filter:
    If no filter, then you are not going to fix that supply.
    Other essential functions, such as OVP, are also probably
    missing. A supply designed for bean counters - people who are
    so often the reasons for failure.

    Noted earlier was how a power supply sells at lower price
    for greater profits. OVP is but another function routinely
    missing on power supplies designed for bean counter
    consumption. Power supply that provides no written
    specifications because so many such 'computer experts' only
    look at two numbers - dollars and watts. People who will
    rationalize their decision rather than take hard and
    conclusive actions.

    If that power supply does not have an AC line filter as
    pbdelete asked to inspect for, then that missing function may
    be part of an iceberg. Your only viable solution is a new
    power supply from a responsible power supply vendor and
    manufacturer. Some computer parts houses don't even stock
    such supplies.

    Meanwhile, what is accomplished by learning AC house
    wiring? Learn which wires are better and worse transmitting
    antennas. Again - and it should be so painfully obvious - no
    grounding was going to solve your RFI. That ground wire
    inside the wall only becomes a transmitting antenna - as was
    stated previously.

    There is no plug-in or power cord solution for a missing
    filter that was required and standard even generations ago.
    There is no "coil to put around the the AC cord". A supply is
    causing RFI interference. Then the entire supply is detective
    by design. Completely unacceptable, a reason for more future
    problems, and a classic example of why people with MBA
    mentalities create so many failures. Patching a solution on
    the power cord is called 'curing symptoms' or a 'kludge'. A
    solution must exist at the problem - inside that power supply
    box - where the solution would have cost so little money.
     
  18. Guest

    Why not shileded power cables?, especially since it's likely not complient to
    disconnect earth on the mains connector. Especially if something goes wrong.
     
  19. Guest

    Use an AM-radio, disturbances => RFI. (simple but working)
    You could also wire a simple directional antenna from steel wires if you
    want to get more accurancy easy.
    Push down a pole >2meter into the ground measure between jack and pole.
    Also check between jack and incoming terminal. Btw, this depend on the specific
    electrical code in your country. So be careful.
    Maybe the english word is cardboard/pasteboard. Thick paper used in packageing
    at least. Combined with household aluminium foil you can create a quick and
    cheap faradays cage to block RFI.

    However as suggested in later posts. IF the powersupply "sends" it's RFI back
    out through the powercord. Then no shielded box etc.. will help. Unless you
    add an line filter. I think many are two coils in parallel combined with
    suitable capacitors to shortcircuit highfrequencys.

    I had a look into this when investigating powercontrol via igbt and mosfet.
    Where one technique is modulated pwm to produce sinus. Which is nice.. except
    for the RFI ;)

    If you have a oscilloscope then you can verify many of these things yourself.
    At least PFC and Line filtering should be visable.
    Btw, does the disturbances affect FM-band too ?
     
  20. Jasen Betts

    Jasen Betts Guest

    IT all depends on what mode the RFI takes,

    If it;s balanced on the live and neutral conductors the only thing that'll
    stop it is a proper EMI filter that filteres each conductor separately.

    if it's common to all three conductors then the toroid etc will stop it
    abd the EMI filter may not (but I count this as unlikely)

    all that shielding does is keep the cable from radiating, if it's plugged
    into an outet shared with other unshielded cables they will radiate the RFI...
     
Ask a Question
Want to reply to this thread or ask your own question?
You'll need to choose a username for the site, which only take a couple of moments (here). After that, you can post your question and our members will help you out.
Electronics Point Logo
Continue to site
Quote of the day

-