Announcement

Collapse
No announcement yet.

Power Supply Design

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

    Power Supply Design

    I replaced all the caps in the Panstar 230W AT power supply (for the FIC VT-501 Pentium system I mentioned in another post) with Panasonic EB and FC parts. Nothing was obviously swollen or visibly degraded, and the two major filter caps were HEC. The power supply works just fine.

    My question has to do with the design of the AC input side, which I imagine is not very different in an ATX supply. There are jumpers installed in place of where LF1 could be. The following parts were not originally installed: LF1, C1, C2, C3, C4, and R1. C3 and C4 have since been added (0.0047uF, 500V ceramic disc) to provide RF filtering. R1, I'm guessing, is to discharge the L-C network, and could be 100k to 1M, so as to not load down the supply. C1 and C2: polystyrene or other non-polarized caps? The board has an outline for LF1. Is this usually a single part, or two separate coils? Recommended values? I'm not sure these parts are really needed, as the power supply ran just fine without them for several years. I'm curious more than anything.

    Also, what is the purpose of the 1-ohm power resistor in series with the input fuse? It's mounted on end, so one of the two leads isn't accessible from the top of the board. I'm thinking that it was for testing, to monitor the input current when the board was being tested at the factory, although it's not a precision 1-ohm resistor, so the voltage developed across it would depend on the line voltage as well as the resistor accuracy.
    Attached Files
    Last edited by AK0R; 10-26-2005, 09:06 PM.

    #2
    Re: Power Supply Design

    C1 and C2 need to be X2-class units, and C3 and C4 need to be Y2 (or Y1) class units. C1-C4 must be mains rated, and therefore are safety critical components.
    R1 safely discharges C1 and C2 when the supply is disconnected from the mains.
    The 1-ohm resistor is a inrush current limiting resistor.
    Since LF1 has jumpers, it may be on a separate PCB, like in the LC-B400ATX.

    I have never heard about a Panstar power supply before.
    My first choice in quality Japanese electrolytics is Nippon Chemi-Con, which has been in business since 1931... the quality of electronics is dependent on the quality of the electrolytics.

    Comment


      #3
      Re: Power Supply Design

      The following parts were not originally installed: LF1, C1, C2, C3, C4, and R1. C3 and C4 have since been added (0.0047uF, 500V ceramic disc) to provide RF filtering.
      It sounds like your P/S either isn't compliant with FCC EMI standards, those omitted components aren't needed to comply, or the PCB is used in another model that does need those components. Those 4700pF GP ceramic caps do not comply with European safety standards, and I'm unsure whether they comply with UL. If that P/S were mine, I'd try to find 4700pF agency-approved "Y-caps" (AC line-to-chassis; Ceramite and and Murata have approved ceramic types; Evox-Rife, Roederstein, Wima and others have metallized paper or film types).
      R1, I'm guessing, is to discharge the L-C network, and could be 100k to 1M, so as to not load down the supply.
      Its purpose is probably to discharge C1 and C2, but it was cheaper to install it.

      End of Part 1.
      PeteS in CA

      Power Supplies should be boring: No loud noises, no bright flashes, and no bad smells.
      ****************************
      To kill personal responsibility, initiative or success, punish it by taxing it. To encourage irresponsibility, improvidence, dependence and failure, reward it by subsidizing it.
      ****************************

      Comment


        #4
        Re: Power Supply Design

        quote]C1 and C2: polystyrene or other non-polarized caps?[/quote]
        Not seeing what you are seeing, I'd guess these are "X-caps" (line-to-line). These would usually be metallized paper, polester, or polypropylene, and agency-approved parts are advisable.
        The board has an outline for LF1. Is this usually a single part, or two separate coils?
        I'm guessing that this would be a common mode inductor. This is an inductor with two windings, one connected to "Line", the other to "Neutral", with the windings phased so that noise that is the same on "Line" and "Neutral" gets cancelled. This is usually a custom part, tailored to the application, and the construction must meet safety agency requirements.

        End of Part 2
        PeteS in CA

        Power Supplies should be boring: No loud noises, no bright flashes, and no bad smells.
        ****************************
        To kill personal responsibility, initiative or success, punish it by taxing it. To encourage irresponsibility, improvidence, dependence and failure, reward it by subsidizing it.
        ****************************

        Comment


          #5
          Re: Power Supply Design

          I'm not sure these parts are really needed, as the power supply ran just fine without them for several years.
          I speculated about this above. AC filters usually aren't essential to P/S operation, but the FCC doesn't want your computer to prevent you and your neighbors from seeing and hear all those wonderful TV and radio ads.

          [quote]Also, what is the purpose of the 1-ohm power resistor in series with the input fuse?[quote]
          Well, I'm not seeing what you're seeing, but my guess is that the P/S mfr cheaped out and used a regular resistor instead of a thermistor to limit inrush current at turn-on. When you first apply power to the P/S (after it's been disconnected from AC for more than a few minutes), only the size of the wires and PCB traces (and the contact resistances of the wall outlet and P/S inlet) limit the magnitude of those first couple of current pulses, which are charging those I/P lytic caps. This only takes .1 seconds or so, but the peak currents can be >>50A.

          End of Part 3
          PeteS in CA

          Power Supplies should be boring: No loud noises, no bright flashes, and no bad smells.
          ****************************
          To kill personal responsibility, initiative or success, punish it by taxing it. To encourage irresponsibility, improvidence, dependence and failure, reward it by subsidizing it.
          ****************************

          Comment


            #6
            Re: Power Supply Design

            A thermistor is usually disc shaped: at room temperature, it can be 5 ohms, 10 ohms, 25 ohms (whatever the designer chooses), but when hot, its resistance can drop to .05 ohms or .01 ohms (depends on the chosen part). These do a good job of limiting the inrush current, but if AC is lost and comes back in a few seconds, they won't limit the inrush current very well. If that 1 ohm resistor is a sub for a thermistor, it won't limit the inrush current very well [(120*1.414)Volts/Ohms], dissipates more power than a thermistor, but will still be 1 ohm in case of a brief AC drop-out. A 5 or 10 ohm thermistor rated for your P/S's maximum I/P current of proper physical size would be the better choice since it is better at inrush limiting and dissipates less power (less heat in your computer).

            End of Part 4

            I'm not trying to be dramatic, but I couldn't post this as a single reply. If the Mods want to merge this into one post, please do.
            PeteS in CA

            Power Supplies should be boring: No loud noises, no bright flashes, and no bad smells.
            ****************************
            To kill personal responsibility, initiative or success, punish it by taxing it. To encourage irresponsibility, improvidence, dependence and failure, reward it by subsidizing it.
            ****************************

            Comment


              #7
              Re: Power Supply Design

              Lower-wattage and many AT PSUs don't have the input filter installed.
              I'd just keep the input section as is. If it ain't break, don't fix it

              Comment


                #8
                Re: Power Supply Design

                Originally posted by Rainbow
                Lower-wattage and many AT PSUs don't have the input filter installed.
                I'd just keep the input section as is. If it ain't break, don't fix it
                I agree that it's probably because it's on the low power side. The design looks as though it could accommodate a higher rating, as the installed heat sinks are rather chintzy (they look a lot like those in your LC-B400ATX).

                My reason for investigating this is because the computer and power supply would become unstable during radio transmissions, either due to the rig loading down the AC input (I find that hard to believe, running 100W output), or more likely, due to "reception" of radiated RF on the power supply cord. Therefore, my primary defense is installation of RF bypass capacitors on line-to-ground and neutral-to-ground, followed by deciding if the AC input filter will provide sufficient common mode rejection.
                Originally posted by PeteS in CA
                Its purpose is probably to discharge C1 and C2, but it was cheaper to install it.
                R1 was not installed by the manufacturer. 100k to 1M is my estimate on an appropriate size to install to discharge the AC input filter network. The inrush resistor looks like one of those "blocks of sand" power resistors (Ohmite?), except that it's white instead of tan colored. Thanks for the info on design. I'll check out your recommended sources for approved parts. I'm not terribly concerned about meeting European standards, but neither do I want to make a modification that makes the power supply unsafe.
                Last edited by AK0R; 10-27-2005, 06:27 PM.

                Comment


                  #9
                  Re: Power Supply Design

                  My reason for investigating this is because the computer and power supply would become unstable during radio transmissions, ... or more likely, due to "reception" of radiated RF on the power supply cord. Therefore, my primary defense is installation of RF bypass capacitors on line-to-ground and neutral-to-ground, followed by deciding if the AC input filter will provide sufficient common mode rejection.
                  Your diagnosis seems reasonable. The Line/Neutral-ground caps you installed provide some common-mode rejection. LF1 would be wound with an equal number of turns in each winding, in-phase, so that both starts would be connected to the same side (AC side or P/S side). If you wind your own, use a ferrite toroid (5000 or 10000 perm would be best; insulate it with mylar or Kapton tape if it isn't coated), a wire gauge that will handle the current, and do separate windings - not bifilar, but opposite each other on the toroid (keep the windings physically separated, I think 8mm, or 3 layers of tape, or a physical barrier is the European standard). The increases the leakage inductance, which becomes differential-mode inductance.
                  I'll check out your recommended sources for approved parts. I'm not terribly concerned about meeting European standards, but neither do I want to make a modification that makes the power supply unsafe.
                  With their higher mains voltage and bureacratization, the Europeans have been leaders in safety and EMI regs, though UL's and CSA's have been similar for 15 years or more. The higher quality parts are readily available and aren't too expensive. Check Digi-Key, as they probably have agency-approved parts made by Panasonic (I forgot them yesterday).

                  Now, will the SW let me post this in one post?
                  Last edited by PeteS in CA; 10-28-2005, 08:02 AM.
                  PeteS in CA

                  Power Supplies should be boring: No loud noises, no bright flashes, and no bad smells.
                  ****************************
                  To kill personal responsibility, initiative or success, punish it by taxing it. To encourage irresponsibility, improvidence, dependence and failure, reward it by subsidizing it.
                  ****************************

                  Comment


                    #10
                    Re: Power Supply Design

                    Originally posted by japlytic
                    I have never heard about a Panstar power supply before.
                    http://www.panstar.com/html/about.html

                    They produce PC cases and power supplies, as well as lighting products. They are based in Thailand.

                    Comment

                    Working...
                    X