Thank you to the guys at HEGE supporting Badcaps [ HEGE ] [ HEGE DEX Chart ]

Announcement

Collapse
No announcement yet.

Modifying the FSP Epsilon Design

Collapse
X
Collapse
 
  • Page of 1
  • Filter
  • Time
  • Show
Clear All
new posts
Previous template Next
    #1

    Modifying the FSP Epsilon Design

    Many of you are probably already familiar with the high ripple and noise the FSP Epsilon design can produce under heavy load. jonnyGURU has documented the issue several times.

    I have one of these sitting before me, and would like to take a stab at cleaning up the output somewhat by upgrading the caps in it. I'm going to be concentrating mostly on the 12v rail, which is serviced by OST caps, but will probably replace the others as well.

    From what I can tell of the design, this thing uses two OST 2200uF caps in parallel to filter the 12v output. However, I'm not sure if there's a proper pi filter in here... there's a third OST of that value, but it connects to the smaller of the two coils that runs to the 5v and 3.3v section. It actually looks like there's a pi filter for the two lower voltages, but not for 12v. I could be wrong though - I'll post a pic.

    My question is, would higher values in good quality caps on the 12v output be likely to improve the ripple situation? I was thinking I'd move to 2700uF if I can get them in there.



    This pic shows all three OST's - one is the one wrapped in heatshrink.
    Attached Files
    Tags: None
    #2
    Re: Modifying the FSP Epsilon Design

    Some more info - I've been studying pics of Zalman's heatpipe based version of this unit, and it looks like they've made some changes since the above unit came out. The 12v section looks very similar, except instead of using two OST's in parallel on the output it looks to have one OST and four 470uF 16v Teapos in parallel. Interesting.

    Should have mentioned before that I'm doing all this for an article I'm writing for jonnyGURU.com - if that represents a conflict of interest, feel free to delete my thread. I plan to give credit where it's due.

    Comment

      #3
      Re: Modifying the FSP Epsilon Design

      Well, reducing the ESR ist alway the easiest way to reduce ripple current (as long as the required capzitance is already there). four 470uF caps have seriousely a far lower esr than one 2200uF. You can imitate this desing in changing the caps with some lower ESR ones. But usually, the last cap in a phi filter is the most important regarding esr. I put some Panasonic FM in my Amacrox Free Earth 500w, wich is probalby the same desing as the Epsilon series. I have increased the values slightly too. I replaced all 1000uF CapXon caps by 1200uF ones. The 12v ones are now 1500uF. I am not shure, if that was a good choice. The inital desing of the Free Earth series is a litle different regarding the o/p caps. Instead of a few large caps, they only are using 1000uF caps at all voltages. But on the 12v line, each rail (this means each wirering pad wich has its own overcurrent shunt resistor) has its own caps. This equals to 2200uF.

      Comment

        #4
        Re: Modifying the FSP Epsilon Design

        Cool, thanks. I think I can get 12.5mm caps to fit in here, so I might try Panasonic FM or Chemicon KY. Maybe nudge them up to 2700uF too.

        Comment

          #5
          Re: Modifying the FSP Epsilon Design

          Hm are all those 16v caps Ost RLP series? The datasheets spec is 12,5mm and an ESR of 0,030 Ohm. You can try Panasonic FM or Rubycon MCZ. I would Prefer Fm, due to 3000h loadlife

          Comment

            #6
            Re: Modifying the FSP Epsilon Design

            RLP is indeed what they are - thanks again.

            Comment

            Previous template Next
            Working...
            X