Connect with us

Who makes this MOSFET?

Discussion in 'Electronic Repair' started by DaveC, Nov 19, 2003.

Scroll to continue with content
  1. DaveC

    DaveC Guest

  2. Tom Wright

    Tom Wright Guest

    DaveC
    You have to face reality, not even Einstein could help you answer this
    question if he was alive.

    You need to do better with your question,

    By the way DaveC, How you ever been or are you a politician in the New
    Zealand Government?




    http://home.covad.net/~peninsula/Amp/FET.tiff

    (All I have is the photo; no actual unit.)

    It's used in a GamuT power amplifier.

    Thanks,
     
  3. Robert Baer

    Robert Baer Guest

    Looks like it could be a bridge rectifier; most power FETs in that
    kind of a package have only three terminals per FET (and there are some
    with multiple FETs).
    Look at IRC, STM and other makers for a replacement.
     
  4. I read in sci.electronics.design that Robert Baer
    I agree. Also, if one wants to have a component identified from its
    picture, it's a good idea to post a picture with a sensible resolution,
    which doesn't mean flagstone-sized pixels.(;-)
     
  5. DaveC

    DaveC Guest

    Thanks so much to the (non-Aus group) contributors who offered some real
    suggestions.

    This photo is all I have. I don't have a device, or hi-res pics, just this
    one crummy photo (not mine).

    I'll start with the IRC, STM, and IXYS mfgrs. and see what they show.

    Thanks again for the leads.
     
  6. DaveC wrote...
    You won't get far without the part number. The SOT-227 parts are
    expensive and hard to get, especially in small quantities, so your
    best path is to get that part number and let us know. Then perhaps
    we can help you find an available replacement part.

    Thanks,
    - Win

    whill_at_picovolt-dot-com
     
  7. DaveC

    DaveC Guest

    Hmm... I have no p/n's. The MOSFET in question is used in an audio amp
    (Danish brand GamuT). I was curious what mfr. was making audio-quality FETs
    of this size. GamuT is not forthcoming with any supplier info.

    So I'm left with virtual "reverse engineering" (don't even have an sample to
    start with...)
     
  8. Aha, finally the reason for the original question starts to emerge... but
    it's still not quite there; we still don't know *why* you want to know what
    the part is. Will you forgive me for suspecting that the right answer to
    your *problem* has nothing to do with the right answer to your *question*?

    Originally it sounded like your problem might have been "what FET should I
    use to replace a blown FET in a device I am repairing." But it sounds like
    that's not the case? You've been clear that you don't have the device in
    your picture, but you've still not really explained what it is you're trying
    to do.

    If your problem is "what FETs should I use in an audio amp I am building,"
    let me suggest that "whatever was used by the designer of some other
    circuit, manufactured in some other country" is not actually a good
    criterion, no matter how good their circuit sounds. The sound of an amp is
    not due to good-sounding FET's, it is due to good design, which matches the
    FET to the circuit demands; and I'm not sure there is such a thing as an
    "audio-quality FET", although there are probably some that are better suited
    than others to operation in the linear range. A good criterion for
    selection would be "whatever meets the specifications implied by my circuit
    and is available through the suppliers I have access to".

    That criterion implies that you are going to need to actually understand the
    circuit you're trying to build, and understand what specifications are
    required (or, at least, how the specs affect the performance). Now *that*
    is a question that many folks on these groups are prepared to answer.
    Having done that, you can then try to search for an FET that meets those
    specifications; and again, folks on these groups are excellent at that type
    of question.

    Of course, if that is not your actual problem, then those questions are not
    the appropriate ones. Can you let us know what the problem you're trying to
    solve is?
     
  9. DaveC

    DaveC Guest

    Well said, Walter.

    I have come across the statement by some "single-FET" amplifier designers
    that when using paralleled output devices that because if their differences
    in turn-on characteristics there is some distortion introduced. This is a
    given, due to the manufacturing tolerance differences. (We won't argue about
    what magnitude this might be, and if it is at all audible.)

    By using a single, high-power FET one can reach equivalent results as a
    multi-FET design.

    I would like to experiment with such a single-FET design, and since not every
    manufacturer makes high-power FETs in such a package, I thought I'd start by
    fiding out who makes such FETs. Then, ruling out those devices whose specs
    are more appropriate for switching duties (rather than linear duties such as
    audio), etc. arrive at a short list of devices to start exploring more
    closely.

    Does that help clarify?

    Of course this may all come to nothing more than intellectual curiosity. But
    that's what curiosity is for, isn't it, to be satisfied.

    Thanks,
     
  10. DaveC

    DaveC Guest

    Hmm... Better to say:

    By using a single, high-power FET, one can reach equivalent *powers* as a
    multi-FET design, but without the inherent distortions.
     

  11. Aha! So, the problem is: "I would like to find a MOSFET that is appropriate
    for a high-power class-A single-FET audio amplifier design. Said MOSFET
    should be able to dissipate a lot of power."

    Now, that's a nice clear problem statement, lacking only quantification of
    "high power" and a particular schematic; and neither of those is especially
    necessary, since the former can just be "how high can I get?" and the latter
    is pretty obvious from the problem statement. (I'm assuming you're thinking
    in terms of, e.g., the "Zen" amp.)

    Group, any suggestions?

    (Dave, if no one answers, you might want to re-post with a new subject line,
    and with your new problem description.)
     
  12. Asimov

    Asimov Guest

    "DaveC" bravely wrote to "All" (20 Nov 03 09:20:32)
    --- on the heady topic of "Re: Who makes this MOSFET?"

    Da> From: DaveC <>

    Da> Well said, Walter.

    Da> I have come across the statement by some "single-FET" amplifier
    Da> designers that when using paralleled output devices that because if
    Da> their differences in turn-on characteristics there is some distortion
    Da> introduced. This is a given, due to the manufacturing tolerance
    Da> differences. (We won't argue about what magnitude this might be, and
    Da> if it is at all audible.)
    Da> By using a single, high-power FET one can reach equivalent results as
    Da> a multi-FET design.

    Da> I would like to experiment with such a single-FET design, and since
    Da> not every manufacturer makes high-power FETs in such a package, I
    Da> thought I'd start by fiding out who makes such FETs. Then, ruling out
    Da> those devices whose specs are more appropriate for switching duties
    Da> (rather than linear duties such as audio), etc. arrive at a short list
    Da> of devices to start exploring more closely.

    Da> Does that help clarify?

    Actually some switching types are better suited to linear operation than
    some linear devices but that's another topic altogether. Here is
    something else to muddle the issue. Some power fets are not a single
    device but are actually made up of an array of many fets all wired up on
    the die in parallel. The theory is that the fets are virtually identical
    and the best possible match that can be had. Some even mix in bipolar
    devices to get the best of both types. There is a lot of choice out
    there these days!
     
  13. Walter Harley wrote...
    I have some comments. First, with the exception of Hitachi's
    lateral MOSFETs, which feature a negative current tempco when
    in the linear region, there are few features that make a FET
    more suited for switching that linear use. For common D-MOS
    power FETs the primary spec of interest is thermal resistance,
    because that's what determines the maximum power dissipation.

    You will find that older FET designs all have a lower thermal
    resistance for a given current capability, because they had a
    larger die area. But this really isn't relevant, because what
    counts is die-size / thermal resistance vs. price. There are
    some rather impressive low-cost high-power-handling MOSFETs
    available nowadays, but you'll find that they aren't available
    at reasonable high-power audio-amplifier power-supply voltage
    ratings, i.e. 150 to 250V. That's because high-volume stuff
    is concentrated in the automotive area, or in off-line motor
    control, etc. For example consider the high-power FETs below:

    part no. watts volts
    -------- ---- ----
    FQA170N06 375 60
    IRFP2907 470 75
    HUF75652 515 100
    FDH44N50 750 500
    STW47NM50 417 500

    (FET thermal resistance ratings are inversely proportional to
    the power rating in watts.) BTW, the power rating assumes a
    25C case temperature, which of course is blatantly impossible.

    All of these parts come in low-cost packages. The low-voltage
    parts have too low a voltage rating, and the high-voltage parts
    are too high. Well, maybe not; why does it matter if they have
    an excessively-high voltage rating? Oops, if one digs deeper
    and examines Ciss, you may have an answer.

    Now, with respect to the SOT-227B parts. These do have a nice
    low thermal resistance, but they cost $25 to $75 each, and are
    very hard to get. Usually a high minimum order is required,
    plus a long lead time for your expensive order.

    A typical power rating for a SOT-227B part is 500 to 680 watts.
    Not so attractive, compared to the lower-cost easier-to-get
    parts listed above.

    Thanks,
    - Win

    whill_at_picovolt-dot-com
     
  14. As you know, I like the Thompson Isotop/SOT-227 package.

    The isolation and the heavyweight connection bolts are both
    already done for you. For 1-off or low volume jobs the simple
    assembly can easily compensate for the higher purchase price.

    And you can bolt a little pcb onto the lugs for close up
    components (current shunts or surge suppressors, etc).

    And maintenance in the field (by the customer) is relatively
    uncomplicated, just a screwdriver operation. No need to
    unsolder/resolder, and no worry about isolation being lost
    by a careless tech.

    And the nice hefty csa of the connecting lugs means that
    you don't have to look out for that little ambush that
    sometimes says (for Ipk) "Exceeds package capability".

    Easy to get (in the UK) from Farnell or Radiospares.
     
  15. Tony Williams wrote...
    Aha, I ignored the very important issue of case-to-heat-sink
    thermal resistance. For a TO-247 package with a thin grease
    layer (no insulator) this is about 0.24 C/W and dramatically
    changes the actual-usage rating. The new calculated values
    are below, using 175C junction temperature and 25C heat sink.

    P = (Tj - Tc) / (R_theta-JC + R_theta-CS)

    part no. watts volts real watts *
    -------- ---- ---- ------------
    FQA170N06 375 60 234
    IRFP2907 470 75 268
    HUF75652 515 100 283
    FDH44N50 750 500 341
    STW47NM50 417 500 232 (150C junction)

    Now the TO-247 FETs look much less attractive compared to
    SOT-227 parts. Yet if we were to redo the calculation with
    0.5 C/W (or worse) insulators, then they'd really look like
    wimps. The best FET above, a FDH44N50, rates only 214 watts
    with a hypothetical 0.5 C/W insulator. This is only 1/3 to
    1/4 of the capability of one of Tony's favorite parts.

    * Note for novice engineers: we can rarely assume 25C heat-
    sink temperatures in real life. Not only is the heat sink
    going to get hot in operation, but the FET contact point will
    be even hotter due to internal thermal resistance in the heat
    sink. For example, a 60C heat-sink surface might be seen in
    a well-designed aggressively-used system. This reduces the
    FDH44N50 rating to 164 watts. Furthermore, most engineers
    will NOT design for operation of a power MOSFET continuously
    at it's full rated junction temperature. If we use Tj = 150C,
    then the FDH44N50 rating drops to only 128 watts.

    128 watts is a realistic maximum operating power dissipation
    for this FET, which carries a data-sheet rating of 750 watts!
    Those insulators can be real performance killers! :>)
    OKEY DOKEY!

    Thanks,
    - Win

    whill_at_picovolt-dot-com
     
  16. Ben

    Ben Guest

     
  17. Ben wrote...... You were saying?

    Thanks,
    - Win

    whill_at_picovolt-dot-com
     
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

-