Connect with us

low voltage drop transistors for power supply

Discussion in 'Electronic Components' started by Albert, Apr 26, 2004.

Scroll to continue with content
  1. Albert

    Albert Guest

    My neighbor brought his 25A (12 volts) power supply over for repair
    last evening. I found the bridge rectifier shorted and one of the 2
    pass transistors shorted. The 20A fuse in series with the output was
    not blown, the 4A line fuse was blown.

    I patched it enough to test it, putting in a 10A bridge rectifier and
    using only one of the pass transistors (they were in parallel).

    I found the transformer which looked hugh, was only putting out 17
    volts no load (dc, measured at the bridge output) and 15.5 at 10A
    load. While the transformer probably has the current capacity, it
    probably loads down under full load (at 20A). I couldn't test at 20A,
    but he confirmed that the ham radio transmitter had some quirky
    behavior, probably due to the voltage drop off at the higher load
    currents.

    It looks like someone (previous owner) disconnected the metering in
    order to get around the voltage drop associated with the meter.

    I'd like to put the thing back together for him, and have it operate
    properly.

    It had a 2N3055 and an MJE2955 pass transistors-not sure why they were
    different types, they were in parallel.

    I need suggestions for low collector to emitter voltage drop pass
    transistors that will handle 15A each. I also wonder if a bridge
    rectifier made from schottky diodes. Suggestions for the schottky
    diodes??

    Thanks,

    A
     
  2. Tam/WB2TT

    Tam/WB2TT Guest

    Albert,

    The 2N3055 is way too wimpy for this use. They run out of gain at around 5
    amps. They should have used 4 of them. Probably what happened is that one
    transistor was carrying all the load, due to uneven load sharing, and
    shorted. I would use a minimum of a pair of 2N3071 transistors.

    Schottky rectifiers would be good. If the transformer is open frame, I would
    see if you can thread a few more turns of wire on there for the secondary. I
    would start with 4 -5 turns. If this reduces the voltage, reverse the
    connections. I have done this with #14 stranded wire, it worked OK.

    I suspect this supply might have started life as a 12.0 V supply, not 13.8
    which is what your friend needs.

    Tam/WB2TT
     
  3. budgie

    budgie Guest

    Sounds like a can of worms. One issue is that the venerable 2N3055 isn't all it
    is cracked up to be, and was often mis0used in serious power supply roles. If
    you still can, get hold of a REAL data sheet with curves, and you'll find that
    hFE drops off to single digit gains at 7A or so. Thius often leaves them
    under-driven and hence the series volt drop becomes a problem.

    Also, if this is an NPN-emitter-follower-in-the-positive lead then the
    configuration is worthy of review. When the filtered bridge output is
    struggling for headroom, high-side drive is going to be hard to achieve when the
    going gets tough, compounding the 2N3055 gain problem. Although you probably
    want a drop-in replacement rather than a major design project, a decent PNP or
    two would make a better regulator in the face of low headroom. At a pinch, look
    at the whole 2955/3055 family with particular attention to the MJE types rather
    than the 2N3055.

    Schottky diodes will "buy" you something less than a volt IF you get a suitable
    high-current type, but a 40A device is unlikely to gain you much at 25A - you'll
    be looking at 80A or bigger devices to get the lean volt drop @25A.
     
  4. Tweetldee

    Tweetldee Guest

    17 volts at the secondary for your transformer sounds quite normal for a 12
    volt supply. The flag that I saw raised was the fact that it drops to 15.5
    volts under moderate load. Sounds like there may be a problem with the
    transformer or a connection to it, either on the primary or secondary side.
    Eliminate defective connections as the source of your problem, especially
    around the high-current secondary.
    If the pass transistors are mounted in sockets, make sure the connections
    are good there as well.

    But there's a bigger problem with your transistors... they probably aren't
    in parallel, since the 2N3055 is an NPN transistor, and the MJE2955 is a PNP
    unit. The MJE2955 is in a totally different case than the 2N3055, so it's
    very unlikely that these are supposed to be in parallel. Recheck your
    wiring and see if maybe the MJE2955 is used to drive the 2N3055. Even if
    the wiring indicates that it's the driver for the pass element, it's strange
    that it is a PNP transistor. Maybe you meant to say MJ2955 instead of
    MJE2955 ????
    Another problem lies in the fact that the 2N3055 has a spec'ed max collector
    current of 15A. If this is the only pass transistor in the supply, it can't
    possibly handle 25A. It's possible that if the MJ(E)2955 that you found in
    the supply should be another 2N3055. That would be much more believable
    than the MJ(E)2955.

    Sounds like the someone who attempted repair on this supply didn't really
    know what he was doing, and may have put some wrong parts in there.
     
  5. Albert

    Albert Guest

    All the connections are good, I checked them mechanically and
    confirmed by measuring the voltage drop across the high current
    connections.
    Ah, yes! You caught me! Yes, it is an MJ2955, my error. Sorry.
    Not sure which transistor should be there, but they are definately in
    parallel! The 2N3055 blew, the MJ2955 still operates and I used it to
    test the supply with.
    I am sure someone with a low level of competancy repaired it the last
    time it blew. Many of the soldered connections left a lot to be
    desired although I did not find any connection that actually failed.

    It says 25A on the front, but looking at the innards, it is clear that
    it probably never produced 25A. I suggested to the owner that we
    rebuild it as a 15 A constant duty with 20A intermittent output
    because it probably can't produce 25A no matter what we do to it. So,
    we are going to rebuild it from scratch.

    My hope was to gain some extra head room with a quick and dirty pass
    transistor and rectifier swap and build a simple regulator from the
    ground up.

    Should I be looking at a switching supply or try to squeek out a few
    extra volts of headroom by maintaining the origianl design with some
    improved parts?

    Thanks for all comments.
     
  6. Reg Edwards

    Reg Edwards Guest

    It is dangerous to connect two different types of power transistor in
    parallel. First one goes because it takes most of the load and then the
    other.
     
  7. There are very few choices in Schottky bridges in this size, but there
    are lots of single diodes big enough for this task. The best choice
    probably has more to do with how you can mount them on the heat sink
    than anything else. This is a good idea just to reduce the heat
    inside the box.

    For the pass transistors you might look into something much beefier,
    like the 2N5684 2N5686 size. These are rated for 50 amps but have
    good low voltage gain to about 10 or 20 amps. They have enough
    capacitance that they might destabilize the regulator.
     
  8. Tim Wescott

    Tim Wescott Guest

    What is it using for a regulator? You may want to consider boosting the
    main supply by putting in a little 6.3-volt transformer + bridge to make
    a base supply for your pass transistor. This will let you drive the
    pass transistor into saturation, which will give you 1/2 to 1 volt more
    headroom.
     
  9. Tweetldee

    Tweetldee Guest


    Well, that only adds to the confusion!!! As I said in my earlier post, the
    MJ2955 is a PNP unit, whereas the 2N3055 is NPN. They can't operate as
    series pass elements in parallel!!!!!!
    Is the output of the supply regulated now? Is it holding the output
    constant with changing load?
    If the supply is operating (albeit limping) with only the MJ2955 in the
    circuit, then I have to ask for a schematic, or at least a sketch of the
    circuit. I'm curious as to how this supply is built. What's the make/model
    of this supply?
    If it's simple enough to sketch, that may allow us to offer more intelligent
    answers to your dilemma.

    Cheers!!!!!
     
  10. Albert

    Albert Guest

    I was just at the Fairchild semi site looking for big diodes. The only
    ones they have are in TO-220 packages and although they have 30 and
    40A ratings, the power dissipation is only 3 watts. How can the thing
    drop .7 volts at 40A and only dissipate 3 watts of heat?? Now I'm
    confused:>:

    For instance:

    http://www.fairchildsemi.com/ds/MB/MBR3050PT.pdf

    Who else makes high current schottky diodes??

    I just received an email from a user suggesting syncronous
    rectification using MOSFETS. Although it's a little more complicated,
    it should be easier to implement at 60 hertz and I hadn't thought of
    this. It sure would give me the head room I need as the drop across
    fully turned on mosfets should be in the millivolt range.

    I'm really tempted to suggest to my friend that we just assemble it as
    a 10A supply with junkbox componentws that I already have on hand!

    Thanks for the suggestion of pass transistors.
     
  11. Tam/WB2TT

    Tam/WB2TT Guest

    TYPO - TYPO

    I meant to say 2N3771, not 3071.

    How are the pass transistors driven? If it is a Darlington configuration,
    you can get back about a Volt by using a PNP driver, but you now need a
    current sink to drive that. Also, the guy with the 6V transformer had a good
    idea.

    Tam
     
  12. That rating applies if you just have the diode sticking up into air,
    with no heat sink. The thermal resistance spec applies when you mount
    it in a heat sink and need to figure out how hot the junction gets
    compared to the heat sink. You need heat sink date to figure out how
    hot the heat sink gets.
    International rectifier makes lots of big ones. So does Diodes
    Incorporated. Digikey sells both these brands.

    (snip)
     
  13. Albert

    Albert Guest

    Makes perfect sense now. Thanks.

    I did have a quick look at IR, but probably the high power diodes
    listings.
     
  14. mikem

    mikem Guest

    I reading this thread, I never saw the filter capacitor mentioned.
    To get 13V at 20+A starting from a 17Vrms tranny, a fullwave Si bridge
    and leaving a little to drop across the series regulator transistor,
    the capacitor ripple voltage would have to be very low, less than
    ~3V.

    C*V=I*t; solve for C.

    V = 3V (ripple drop)
    I= 20A
    t=8.33ms (60Hz line, full wave rectifier)

    C= (20 * 0.0083)/3 = 0.0055 FARAD!

    You better have a 60,000uF 25V capacitor in there, do you?

    MikeM
     
  15. Albert

    Albert Guest

    No, it's a 6500uF at 18 volts. As I said in the original post, the
    transceiver sounded chirpy at high power and sometime the PLL would go
    out of lock and latch. Although it is a 50 watt output transceiver,
    I'm sure the quirky behavior can be attributed to the under design of
    the power supply. I'll put a larger cap on the shopping list, thanks
    for pointing this out!!
     
  16. Might be worth quietly checking that the bridge will run into that value
    capacitor at start-up without a big bang....

    We build big battery chargers and leave linear stuff behind at 200-300watts
    output, but one thing we have learned over the years is to pay close attention
    to startup surge ratings of the bridge into an effective short circuit
    (discharged capacitor) makes life a lot easier :))

    We ended up rating for surge rather than carry current, usually 4X output
    current was OK for continuous operation with switching on/off periodically. 25A
    bridges are great for 0-10A, but then we go higher if needed. Does not apply
    with smaller capacitors, check the ESR ratings etc.

    My 2c worth

    Peter
     
  17. I was thinking about this the other day as I pulled another heatsink
    with dual diodes off an old AT power supply. Since the cathodes of
    those dual 25A schottky diodes are connected together, why not just
    connect the two anode leads, and make a single 50A diode. Since they're
    on the same header and heatsink, if one hogs a bit more of the current,
    the other will get heated up also, which distributes the amperage more
    evenly than if they were two separate diodes.

    Since there are often a pair of these on the same heatsink, one for the
    high current 5V and one for the high current 12V, I could parallel them
    again and make some kind of arc welder. Hah-hah, just kidding. But I
    could get some seerious current with them paralleled. They might even
    work okay for the app mentioned above.
     
  18. Tweetldee wrote:

    [snip]
    That's exactly what seemed to have happened to an old HP power supply I
    picked up from Ebay recently. The thing wasn't putting out any current,
    and the pass transistors were burned out. Someone had subbed an
    inadequate transistor in there, and changed some resistor values.

    Last time I checked, judgment didn't need an e after the G, at least not
    in the U.S., anyway. But the saying rings true.
     
  19. Albert wrote:

    [snip]

    Has anyone tried to do this? Series aiding secondary to boost the ACV.
    (view with courier font)


    o-----+--+
    120VAC | | +-------o
    | )|| |
    120VAC Pr'y | )||(
    12VAC 3A Sec'y | )||(
    | )||(
    | )|| | 132VAC
    | | | 3A max.
    o--+--)---+ | to PS
    | | |
    | +--------+
    |
    +--------------------o


    One time I used a similar setup to make a 220 VAC guitar amp run off 120VAC.
     
  20. Same thing happens when they're matched. ;-)
     
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

-