Loading ...
Sorry, an error occurred while loading the content.

Death to MOSFET's - prevention?

Expand Messages
  • luthjej
    Hi, I m now on my third set of TIP3055 s (8 of them per set) - as I have killed the previous 2 sets. Bascially I do something stupid (like apply a reverse
    Message 1 of 10 , Jul 2, 2006
    • 0 Attachment
      Hi,
      I'm now on my third set of TIP3055's (8 of them per set) - as I
      have killed the previous 2 sets. Bascially I do something stupid (like
      apply a reverse polarity to the output or short out random pieces of
      the circuit).

      The symptom of the NPN's is that something internal shorts out and
      starts dumping volts from the supply leg (C) to the base (B), which
      turns all of the NPN's on ... making it impossible to tell which is
      the culprit, making it easier just to pull the whole lot off and
      replace then with new ones.

      There are 8 NPN's in total in this variable PSU circuit, the output
      legs connected via 0.1ohm 5W resistors to common rails which supply
      the current.

      Is there any way to prevent this from happening? It's pretty high
      current (40A @ 25VDC), so I'm not convinced that a fast-blow fuse
      would protect from this sort of fault.

      Not convinced excess gate voltage is the culprit either, in fact I'm
      nearly convinced that it isn't.

      Any ideas? A strategically placed diode / zener? One fuse per NPN?

      Give up on bi-polars and go FET instead?
    • Roy J. Tellason
      ... I don t know what your circuit looks like, but is it possible that you re exceeding the collector-base breakdown voltage rating? ... Which pins? A lot of
      Message 2 of 10 , Jul 2, 2006
      • 0 Attachment
        On Monday 03 July 2006 01:15 am, luthjej wrote:
        > Hi,
        > I'm now on my third set of TIP3055's (8 of them per set) - as I
        > have killed the previous 2 sets. Bascially I do something stupid (like
        > apply a reverse polarity to the output or short out random pieces of
        > the circuit).
        >
        > The symptom of the NPN's is that something internal shorts out and
        > starts dumping volts from the supply leg (C) to the base (B), which
        > turns all of the NPN's on ... making it impossible to tell which is
        > the culprit, making it easier just to pull the whole lot off and
        > replace then with new ones.

        I don't know what your circuit looks like, but is it possible that you're
        exceeding the collector-base breakdown voltage rating?

        > There are 8 NPN's in total in this variable PSU circuit, the output
        > legs connected via 0.1ohm 5W resistors to common rails which supply
        > the current.

        Which pins? A lot of high-power audio amp designs that I've seen have tried
        to just run stuff in parallel like that, but a lot of other designs have used
        small-value resistors like what you mention in each emitter lead, and also
        sometimes other small-value resistors (say 10 ohms or less) in the base lead
        as well.

        > Is there any way to prevent this from happening? It's pretty high
        > current (40A @ 25VDC), so I'm not convinced that a fast-blow fuse
        > would protect from this sort of fault.
        >
        > Not convinced excess gate voltage is the culprit either, in fact I'm
        > nearly convinced that it isn't.
        >
        > Any ideas? A strategically placed diode / zener? One fuse per NPN?
        >
        > Give up on bi-polars and go FET instead?

        Hard to say without knowing the circuit. And the load you're trying to drive
        with it...

        --
        Member of the toughest, meanest, deadliest, most unrelenting -- and
        ablest -- form of life in this section of space, a critter that can
        be killed but can't be tamed. --Robert A. Heinlein, "The Puppet Masters"
        -
        Information is more dangerous than cannon to a society ruled by lies. --James
        M Dakin
      • Shawn Upton
        Well, your title is a bit misleading--you re asking about killing bipolars not FET s. Anyway... I wonder if you re getting into secondary breakdown, or
        Message 3 of 10 , Jul 3, 2006
        • 0 Attachment
          Well, your title is a bit misleading--you're asking
          about killing bipolars not FET's.

          Anyway...

          I wonder if you're getting into secondary breakdown,
          or violation of SOA--if you look at a load line for
          bipolars, you'll see the typical VI curve at low
          powers; but at some high powers you'll see a secondary
          curve, at a much different slope. I believe it is due
          to hot spots and/or current hogging: big transistors
          are invariably made by paralleling lots of smaller
          transistors on the die. MOSFET's don't have this
          issue, while bipolars do; I believe it arises from the
          fact that the Vt of MOSFET's go up with temperature,
          while Vbe of bipolars go down. As in, heat up a
          MOSFET, and the drive requirements go up; heat up a
          bipolar and the drive requirements go down. Bipolars
          can "run away" while FET's tend to drop out and
          current share better.

          You may have insufficent heatsinking. Although this
          supply is rated at 40A at 25V, I'll be that it's rated
          to like 16V at 5V. Why? I'll bet again that the Vdc
          unregulated from the diode rectifier section is about
          30ish volts. Maybe a bit higher even. (30-25)*40 is
          200W. Alot of heat. But (30-5)*40 is 1000W. If this
          supply doesn't have big fans, lots of heatsink surface
          and/or overtemp protection, I'd guess you're literally
          cooking the transistors.

          Until you get it figured out, you should invest in an
          overvoltage clamp. I believe they are typically made
          with like a resistor divider and a SCR, meant to fire
          above a certain voltage and short out the supply--with
          the intention of tripping the fuse. I'm going to
          guess whatever you are driving does not what to see
          30+ volts. I don't have a schematic on hand, but a
          bit of searching (perhaps looking at commercial
          supplies) should yield what you need.

          Are you driving inductive and/or capacitive loads?
          This is where you often run into SOA issues, albeit
          with op-amps and non-DC applications. You may need
          some snubber or TVS diodes here, as an inductive
          kickback could violate emitter to base breakdown
          voltages. Alternatively, sometimes this is very easy
          to do if you have insufficent buss capacitance after
          the rectifier stage; if that cap discharges fast (in
          particular if there is more than just this supply
          seeing that cap), sometimes you can run into a Veb
          violation due to Vout being higher than the collector
          voltage.

          Is this a homebrew supply? Maybe you could use some
          sort of overcurrent protection, where you sense
          voltage off a 0.01 ohm resistor and either blow the
          fuse or kill the base drive when more than 40A is
          delivered. Output shorts can be hard to protect
          against, you may just need to add a 50A fuse for this;
          and then perhaps a 30V TVS (25 if it doesn't get too
          warm in practice).

          Shawn

          --- luthjej <luthjej@...> wrote:

          > Hi,
          > I'm now on my third set of TIP3055's (8 of them
          > per set) - as I
          > have killed the previous 2 sets. Bascially I do
          > something stupid (like
          > apply a reverse polarity to the output or short out
          > random pieces of
          > the circuit).
          >
          > The symptom of the NPN's is that something internal
          > shorts out and
          > starts dumping volts from the supply leg (C) to the
          > base (B), which
          > turns all of the NPN's on ... making it impossible
          > to tell which is
          > the culprit, making it easier just to pull the whole
          > lot off and
          > replace then with new ones.
          >
          > There are 8 NPN's in total in this variable PSU
          > circuit, the output
          > legs connected via 0.1ohm 5W resistors to common
          > rails which supply
          > the current.
          >
          > Is there any way to prevent this from happening?
          > It's pretty high
          > current (40A @ 25VDC), so I'm not convinced that a
          > fast-blow fuse
          > would protect from this sort of fault.
          >
          > Not convinced excess gate voltage is the culprit
          > either, in fact I'm
          > nearly convinced that it isn't.
          >
          > Any ideas? A strategically placed diode / zener? One
          > fuse per NPN?
          >
          > Give up on bi-polars and go FET instead?
          >
          >
          >
          >
          >
          >
          >


          Shawn Upton, KB1CKT

          __________________________________________________
          Do You Yahoo!?
          Tired of spam? Yahoo! Mail has the best spam protection around
          http://mail.yahoo.com
        • luthjej
          Hi Shawn, Thanks for the reply, a few interesting bits of info there I wasn t aware of. This is a homebrew 3-stage battery charger, so no inductive loads to
          Message 4 of 10 , Jul 3, 2006
          • 0 Attachment
            Hi Shawn,
            Thanks for the reply, a few interesting bits of info there I wasn't
            aware of. This is a homebrew 3-stage battery charger, so no inductive
            loads to deal with!

            Sorry about the title ... I was halfway into the post when I realized
            that a TIP3055 is actually bipolar, not FET .. :)

            Heatsinking is very adequate for these, they are mounted on 2 x fan
            forced heatsinks (4 on each) at 0.30C/W each half. Total supply
            current (out) is up to 15v @ 40A, bridge supply is 25V, the base of
            the NPN's is driven via a 30V rail delivered from a voltage multiplier.

            The "problem" as such is not in the operation of the supply, just in
            me doing stupid things (the last "stupid" thing was connecting a
            battery in reverse polarity) which caused the failure of the last
            batch of NPN's), so really I guess my question is more in the
            protection of the NPN's rather than destruction via overloading.

            There is also thermal protection on the heatsinks which shuts down
            drive to the transistors if the temp goes over 90degC.

            I've incorporated reverse polarity protection by way of a diode
            network on the output which will start conducting (and theoretically
            trip the breaker) if the output is connected in reverse ... but sadly
            this doesn't seem to be the case.

            I have a 50amp thermal circuit breaker inline with the output - which
            didn't trip, I realize that these thermals can be quite slow to react.

            So you are suggesting that a 50A wire or ceramic fuse (HRC?) will blow
            faster and give better protection? I'd rather change a fuse than
            half-a-dozen NPN's! :)

            These NPN's just seem to be very fragile - especially given their
            rated instantaneous current and wattage rating, and considering I'm
            not dealing with anywhere near their rated voltage & it would be nice
            to have some sort of fail-safe to protect against idiots (like me!)

            Thanks again,

            Jon

            --- In Electronics_101@yahoogroups.com, Shawn Upton <kb1ckt@...> wrote:
            >
            > Well, your title is a bit misleading--you're asking
            > about killing bipolars not FET's.
            >
            > Anyway...
            >
            > I wonder if you're getting into secondary breakdown,
            > or violation of SOA--if you look at a load line for
            > bipolars, you'll see the typical VI curve at low
            > powers; but at some high powers you'll see a secondary
            > curve, at a much different slope. I believe it is due
            > to hot spots and/or current hogging: big transistors
            > are invariably made by paralleling lots of smaller
            > transistors on the die. MOSFET's don't have this
            > issue, while bipolars do; I believe it arises from the
            > fact that the Vt of MOSFET's go up with temperature,
            > while Vbe of bipolars go down. As in, heat up a
            > MOSFET, and the drive requirements go up; heat up a
            > bipolar and the drive requirements go down. Bipolars
            > can "run away" while FET's tend to drop out and
            > current share better.
            >
            > You may have insufficent heatsinking. Although this
            > supply is rated at 40A at 25V, I'll be that it's rated
            > to like 16V at 5V. Why? I'll bet again that the Vdc
            > unregulated from the diode rectifier section is about
            > 30ish volts. Maybe a bit higher even. (30-25)*40 is
            > 200W. Alot of heat. But (30-5)*40 is 1000W. If this
            > supply doesn't have big fans, lots of heatsink surface
            > and/or overtemp protection, I'd guess you're literally
            > cooking the transistors.
            >
            > Until you get it figured out, you should invest in an
            > overvoltage clamp. I believe they are typically made
            > with like a resistor divider and a SCR, meant to fire
            > above a certain voltage and short out the supply--with
            > the intention of tripping the fuse. I'm going to
            > guess whatever you are driving does not what to see
            > 30+ volts. I don't have a schematic on hand, but a
            > bit of searching (perhaps looking at commercial
            > supplies) should yield what you need.
            >
            > Are you driving inductive and/or capacitive loads?
            > This is where you often run into SOA issues, albeit
            > with op-amps and non-DC applications. You may need
            > some snubber or TVS diodes here, as an inductive
            > kickback could violate emitter to base breakdown
            > voltages. Alternatively, sometimes this is very easy
            > to do if you have insufficent buss capacitance after
            > the rectifier stage; if that cap discharges fast (in
            > particular if there is more than just this supply
            > seeing that cap), sometimes you can run into a Veb
            > violation due to Vout being higher than the collector
            > voltage.
            >
            > Is this a homebrew supply? Maybe you could use some
            > sort of overcurrent protection, where you sense
            > voltage off a 0.01 ohm resistor and either blow the
            > fuse or kill the base drive when more than 40A is
            > delivered. Output shorts can be hard to protect
            > against, you may just need to add a 50A fuse for this;
            > and then perhaps a 30V TVS (25 if it doesn't get too
            > warm in practice).
            >
            > Shawn
            >
            > --- luthjej <luthjej@...> wrote:
            >
            > > Hi,
            > > I'm now on my third set of TIP3055's (8 of them
            > > per set) - as I
            > > have killed the previous 2 sets. Bascially I do
            > > something stupid (like
            > > apply a reverse polarity to the output or short out
            > > random pieces of
            > > the circuit).
            > >
            > > The symptom of the NPN's is that something internal
            > > shorts out and
            > > starts dumping volts from the supply leg (C) to the
            > > base (B), which
            > > turns all of the NPN's on ... making it impossible
            > > to tell which is
            > > the culprit, making it easier just to pull the whole
            > > lot off and
            > > replace then with new ones.
            > >
            > > There are 8 NPN's in total in this variable PSU
            > > circuit, the output
            > > legs connected via 0.1ohm 5W resistors to common
            > > rails which supply
            > > the current.
            > >
            > > Is there any way to prevent this from happening?
            > > It's pretty high
            > > current (40A @ 25VDC), so I'm not convinced that a
            > > fast-blow fuse
            > > would protect from this sort of fault.
            > >
            > > Not convinced excess gate voltage is the culprit
            > > either, in fact I'm
            > > nearly convinced that it isn't.
            > >
            > > Any ideas? A strategically placed diode / zener? One
            > > fuse per NPN?
            > >
            > > Give up on bi-polars and go FET instead?
            > >
            > >
            > >
            > >
            > >
            > >
            > >
            >
            >
            > Shawn Upton, KB1CKT
            >
            > __________________________________________________
            > Do You Yahoo!?
            > Tired of spam? Yahoo! Mail has the best spam protection around
            > http://mail.yahoo.com
            >
          • Shawn Upton
            Well, if you are using up to 40A to charge a battery, I ll wager that those batteries can source 250A easily. More than enough to blow most any transistor
            Message 5 of 10 , Jul 3, 2006
            • 0 Attachment
              Well, if you are using up to 40A to charge a battery,
              I'll wager that those batteries can source 250A
              easily. More than enough to blow most any transistor
              your or I would want to pay money for.

              So, when you reverse battery'd the thing, with no
              current limit, you had 30V - (-15) or about 45V, at
              40A+. Maybe way more than that. But a very large
              power for those NPN's.

              It sounds like you have enough voltage here; you may
              want a series output diode, a very large one at that.
              I don't know enough about power supply design, so I
              cannot recommend putting one in each emitter leg, as
              it the diodes will probably have the same negative
              tempco as the Vbe of the transistors, and may cause
              one transistor to hog current, even with the emitter
              ballasts. You'll have to do the math on it, and make
              sure that it is well balanced if you do that (which
              may be more preferable than finding a 40A schotkey);
              you may need bigger ballast resistors.

              I think your problem is as your described them: it
              works fine until you abuse the supply. I'm not sure
              how you are regulating the supply; is it a typical
              LM723 design? IIRC, that IC does have current sense
              capability; you may be just fine using the voltage
              generated across 1 ballast resistor instead of using a
              new current sense one. Afterall, if the ballasts are
              working properly, then each one should be seeing the
              same fraction of current.

              You may be surprised at how fast these are blowing in
              practice; you can blow IC's from heat dissapation
              before you can feel the heat. It may actually be
              doing ok for 100ms, which is long enough for a fuse to
              blow or the LM723 to respond.

              Shawn

              --- luthjej <luthjej@...> wrote:

              > Hi Shawn,
              > Thanks for the reply, a few interesting bits of
              > info there I wasn't
              > aware of. This is a homebrew 3-stage battery
              > charger, so no inductive
              > loads to deal with!
              >
              > Sorry about the title ... I was halfway into the
              > post when I realized
              > that a TIP3055 is actually bipolar, not FET .. :)
              >
              > Heatsinking is very adequate for these, they are
              > mounted on 2 x fan
              > forced heatsinks (4 on each) at 0.30C/W each half.
              > Total supply
              > current (out) is up to 15v @ 40A, bridge supply is
              > 25V, the base of
              > the NPN's is driven via a 30V rail delivered from a
              > voltage multiplier.
              >
              > The "problem" as such is not in the operation of the
              > supply, just in
              > me doing stupid things (the last "stupid" thing was
              > connecting a
              > battery in reverse polarity) which caused the
              > failure of the last
              > batch of NPN's), so really I guess my question is
              > more in the
              > protection of the NPN's rather than destruction via
              > overloading.
              >
              > There is also thermal protection on the heatsinks
              > which shuts down
              > drive to the transistors if the temp goes over
              > 90degC.
              >
              > I've incorporated reverse polarity protection by way
              > of a diode
              > network on the output which will start conducting
              > (and theoretically
              > trip the breaker) if the output is connected in
              > reverse ... but sadly
              > this doesn't seem to be the case.
              >
              > I have a 50amp thermal circuit breaker inline with
              > the output - which
              > didn't trip, I realize that these thermals can be
              > quite slow to react.
              >
              > So you are suggesting that a 50A wire or ceramic
              > fuse (HRC?) will blow
              > faster and give better protection? I'd rather change
              > a fuse than
              > half-a-dozen NPN's! :)
              >
              > These NPN's just seem to be very fragile -
              > especially given their
              > rated instantaneous current and wattage rating, and
              > considering I'm
              > not dealing with anywhere near their rated voltage &
              > it would be nice
              > to have some sort of fail-safe to protect against
              > idiots (like me!)
              >
              > Thanks again,
              >
              > Jon
              >

              Shawn Upton, KB1CKT

              __________________________________________________
              Do You Yahoo!?
              Tired of spam? Yahoo! Mail has the best spam protection around
              http://mail.yahoo.com
            • Daniel Nicoson
              Jon, I don t think bi-polars are very good in parallel. I think Shawn mentioned that in his first response. If you are pulling significant current I d bet
              Message 6 of 10 , Jul 3, 2006
              • 0 Attachment
                Jon,

                I don't think bi-polars are very good in parallel. I think Shawn mentioned
                that in his first response. If you are pulling significant current I'd bet
                you won't get it to work.

                I'm currently working on a switch mode power supply that pulls about 20 amps
                through two mosfets in parallel. I was actually doing it with one mosfet
                but was right on the margin and know my cooling situation wouldn't always be
                ideal. It is a two phase unit so right now I have 4 mosfets in place. That
                starts to get expensive when you make them emit magic smoke! (I've done it
                a couple times, mostly from stupid moves...)

                If you are serious about the heavy current you mention, I would recommend
                taking the time to learn about power mosfets. They have a much lower
                voltage drop across them when under load which means more efficiency and
                higher load per unit. RFP50N06 is the mosfet I am using for my project. I
                think I pay less than $1.50 each and I am counting on them for up to 20 amps
                each (they are "rated" for 50 amps).

                Check out the files section, there is a good write-up on how to determine
                what current your TO-220 component can really handle when you factor in the
                heat sinks.

                If you post a schematic, this list can be very helpful because then they can
                fully understand what you are trying to do.

                Good luck,

                Dan Nicoson

                -----Original Message-----
                From: Electronics_101@yahoogroups.com
                [mailto:Electronics_101@yahoogroups.com]On Behalf Of Shawn Upton
                Sent: Monday, July 03, 2006 10:41 AM
                To: Electronics_101@yahoogroups.com
                Subject: Re: [Electronics_101] Re: Death to MOSFET's - prevention?

                Well, if you are using up to 40A to charge a battery,
                I'll wager that those batteries can source 250A
                easily. More than enough to blow most any transistor
                your or I would want to pay money for.

                So, when you reverse battery'd the thing, with no
                current limit, you had 30V - (-15) or about 45V, at
                40A+. Maybe way more than that. But a very large
                power for those NPN's.

                It sounds like you have enough voltage here; you may
                want a series output diode, a very large one at that.
                I don't know enough about power supply design, so I
                cannot recommend putting one in each emitter leg, as
                it the diodes will probably have the same negative
                tempco as the Vbe of the transistors, and may cause
                one transistor to hog current, even with the emitter
                ballasts. You'll have to do the math on it, and make
                sure that it is well balanced if you do that (which
                may be more preferable than finding a 40A schotkey);
                you may need bigger ballast resistors.

                I think your problem is as your described them: it
                works fine until you abuse the supply. I'm not sure
                how you are regulating the supply; is it a typical
                LM723 design? IIRC, that IC does have current sense
                capability; you may be just fine using the voltage
                generated across 1 ballast resistor instead of using a
                new current sense one. Afterall, if the ballasts are
                working properly, then each one should be seeing the
                same fraction of current.

                You may be surprised at how fast these are blowing in
                practice; you can blow IC's from heat dissapation
                before you can feel the heat. It may actually be
                doing ok for 100ms, which is long enough for a fuse to
                blow or the LM723 to respond.

                Shawn

                --- luthjej <luthjej@...> wrote:

                > Hi Shawn,
                > Thanks for the reply, a few interesting bits of
                > info there I wasn't
                > aware of. This is a homebrew 3-stage battery
                > charger, so no inductive
                > loads to deal with!
                >
                > Sorry about the title ... I was halfway into the
                > post when I realized
                > that a TIP3055 is actually bipolar, not FET .. :)
                >
                > Heatsinking is very adequate for these, they are
                > mounted on 2 x fan
                > forced heatsinks (4 on each) at 0.30C/W each half.
                > Total supply
                > current (out) is up to 15v @ 40A, bridge supply is
                > 25V, the base of
                > the NPN's is driven via a 30V rail delivered from a
                > voltage multiplier.
                >
                > The "problem" as such is not in the operation of the
                > supply, just in
                > me doing stupid things (the last "stupid" thing was
                > connecting a
                > battery in reverse polarity) which caused the
                > failure of the last
                > batch of NPN's), so really I guess my question is
                > more in the
                > protection of the NPN's rather than destruction via
                > overloading.
                >
                > There is also thermal protection on the heatsinks
                > which shuts down
                > drive to the transistors if the temp goes over
                > 90degC.
                >
                > I've incorporated reverse polarity protection by way
                > of a diode
                > network on the output which will start conducting
                > (and theoretically
                > trip the breaker) if the output is connected in
                > reverse ... but sadly
                > this doesn't seem to be the case.
                >
                > I have a 50amp thermal circuit breaker inline with
                > the output - which
                > didn't trip, I realize that these thermals can be
                > quite slow to react.
                >
                > So you are suggesting that a 50A wire or ceramic
                > fuse (HRC?) will blow
                > faster and give better protection? I'd rather change
                > a fuse than
                > half-a-dozen NPN's! :)
                >
                > These NPN's just seem to be very fragile -
                > especially given their
                > rated instantaneous current and wattage rating, and
                > considering I'm
                > not dealing with anywhere near their rated voltage &
                > it would be nice
                > to have some sort of fail-safe to protect against
                > idiots (like me!)
                >
                > Thanks again,
                >
                > Jon
                >

                Shawn Upton, KB1CKT

                __________________________________________________
                Do You Yahoo!?
                Tired of spam? Yahoo! Mail has the best spam protection around
                http://mail.yahoo.com



                Yahoo! Groups Links
              • Shawn Upton
                Not true. While it is harder to parallel bipolars, remember, it was done for years. All your high current linear regulated power supplies did this for a few
                Message 7 of 10 , Jul 4, 2006
                • 0 Attachment
                  Not true. While it is harder to parallel bipolars,
                  remember, it was done for years. All your high
                  current linear regulated power supplies did this for a
                  few decades; and it is still being done. Typically,
                  up to 5A you'll see a single 2N3055; 5 to 10A or so,
                  it will be a pair of 2N3055's, with 1 2N3055 driving
                  the output pair. [With a beta of 10 no higher than
                  20, base current gets pretty high!]

                  10A supply I have here, has I think (8) TO-3 NPN's in
                  it. I think 2 are drivers leaving 6 as the output
                  transistors. The ballast resistors look like nichrome
                  wire (not sure, they certainly aren't "commercial"
                  resistors).

                  Anyhow, while it is harder and requires more parts,
                  the point is, it is very possible to do. The emitter
                  ballast resistors makes it possible; and in a linear
                  regulated supply, the ballasts really don't impact
                  efficency (since it's not that high to begin with).

                  Also, since this is a linear regulated supply,
                  thoughts of efficency are moot. The whole point is to
                  drop voltage across the regulator. I don't know what
                  the saturation voltage of a 2N3055 is; probably no
                  better than a half volt. But, since he is turning 30V
                  into 15V, well, there will always be at least 14V
                  across the NPN (minus ballast drop, which I think was
                  0.5V in this design). A typical linear supply
                  wouldn't use this high of an unregulated voltage; it's
                  high for a 15V output. But it has to be a couple of
                  volts higher than the output level at the minimum AC
                  input voltage--if the ac input line droops under load,
                  that droopage shows up after the rectifier; and it
                  better still be high enough for the supply to regulate
                  to what is needed.

                  Now, SMPS, yes, bipolars are quite hard to use. At
                  high voltage and slower switching speeds, bipolars can
                  make sense though; 10-20kHz and several hundred volts
                  (boost design), one might find a bipolar. These
                  designs can tolerate the high saturation voltage of a
                  bipolar, as the real problem is high voltage seen at
                  the collector. The high saturation voltage is not
                  nearly as much of a loss as it would be in a 5V
                  booster.

                  Shawn

                  --- Daniel Nicoson <A6intruder@...> wrote:

                  > Jon,
                  >
                  > I don't think bi-polars are very good in parallel.
                  > I think Shawn mentioned
                  > that in his first response. If you are pulling
                  > significant current I'd bet
                  > you won't get it to work.
                  >
                  > I'm currently working on a switch mode power supply
                  > that pulls about 20 amps
                  > through two mosfets in parallel. I was actually
                  > doing it with one mosfet
                  > but was right on the margin and know my cooling
                  > situation wouldn't always be
                  > ideal. It is a two phase unit so right now I have 4
                  > mosfets in place. That
                  > starts to get expensive when you make them emit
                  > magic smoke! (I've done it
                  > a couple times, mostly from stupid moves...)
                  >
                  > If you are serious about the heavy current you
                  > mention, I would recommend
                  > taking the time to learn about power mosfets. They
                  > have a much lower
                  > voltage drop across them when under load which means
                  > more efficiency and
                  > higher load per unit. RFP50N06 is the mosfet I am
                  > using for my project. I
                  > think I pay less than $1.50 each and I am counting
                  > on them for up to 20 amps
                  > each (they are "rated" for 50 amps).
                  >
                  > Check out the files section, there is a good
                  > write-up on how to determine
                  > what current your TO-220 component can really handle
                  > when you factor in the
                  > heat sinks.
                  >
                  > If you post a schematic, this list can be very
                  > helpful because then they can
                  > fully understand what you are trying to do.
                  >
                  > Good luck,
                  >
                  > Dan Nicoson
                  >

                  Shawn Upton, KB1CKT

                  __________________________________________________
                  Do You Yahoo!?
                  Tired of spam? Yahoo! Mail has the best spam protection around
                  http://mail.yahoo.com
                • Stefan Trethan
                  ... Quite frequently actually, in PC supplies for example. ST
                  Message 8 of 10 , Jul 4, 2006
                  • 0 Attachment
                    On Tue, 04 Jul 2006 15:00:22 +0200, Shawn Upton <kb1ckt@...> wrote:

                    >
                    > Now, SMPS, yes, bipolars are quite hard to use. At
                    > high voltage and slower switching speeds, bipolars can
                    > make sense though; 10-20kHz and several hundred volts
                    > (boost design), one might find a bipolar.


                    Quite frequently actually, in PC supplies for example.

                    ST
                  • Shawn Upton
                    Really? I wouldn t have guessed--but I ll guess it s a cost move. That or the topology isn t a simple buck. Cost is more important than efficency, I
                    Message 9 of 10 , Jul 4, 2006
                    • 0 Attachment
                      Really? I wouldn't have guessed--but I'll guess it's
                      a cost move. That or the topology isn't a simple
                      buck. Cost is more important than efficency, I
                      suppose, in these apps.

                      Shawn

                      --- Stefan Trethan <stefan_trethan@...> wrote:

                      > On Tue, 04 Jul 2006 15:00:22 +0200, Shawn Upton
                      > <kb1ckt@...> wrote:
                      >
                      > >
                      > > Now, SMPS, yes, bipolars are quite hard to use.
                      > At
                      > > high voltage and slower switching speeds, bipolars
                      > can
                      > > make sense though; 10-20kHz and several hundred
                      > volts
                      > > (boost design), one might find a bipolar.
                      >
                      >
                      > Quite frequently actually, in PC supplies for
                      > example.
                      >
                      > ST
                      >
                      >
                      >
                      > Yahoo! Groups Links
                      >
                      >
                      > Electronics_101-unsubscribe@yahoogroups.com
                      >
                      >
                      >
                      >
                      >
                      >


                      Shawn Upton, KB1CKT

                      __________________________________________________
                      Do You Yahoo!?
                      Tired of spam? Yahoo! Mail has the best spam protection around
                      http://mail.yahoo.com
                    • Stefan Trethan
                      ... The older PC supplies are a simple flyback, but the newer ones are a more complicated push/pull type drive with two transistors. I ve seen bipolar in both
                      Message 10 of 10 , Jul 5, 2006
                      • 0 Attachment
                        On Wed, 05 Jul 2006 04:06:36 +0200, Shawn Upton <kb1ckt@...> wrote:

                        > Really? I wouldn't have guessed--but I'll guess it's
                        > a cost move. That or the topology isn't a simple
                        > buck. Cost is more important than efficency, I
                        > suppose, in these apps.
                        > Shawn


                        The older PC supplies are a simple flyback, but the newer ones are a more
                        complicated push/pull type drive with two transistors.
                        I've seen bipolar in both much more often than FETs. OTOH small switchers
                        like in TVs and stuff tend to have a FET in the flyback switcher.

                        Not sure why, just an observation.

                        ST
                      Your message has been successfully submitted and would be delivered to recipients shortly.