OT: S-100 card power -- voltage regulators in parallel? - CP/M

This is a discussion on OT: S-100 card power -- voltage regulators in parallel? - CP/M ; If this is the wrong group to post S-100 hardware related questions in, please forgive me and advise me of the correct group. Thank you. I have an operational 4 card S-100 system running CP/M. As you know, each S-100 ...

+ Reply to Thread
Page 1 of 2 1 2 LastLast
Results 1 to 20 of 28

Thread: OT: S-100 card power -- voltage regulators in parallel?

  1. OT: S-100 card power -- voltage regulators in parallel?


    If this is the wrong group to post S-100 hardware related questions
    in, please forgive me and advise me of the correct group. Thank you.

    I have an operational 4 card S-100 system running CP/M.

    As you know, each S-100 card has 1 or more linear voltage
    regulators. For specific reasons not important to my question, I
    plan to disable all 5v regulators on these boards, and to build
    another board which will, among other things, provide 5v to each of
    the other 4 boards via dedicated wiring. 8v will be provided to this
    new board directly from the 8v supply (not from the S-100 bus).

    My system uses 4.4amps of 5v. I plan for 6 amps.

    My question is -- can I use six (6) 7805 regulators in parallel to
    produce 6A capacity? Load balancing would be achieved by a small .1-.
    47ohm resistor in series to the IN of each regulator. Each 7805
    input + output would have its own 1-2uf tantalum cap for noise
    reduc. All outputs would be commoned to 600uf electrolytic cap.

    That's pretty much all I have in mind at the moment, and I hope it can
    be this simple.

    Thank you!
    ~ J


  2. Re: OT: S-100 card power -- voltage regulators in parallel?

    MdntTrain wrote:

    (snip)

    > My question is -- can I use six (6) 7805 regulators in parallel to
    > produce 6A capacity? Load balancing would be achieved by a small .1-.
    > 47ohm resistor in series to the IN of each regulator. Each 7805
    > input + output would have its own 1-2uf tantalum cap for noise
    > reduc. All outputs would be commoned to 600uf electrolytic cap.


    The data sheet for the regulator explains how to use it for
    more current. I believe the usual way is to drive a pass
    transistor, though maybe with something other than the 7805.

    I haven't looked recently so I can't directly answer, but
    I know it is there. See: http://www.datasheets.org.uk/

    -- glen

    -- glen


  3. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Aug 15, 4:25 pm, glen herrmannsfeldt wrote:
    > MdntTrain wrote:
    >
    > (snip)
    >
    > > My question is -- can I use six (6) 7805 regulators in parallel to
    > > produce 6A capacity? Load balancing would be achieved by a small .1-.
    > > 47ohm resistor in series to the IN of each regulator. Each 7805
    > > input + output would have its own 1-2uf tantalum cap for noise
    > > reduc. All outputs would be commoned to 600uf electrolytic cap.

    >
    > The data sheet for the regulator explains how to use it for
    > more current. I believe the usual way is to drive a pass
    > transistor, though maybe with something other than the 7805.
    >
    > I haven't looked recently so I can't directly answer, but
    > I know it is there. See:http://www.datasheets.org.uk/
    >
    > -- glen
    >
    > -- glen


    Hi, Glenn. Thanks. I've read the datasheet. These devices are quite
    flexible.. and I've seen a variety of solutions posted on the net by
    users including the one I mentioned.. but I'm really looking for more
    solid opinions on whether the particular configuration I've described
    will work or not.

    ~ J


  4. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Aug 15, 4:49 pm, MdntTrain wrote:
    > If this is the wrong group to post S-100 hardware related questions
    > in, please forgive me and advise me of the correct group. Thank you.
    >
    > I have an operational 4 card S-100 system running CP/M.
    >
    > As you know, each S-100 card has 1 or more linear voltage
    > regulators. For specific reasons not important to my question, I
    > plan to disable all 5v regulators on these boards, and to build
    > another board which will, among other things, provide 5v to each of
    > the other 4 boards via dedicated wiring. 8v will be provided to this
    > new board directly from the 8v supply (not from the S-100 bus).
    >
    > My system uses 4.4amps of 5v. I plan for 6 amps.
    >
    > My question is -- can I use six (6) 7805 regulators in parallel to
    > produce 6A capacity? Load balancing would be achieved by a small .1-.
    > 47ohm resistor in series to the IN of each regulator. Each 7805
    > input + output would have its own 1-2uf tantalum cap for noise
    > reduc. All outputs would be commoned to 600uf electrolytic cap.
    >
    > That's pretty much all I have in mind at the moment, and I hope it can
    > be this simple.
    >
    > Thank you!
    > ~ J


    Why not just use a single high-current regulator? 6A should be no
    problem with a nice heat sink.

    mike


  5. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Wed, 15 Aug 2007 12:49:44 -0700, MdntTrain wrote:

    >
    >If this is the wrong group to post S-100 hardware related questions
    >in, please forgive me and advise me of the correct group. Thank you.
    >
    >I have an operational 4 card S-100 system running CP/M.
    >
    >As you know, each S-100 card has 1 or more linear voltage
    >regulators. For specific reasons not important to my question, I
    >plan to disable all 5v regulators on these boards, and to build
    >another board which will, among other things, provide 5v to each of
    >the other 4 boards via dedicated wiring. 8v will be provided to this
    >new board directly from the 8v supply (not from the S-100 bus).
    >
    >My system uses 4.4amps of 5v. I plan for 6 amps.
    >
    >My question is -- can I use six (6) 7805 regulators in parallel to
    >produce 6A capacity? Load balancing would be achieved by a small .1-.
    >47ohm resistor in series to the IN of each regulator. Each 7805
    >input + output would have its own 1-2uf tantalum cap for noise
    >reduc. All outputs would be commoned to 600uf electrolytic cap.
    >
    >That's pretty much all I have in mind at the moment, and I hope it can
    >be this simple.


    No. the design of the regulators will not assure good regulation and
    no current hogging. Adding the resistors makes that worse.

    Either use an external pass transistor (or two) or find a suitable
    higher current regulator.

    Allison

    >
    >Thank you!
    >~ J



  6. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Aug 15, 8:42 pm, MikeS wrote:
    > Why not just use a single high-current regulator? 6A should be no
    > problem with a nice heat sink.
    >
    > mike


    Mike, at your suggestion, I have located a 5A one. If that single
    regulator failed, might it let the whole 8V pass and ruin all my
    cards?

    ~ J


  7. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Aug 15, 8:49 pm, no.s...@no.uce.bellatlantic.net wrote:
    > Either use an external pass transistor (or two) or find a suitable
    > higher current regulator.
    >
    > Allison


    On Aug 15, 8:49 pm, no.s...@no.uce.bellatlantic.net wrote:

    > No. the design of the regulators will not assure good regulation and
    > no current hogging. Adding the resistors makes that worse.
    >
    > Either use an external pass transistor (or two) or find a suitable
    > higher current regulator.
    >
    > Allison


    :: sigh :: I feel like I should put you on retainer, Allison. ;-)

    In the datasheet, I see the external pass transistor(s) as the
    suggested high current method. I want to understand how the circuit
    works.

    I see that up to a certain current drawn through a resistor (tied to
    their bases) by the regulator, the transistors remain in cutoff. It's
    what happens after Vbe-on is reached that I don't understand... don't
    they all then start passing 8volts to the load?

    I have a schematic for anyone interested.

    ~ J



  8. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Aug 15, 8:49 pm, no.s...@no.uce.bellatlantic.net wrote:
    > Either use an external pass transistor (or two) or find a suitable
    > higher current regulator.
    >
    > Allison


    :: sigh :: I feel I should put you on retainer, Allison. ;-)

    In the datasheet, I see the pass transistor(s) as the suggested high
    current method. I want to understand how the circuit works.

    I see that up to a certain current drawn through a resistor (tied to
    their bases) by the regulator, the transistors remain in cutoff. It's
    what happens after the Vbe-on is reached that I don't understand...
    don't they all then start passing 8volts to the load?

    I have a schematic for anyone interested.

    ~ J



  9. Re: OT: S-100 card power -- voltage regulators in parallel?

    MdntTrain wrote:
    (snip)

    > In the datasheet, I see the pass transistor(s) as the suggested high
    > current method. I want to understand how the circuit works.


    > I see that up to a certain current drawn through a resistor (tied to
    > their bases) by the regulator, the transistors remain in cutoff. It's
    > what happens after the Vbe-on is reached that I don't understand...
    > don't they all then start passing 8volts to the load?


    If it is the one I think it is, the regulator connects with its
    input on the base of the pass transistor, and its output on the
    collector of the pass transistor, which is also the output of the
    combined regulator.

    The three terminal regulator passes just enough current to keep its
    output at the specified voltage. In this case, it pulls that current
    from the base of the pass transistor, which then increases the collector
    current of the pass transistor. The output current, then, is that
    through the regulator plus that through the pass transistor. The
    pass transistor current is beta times the base current, which all
    comes from the three terminal regulator. The total, then,
    is (1+beta) times the three terminal regulator current which is
    usually up to one amp with the appropriate heat sink.

    The three terminal regulators are designed to limit the current
    to keep from overheating. I would expect as far as possible to
    fail open, though it is possible one might fail as a short.

    I think it is supposed to be that the regulator and pass transistor
    are on the same heat sink, but I am not sure about that one.

    -- glen


  10. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Thu, 16 Aug 2007 11:15:29 -0700, MdntTrain wrote:

    >On Aug 15, 8:49 pm, no.s...@no.uce.bellatlantic.net wrote:
    >> Either use an external pass transistor (or two) or find a suitable
    >> higher current regulator.
    >>
    >> Allison

    >
    >On Aug 15, 8:49 pm, no.s...@no.uce.bellatlantic.net wrote:
    >
    >> No. the design of the regulators will not assure good regulation and
    >> no current hogging. Adding the resistors makes that worse.
    >>
    >> Either use an external pass transistor (or two) or find a suitable
    >> higher current regulator.
    >>
    >> Allison

    >
    >:: sigh :: I feel like I should put you on retainer, Allison. ;-)
    >
    >In the datasheet, I see the external pass transistor(s) as the
    >suggested high current method. I want to understand how the circuit
    >works.
    >
    >I see that up to a certain current drawn through a resistor (tied to
    >their bases) by the regulator, the transistors remain in cutoff. It's
    >what happens after Vbe-on is reached that I don't understand... don't
    >they all then start passing 8volts to the load?


    No. The transistor is not a switch in this case so as more current
    passes through the resistor the more the transistor conducts, however
    if the load drops or the transistor is conducting too had the current
    through that resistor drops and by servo action the transistor stops
    working harder. Basic power supply theory.

    In any case where there is a votage regulator and it can fail
    destructively to the locif the solution is a a crowbar circuit.
    The circuit senses V+ and if it exceeds say 5.6V it triggers an SCR
    across the 5V rail to short the supply and protect the logic.

    Allison
    >
    >I have a schematic for anyone interested.
    >
    >~ J
    >



  11. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Thu, 16 Aug 2007 12:44:14 -0800, glen herrmannsfeldt
    wrote:

    >MdntTrain wrote:
    >(snip)
    >
    >> In the datasheet, I see the pass transistor(s) as the suggested high
    >> current method. I want to understand how the circuit works.

    >
    >> I see that up to a certain current drawn through a resistor (tied to
    >> their bases) by the regulator, the transistors remain in cutoff. It's
    >> what happens after the Vbe-on is reached that I don't understand...
    >> don't they all then start passing 8volts to the load?

    >
    >If it is the one I think it is, the regulator connects with its
    >input on the base of the pass transistor, and its output on the
    >collector of the pass transistor, which is also the output of the
    >combined regulator.
    >
    >The three terminal regulator passes just enough current to keep its
    >output at the specified voltage. In this case, it pulls that current
    >from the base of the pass transistor, which then increases the collector
    >current of the pass transistor. The output current, then, is that
    >through the regulator plus that through the pass transistor. The
    >pass transistor current is beta times the base current, which all
    >comes from the three terminal regulator. The total, then,
    >is (1+beta) times the three terminal regulator current which is
    >usually up to one amp with the appropriate heat sink.
    >
    >The three terminal regulators are designed to limit the current
    >to keep from overheating. I would expect as far as possible to
    >fail open, though it is possible one might fail as a short.


    Either can happen and a crowbar circuit should be used.

    >I think it is supposed to be that the regulator and pass transistor
    >are on the same heat sink, but I am not sure about that one.


    Does not have to be. Since the regulator may be running some power
    it could have a heatsink. They might as well share the heatsink. wich
    should be farily large.

    Another good addition to that circuit (often the datasheets have this)
    is an over current protect for the high current pass transistor.


    Allison
    >
    >-- glen



  12. Re: OT: S-100 card power -- voltage regulators in parallel?



    > If it is the one I think it is, the regulator connects with its
    > input on the base of the pass transistor, and its output on the
    > collector of the pass transistor, which is also the output of the
    > combined regulator.


    Pretty much, yes. The emitters are connected individually through
    0.1ohm resistors to 8V, then the bases + the regulator IN are
    connected to 8V through a 100ohm resistor.

    The outs are combined as you say.


    > The three terminal regulator passes just enough current to keep
    > its output at the specified voltage. In this case, it pulls that
    > current from the base of the pass transistor, which then
    > increases the collector current of the pass transistor. The


    Ok, I follow so far.


    > output current, then, is that through the regulator plus that
    > through the pass transistor. The pass transistor current is beta
    > times the base current, which all comes from the three terminal
    > regulator. The total, then, is (1+beta) times the three
    > terminal regulator current which is usually up to one amp.


    I follow that.

    But I still don't get how the 8v at the emitters is dropping to 5
    volts at the collectors. The 0.1 resistors are way too small to be
    dissipating the extra voltage. So is the regulator compensating
    somehow?

    ~ J



  13. Re: OT: S-100 card power -- voltage regulators in parallel?

    These regulators are designed to "fail open"; so what you described
    should not happen, but when it comes to failures, there are no absolute
    guarantees.


    MdntTrain wrote:
    > On Aug 15, 8:42 pm, MikeS wrote:
    >> Why not just use a single high-current regulator? 6A should be no
    >> problem with a nice heat sink.
    >>
    >> mike

    >
    > Mike, at your suggestion, I have located a 5A one. If that single
    > regulator failed, might it let the whole 8V pass and ruin all my
    > cards?
    >
    > ~ J
    >


  14. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Aug 16, 4:19 pm, MdntTrain wrote:
    > > If it is the one I think it is, the regulator connects with its
    > > input on the base of the pass transistor, and its output on the
    > > collector of the pass transistor, which is also the output of the
    > > combined regulator.

    >
    > Pretty much, yes. The emitters are connected individually through
    > 0.1ohm resistors to 8V, then the bases + the regulator IN are
    > connected to 8V through a 100ohm resistor.
    >
    > The outs are combined as you say.
    >
    > > The three terminal regulator passes just enough current to keep
    > > its output at the specified voltage. In this case, it pulls that
    > > current from the base of the pass transistor, which then
    > > increases the collector current of the pass transistor. The

    >
    > Ok, I follow so far.
    >
    > > output current, then, is that through the regulator plus that
    > > through the pass transistor. The pass transistor current is beta
    > > times the base current, which all comes from the three terminal
    > > regulator. The total, then, is (1+beta) times the three
    > > terminal regulator current which is usually up to one amp.

    >
    > I follow that.
    >
    > But I still don't get how the 8v at the emitters is dropping to 5
    > volts at the collectors. The 0.1 resistors are way too small to be
    > dissipating the extra voltage. So is the regulator compensating
    > somehow?
    >
    > ~ J



    Hi
    The method described doesn't work.
    Typically, multiple regulators are balanced by placing them on the
    same
    thermal heat sink but the output drives go to different sections of
    the load.
    This way, the resistance of the traces on the board tend to help
    balance the
    loads.
    If the load is constant enough, you can balance with resistors on the
    outputs of the regulators. The only problem here is if the load goes
    down too much, the voltage will go up as well. It sounds like you
    have enough load that this may be practical.
    Another option is to build a sensed regulator circuit and not use
    three terminal regulators at all.
    Yet another option might be to put a current sense circuit on the
    inputs
    to the regulators. This could cause the ground lead of the regulator
    to
    go below ground when that regulator exceeded some specific amount
    of current.
    I think such a circuit could be made with 5 resistors, 2 diodes, a
    NPN and
    a PNP transistor.
    Yet another way that might work with the normal tolerances is to
    place a
    diode in series with the output of each regulator. Also place a diode
    in
    series with the ground lead of the regulator. Place a resonable sized
    resistor from the input to the regulator to the reglator ground lead,
    to bias the diode.
    Isolate the output diode thermally from the regulator but make sure
    it is well heat sunk. Place the diode on the ground lead of the
    regulator
    such that it shares the heat of the regulator. It is better to mount
    it right
    on the regulator so that it sees the heat more directly.
    The one in series, with the output, will need to be a high current
    one ( at least
    3 amp ) but the one on the ground could be a 1n4148 type small signal
    diode.
    The resistor to bias the signal diode can be optimized but something
    on the
    order of 1500 to 10K ohms should be enough.
    Dwight


  15. Re: OT: S-100 card power -- voltage regulators in parallel?

    MdntTrain wrote:
    (snip on regulators)

    > But I still don't get how the 8v at the emitters is dropping to 5
    > volts at the collectors. The 0.1 resistors are way too small to be
    > dissipating the extra voltage. So is the regulator compensating
    > somehow?


    The right amount of base current is supplied such that the collector
    voltage is 5V. It is a negative feedback system. If the load increases
    the voltage would go down. The regulator pulls more current through
    the base, increasing the current through the transistor the exact amount
    to get the collector back to 5V. The power dissipated in the transistor
    is the voltage drop (8V-5V in your case) times the collector current
    (most of the load current).

    That is also pretty much what happens inside the regulator. There is
    a big transistor from the input to the output (emitter to collector)
    and circuitry to compare the voltage between the output and ground,
    and adjust the base current accordingly. A few milliamps go through
    the ground terminal to run the voltage reference and comparison
    circuit. There is a very small resistor on the output used for the
    overcurrent protection, and some thermal sensing shutdown circuitry,
    also.

    -- glen


  16. Re: OT: S-100 card power -- voltage regulators in parallel?

    no.spam@no.uce.bellatlantic.net wrote:
    (I wrote)

    >>I think it is supposed to be that the regulator and pass transistor
    >>are on the same heat sink, but I am not sure about that one.


    > Does not have to be. Since the regulator may be running some power
    > it could have a heatsink. They might as well share the heatsink. wich
    > should be farily large.


    I was thinking it was needed so that the thermal protection
    will include the pass transistor, but I am still not sure either
    way.

    -- glen


  17. Re: OT: S-100 card power -- voltage regulators in parallel?

    MdntTrain wrote:
    (snip)

    > But I still don't get how the 8v at the emitters is dropping to 5
    > volts at the collectors. The 0.1 resistors are way too small to be
    > dissipating the extra voltage. So is the regulator compensating
    > somehow?


    I didn't figure out your question until I read Allison's post.

    It is not normal that transistors are switches. That is the way
    they are used in digital logic, but not, for example, a stereo
    amplified. (Except for rare Class D amps.) When not saturated,
    the collector current is approximately proportional to the base
    current, with the constant being called beta. That is, the
    current gain in common emitter configuration. The circuit
    adjusts the base current such that the collector voltage
    (relative to ground) is 5V.

    -- glen


  18. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Aug 17, 2:10 am, glen herrmannsfeldt wrote:

    > I didn't figure out your question until I read Allison's post.
    >
    > It is not normal that transistors are switches. That is the way
    > they are used in digital logic, but not, for example, a stereo
    > amplified.


    I see, Glen. I normally work with tubes (audio) and haven't done much
    with transistors except recently using a Darlington as a large current
    *switch*. I'm having a rough go of the mental transition -- tubes
    make much more sense to me.. and I rarely if ever would dare to draw
    current out the grid (base equiv) of a tube! :-D


    > When not saturated,
    > the collector current is approximately proportional to the base
    > current, with the constant being called beta. That is, the current gain
    > in common emitter configuration.


    Yep, I've gotten that much so far. That is a weird concept as I'm
    used to controlling tube current with grid *voltage*. I thought
    transistors were only voltage controlled too.


    > The circuit adjusts the base current such that the collector voltage
    > (relative to ground) is 5V.


    It's embarassing, but it's SLOWLY starting to dawn on me how this is
    working. Tell me this: if the collector weren't tied to the
    regulator's 5v out (really to ground through the load of the S-100
    card (about 2-3 ohms))... but instead to like a 8kohm resistor, the
    collector voltage WOULD the 8volts - a forward voltage drop, isn't
    that right?

    ~ J



  19. Re: OT: S-100 card power -- voltage regulators in parallel?

    On Aug 16, 6:48 pm, Barry Watzman wrote:
    > These regulators are designed to "fail open"; so what you described
    > should not happen, but when it comes to failures, there are no absolute
    > guarantees.


    I've heard that said about life, too, Barry.. ;-)

    ~ J


  20. Re: OT: S-100 card power -- voltage regulators in parallel?

    MdntTrain wrote:
    (snip)

    > I see, Glen. I normally work with tubes (audio) and haven't done much
    > with transistors except recently using a Darlington as a large current
    > *switch*. I'm having a rough go of the mental transition -- tubes
    > make much more sense to me.. and I rarely if ever would dare to draw
    > current out the grid (base equiv) of a tube! :-D

    (snip)

    Tubes are somewhat similar to FETs, but very different from
    bipolar transistors. If you add a resistor, though, the
    base voltage can vary with an input current. Tubes don't
    make a good analogy for a PNP transistor, though.

    > Yep, I've gotten that much so far. That is a weird concept as I'm
    > used to controlling tube current with grid *voltage*. I thought
    > transistors were only voltage controlled too.


    (snip)

    Bipolar transistors are pretty much current controlled current
    sources.

    > It's embarassing, but it's SLOWLY starting to dawn on me how this is
    > working. Tell me this: if the collector weren't tied to the
    > regulator's 5v out (really to ground through the load of the S-100
    > card (about 2-3 ohms))... but instead to like a 8kohm resistor, the
    > collector voltage WOULD the 8volts - a forward voltage drop, isn't
    > that right?


    The quiescent current is about 5ma. The data sheet I am looking at
    has a 3 ohm resistor between the base and emitter, which will take
    about 0.7V/3 ohms, or 230ma. That should be enough to keep the
    transistor in the linear region, even with an 8K load.
    There might still be some leakage current, but I think 8K is
    small enough.

    There is also a high current circuit with an extra transistor and
    resistor to limit the short circuit current. The regulator protects
    itself, but not the pass transistor.

    http://www.datasheets.org.uk/datashe...article=623405

    For a circuit diagram of the inside of the regulator, see:

    http://www.datasheets.org.uk/datashe...article=623408

    -- glen


+ Reply to Thread
Page 1 of 2 1 2 LastLast