Connect with us

tranformers again and again

Discussion in 'Electronic Design' started by Abstract Dissonance, Feb 6, 2006.

Scroll to continue with content
  1. Sorry for all the posts but I'm having some problems in designing my PS.

    I'm using a CT tranformer that is rated at 25.2VCT/2A(it says that on the
    transformer). I am using several voltage regulators to get differen fixed
    and variable outputs. To reduce power dissipation on the lower fixed
    voltage regulators I run them from the center tap. Now I have the 7912 which
    is a 12V negative fixed voltage regulator... since I need a few volts above
    12 to run it I'm not sure if I can run it off the CT or not... if the CT
    gives me exactly 12V or even 13V then it won't be good enough and I have to
    run it off the 25.2V which means I'm dissipating ~18W instead of about 6W(if
    I run it off the CT then it reduces the voltage by about 12V and hence that
    power isn't wasted in running the regulator). Maybe 12W isn't much different
    but I have many of these regulators and I don't want to worry to much about
    heat issues(everything will have heat sinks though).

    One more issue I'm worried about is connecting my regulators in series and
    parallel. What happens when one load ends up changing can it cause
    oscillations or drops in the other regulators?

    This is what I've done so far.

    The regulators on the LHS are the 7924, 7918, 7915, 7912 and the RH are the
    08 and 05. I'll add two variable LM317's next...

    The Bus just contains the connectors to power so it was easy to work with.

    I was thinking of running the 24-12 from the LM317 so I'd get better
    regulation and protection(since it can handle about 1.5A which is close to
    what my transformer is).

    The problem I had before about running everything in cascade was that it
    would limit the max current I could draw from stages(so I could draw 1A from
    the 4th stage and 1A from the 3rd stage cause I'd have 2A going through the
    first and second stage).

    Anyways, Thanks for any help,
  2. Robert Baer

    Robert Baer Guest

    Since the CT is at ground, you cannot use it to power anything.
    D2 and D4 power U2, U3, U4 and U5 which are positive regulators; D1
    and D3 power U6 and U7 which are negative regulators.
    From the CT, each side of the transformer is rated at 12.6V AC RMS
    "nominal" meaning as if the line voltage was 110V AC RMS.
    In most places in the US, the typical line voltage is 123V AC RMS.
    Furthermore, with a large capacitor, the average DC voltage at the
    rectifiers will be "close" to the peak which is 1.4 times the RMS,
    translating to a max near 17V and a crude estimate of the DC average
    near 15V.
    That is not enough of "headroom" for a three terminal 12V regulator,
    but plenty for an LDO (low drop out) type of regulator.
    If any desired regulated voltage needs to be larger, then consider a
    switcher; for 5V and lower - especially at large currents, then again
    consider a switcher for better efficiency.
  3. I don't follow this? The potential difference between CT the positive is
    ~12V and hence can power stuff...
    All of the voltage regulators are negative regulators... I'm not sure what
    you are getting at?

    The "ground" at U2-U5 is ~12V and the power is at -12V which gives me
    ~-12-12 =-24V to use to regulate with.

    U6 and U7 work from G-P ~= -12V and is used to regulate the 8V and 5V
    supplies(U7 actually goes from N to G but the potential difference is the
    same with it being -12-0 = -12V)

    Why isn't it enough? if its near 15V then its 3V above the 12V which means
    that it is good enough from what I have read?
    At this point in time I just need a PS that can supply +-12V and +-5 so I
    can work with my MCU's. But Since its just a matter of adding some extra
    regulators(and I have many) I might as well add some other fixed and 2
    variable just incase I need them...

    Eventually I might make or just buy a SMPS but for now I need something to
    work with that is decent... I'm not to worried about efficiency but I figure
    that I might as well do what I can do make it a little more efficient..
    (i.e., using 12 volts across the 8 and 5 regulators instead of 24V to save
    about 13W+12W = 25W of wasted power).

  4. Robert Baer

    Robert Baer Guest

    AS i read it, the CT *is* ground.
    Therefore, D2 and D4 give a *positive* voltage, and powers U2, U3, U4
    and U5 which must be positive regulators.
    And, D1 and D3 give a *negative* voltage, and powers U6 and U7 which
    must be negative regulators.
    Yes, according to the spec, 3V is the minimum that 3-terminal
    regulators work at.
    So, say there is a heavy load; that will cause a larger ripple on the
    input capacitors to the regulators than a lighter load.
    The low part of that ripple might go below the minimum for regulation.
  5. hmm... thats in volts.. *2A to get 50W wasted ;)
  6. So? Also CT is not earth ground but it shouldn't matter. For the schematic
    both sides of the transformer must be grounded so it can use probes and
    stuff. (since its done in a simulator)

    D2 and D4 might give a "postive" voltage but it is used as the ground of the

    Potential is relative. A negative regulator only needs its In terminal to be
    lower than its ground terminal... since I used the "positive side" of the
    bridge as the ground I get 12V on it... the negative terminal as -12 for a
    total of -12-12 = -24V that the negative regulator see's...

    if I saw them around and use a positive regulator I'll get 24V (but then I
    can swap its output with its relative ground to get -24V again.

    Negative and Positive regulators can be used interchanabe... just requires
    you to swap GND with In if you want to swap the polarity of it.
    Maybe... I'm not sure... hopefully the regulators work quite well over there
    maximum current rating. I don't plan on using any heavy loads I think.. I'll
    have to do some testing to find out how well it works though.

    I'm sure that having heavy loads on each regulator can cause some major
    problems but I doubt I will be using more than 2 regulators at a time.

  7. If you want to run both positive and negative regulators from the same
    transformers, and use the resulting posive and negative outputs
    simultaneously, _all_ the regulator ground terminals _must_ be
    connected to the transformer center tap - otherwise you will get short
    circuits when you try to connect the "ground" of the positive supply
    to the "ground" of the negative.

    If you just want a single supply that you can use as either negative
    or positive, you just need one regulator - either polarity will work -
    then call whichever end you wish "ground".
  8. Robert Baer

    Robert Baer Guest

    So the schematic you show is incorrect? That the CT is not ground?
    *IF* the regulators were as i stated and connected as i stated, then
    there would be no problems, as the load on each regulated polarity could
    be connected to ground, instead of floating.
    I suggest that you make a monochrome schematic and make it complete
    and accurate (include all loads and multiple level regulators).
  9. One trick I use is a voltage doubler circuit so a 12 VAC tap on a
    transformer will give about 30 VDC. You use one electrolytic cap C1 with
    negative to the transformer, and use a diode D1 cathode to C1(+), and D1
    anode to CT. Then another diode D2 anode to C1(+), and D2 cathode to another
    capacitor C2(+). Tie C2(-) to the CT, and you will have about 30 volts to CT
    (gnd). The inverse of this circuit on the other end of the transformer gives
    you -30 VDC. As you load it down, the voltage drops fairly quickly, so you
    just use capacitors that are just large enough to keep the VREG input above
    its limit at full load. It also tends to be fairly forgiving for overloads
    and short circuits. It will draw fairly high current surges from the
    transformer, however, so check for overheating at heavy loads. The true
    RMS(AC+DC) current will be much more than what you read on an ordinary

    If you really just need +/-12 VDC and +/-5 VDC for MCUs, why not find an old
    junk PC and use its switching power supply? You can probably get all you
    need for free or for a couple dollars at a swap meet or computer recycler.

    Paul E. Schoen
  10. ehsjr

    ehsjr Guest

    Not according to your schematic.

    Now I have the 7912 which
    Ignoring your supply for the moment and just looking
    at the concept: the DC output from the rectifier will be
    the ac peak voltage - about 1.4 times the AC rms voltage -
    minus two diode drops in the bridge.
    Series is OK, where a higher Vout regulator is used to
    provide Vin to a lower voltage regulator, as long as
    there's enough Vin "headroom". Parallel is not good.
    But your transformer is limited to 2A - so the scenario you
    paint above is transformer limited, not "cascade limited".
    You can get a total of 2 amps out, however you divide it.
    In theory, you could cascade the 24,18,15, all to feed into
    the 12, and you could pull 2A from it. You could not pull
    2A from it and connect another 1A load to one of the higher
    voltage regs at the same time, whether or not the thing
    was wired in cascade.
    All of the regulators shown except U7 use positive as the common.
    The powered circuit will have its ground connected to + by
    regs 1 - 6. That's fine. But - if you supply the same circuit
    simultaneously with U7, you have the CT connected to the + which
    is not good. Essentially, you are shorting the CT to the + output
    of the bridge through the regulator and the powered circuit ground.
    I would suggest that you avoid connecting the CT to anything, bite
    the bullet and dissipate the heat. And you can cascade to
    distribute the heat across multiple regs instead of just one.

    If you want to use the transformer to make a + gnd - supply,
    and the kind of supply (minus U7) in the existing diagram,
    put a switch in there to totally isolate it from the circuit
    you drew and feed it to a + gnd - regulator designed for the
    purpose. Otherwise, I fear you will "clever" yourself into
    a state of confusion.

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