Idea for triple output power supply

Discussion in 'The Projects Forum' started by witssq, Dec 11, 2009.

  1. witssq

    Thread Starter Active Member

    Mar 29, 2009
    I would like to make triple output linear regulated power supply.

    +15V 1.5A(For Analog)
    -15V 1.5A(For Analog)
    5V 5A (For digital)

    Please advise me which choice is best for transformer selection.

    One transformer centertaped(ex. Parellel 18VAC) for +/- 15v
    Another transformer(ex 12VAC) for +/- 5V

    2) On big transformer centertaped(18VAC)
    One region for +15V, + 5V
    Another region for -15V (this seems to make to much loss)

    3) On big transformer centertaped(18VAC)
    One region for +15V and -15 using voltage doubler circuit
    Another region for 5V

    I think that choice 3 is best for transformer power distribution.
    But, many materials says that voltage doubler is not good.

    And, I found that dual centertaped triple output at digikey as attached.
    But rating is too low.

    Please advice for selection.

    Last edited: Dec 11, 2009
  2. studiot

    AAC Fanatic!

    Nov 9, 2007
    Choice 1 is by far the best.

    If you consistantly load only part of the winding heavily, as you will drawing 5A for the 5V supply, you will cause magnetic unbalance within the transformer. This leads undersireable effects including a phenomenon known as magnetisation. This is not a problem if you do it occasionally and randomly so the effects are balanced. It is also not a problem if one winding is very lightly loaded as in the days of valves (tubes) when the heater winding drew a substantial current, but the anode supplies drew a small current.
  3. witssq

    Thread Starter Active Member

    Mar 29, 2009
    Thanks for your advice.

    I found one sample schematic for dual supply.

    If transformer current rating 24VAC-1A(24VA), and final regulated voltage
    15V, how much current can be drawn at load?
    ex. 1A or (24(VAC)*1.414*1A / 15VDC) or any other.

    Please advice me calculation details about power supply current rating
    for 15Vx1.5A regulated DC.

  4. studiot

    AAC Fanatic!

    Nov 9, 2007
    Well firstly the diodes are rated at 1amp RMS so you would need to go for a higher rating if you required more current.

    Secondly 24 volts is too high to provide 15 volts final output. There would be excessive dissipation in the regulating device.

    24 volts yields 33.5 volts peak. the reservoir capacitors will charge up to this value. However they will fall to maybe 26 volts at the low point of the ripple waveform if you draw any significant amount of current.

    24 volts at 12- 0 -12 would make a reasonable 15 -0 - 15 DC supply up to about 0.5 amps.
    Beyond that I would go for a 30 to 36 volter (15- 0 15 or 18 - 0 -18).

    You estimate by realising that the AV power = DC power (approx) so RMS volts x RMS amps in will be roughly equal to DCvolts x DC amps out. Then it is up to you to arange the circuitry to come as close to this as possible for efficiency reasons.
  5. witssq

    Thread Starter Active Member

    Mar 29, 2009
    Thanks for your kindness.

    As to your advice, if transformer parellel 18VAC@3.12A is rated (18-0-18) and the needed regulated DC is +/-15V, the maximum regulated DC current is roughly (18*1.56/15 = 1.872A), with 14A-56-36 data sheet attached. Is this right?

    Best rgds,
    SunSung Hwang
    • 14A.pdf
      File size:
      261.8 KB
  6. studiot

    AAC Fanatic!

    Nov 9, 2007
    I'm afraid it really doesn't work like that.

    The maximum current is not determined by the transformer but by the load and the reservoir capacitor.

    You can continue to increase the secondary current draw from a transformer until well past the rated output.

    As you do this three things happen

    Firstly the core moves into magnetic saturation, reducing the extra flux generated by each increase in primary amperage.
    This results in a lower magnetic linkage in the secondary which results in lower induced emf. So the secondary output voltage drops increasingly with increasing secondary current.
    Finally the transformer heats up and may even overheat.

    The reservoir capacitor charges up to the peak of AC incoming from the transformer.
    It discharges through the load so its voltage falls.

    The difference between this cyclic peak voltage and the low voltage is called ripple voltage.

    Whilst it is easy to calculate and depict the shape and size to the ripple voltage it is more difficult to show the resulting ripple current.

    It is however important to realise that the reservoir capacitor could not be charged without it, nor the load current supplied.

    A simple analysis is to assume that the average ripple current is constant and equal to the maximum current supplied to the load. This allows calculations of ripple voltage and required reservoir capacitance based on simple formulae.

    In fact whilst the ripple voltage is very similar to a sawtooth waveform the ripple current actually occurs in short bursts of very high intensity. Curent can only flow from the transformer to the capacitor whilst the transformer voltage is greater than that of the capacitor, an event whcih only occurs for a short period in each cycle.
    So the actual peak ripple current flows may be 10 - 20 times that of the current supplied to the load.

    So start you design from the other end - what supply do you require and therefore what components do you need to achieve this - not what can I get out of these components.