Bench PSU project

Discussion in 'The Projects Forum' started by kdillinger, Mar 28, 2014.

  1. kdillinger

    Thread Starter Active Member

    Jul 26, 2009
    I do a lot of work with op-amps and I have been using a patch work of batteries to supply a dual supply on my home bench.

    I am designing a small power supply module for my bench and I have attached a PDF of a preliminary schematic. It is a transformer based design implementing an LM317 and LM337.

    Would the group mind taking a look-see? I have a resistor setting the voltage for both LM317 and LM337 at the moment, but will change this to a pot.

    The transformer is a 4:1 stepdown, but I have no identified a part number. Any suggestions here?
  2. AnalogKid

    AAC Fanatic!

    Aug 1, 2013
    Pretty straightforward stuff. If you surf for power supply schematic, you'll get a ton of very similar designs. You can save four diodes by connecting the two secondaries for form a single center-tapped secondary, and ground the center tap.

    4:1 in the US means 28.75 Vrms out, or 39V across the filter cap at no load. At high line, no load, and minimum DC output (1.24V), you are right on the hairy edge of over-voltaging the regulators. The LM317 maximum in-out differential is 43 V.

    At 1 A there will be less than 4 Vp-p of ripple across each cap, and each regulator will be dissipating almost 25 W. That's almost fan territory. Not a crisis, but something to consider then you get to the heatsinks.

    10 uF aluminum and 0.1uF ceramic caps from each adjust pin to gnd will reduce output noise, especially while adjusting.

    Depending on how much bipolar testing you do, one thing to consider is tying the two regulators together so they track with a single pot.

  3. ericgibbs

    AAC Fanatic!

    Jan 29, 2010
    If your mains is rated at 110V , this will give a Vrms at the secondary of ~28V, the Vpeak allowing for diode drops will be 37Vdc, which close to the LM317 limit.

    You also require a higher value capacitor of the input to the LM317/337 regs.

    Capacitor decoupling on the output side , say a 100nF and 470uF.

    What voltage and current ranges are you designing for.??
  4. #12


    Nov 30, 2010
    Be careful about the power line voltage. At my house, I get 125 volts. That would put your rectified voltage at 43 volts and the regulators go, "poof". Especially if your transformer is labeled for 110 volts, high power line voltages can cause trouble.

    There is an LM317H that can survive 60 volts, but I can't find a negative regulator that will go above 40V.

    It seems you are a candidate for a different transformer or a switching regulator. They don't get so hot while operating.

    ps, please uf for microfarads. The, "m" means millifarads and can be confusing.
  5. kdillinger

    Thread Starter Active Member

    Jul 26, 2009
    A lot of replies so I will summarize here.

    I can certainly crank up the turns ratio, say 6:1 as my goal is a +/-15V supply rail that can provide 100mA which is well more than enough for the time being.

    Many of the old op-amps I have here are of the LM741 variety, so they are terrible on low voltage supplies; e.g., have to watch that input and output compliance.

    If I implement a 6:1 transformer with 2 secondaries and center tap, what will the ratings of the secondary coils have to be? Roughly 40V?
  6. #12


    Nov 30, 2010
    More like 7 to 1.
    Allowing 1.4 volts for the rectifiers and 2.7 volts for the regulator chip,
    assuming you are in the USA,

    For an input voltage from 105 to 125 RMS,
    (105V/7 x sqrt2) -1.4 -17.7 = 2.11 volts of ripple allowed on the filter capacitors.
    (125/7 x sqrt2) -1.4 -17.7 = 6.15 volts of ripple allowed on the filter capacitors.

    sqrt2 C Er F = I
    .1/(1.414 2.11 120) = 279 uf minimum capacitance
    and...more than 25 volt rating. Probably 35 volts, but 50 volts will help the lifetime expectancy.

    The copper for the secondary must be rated for 162 ma.

    If my numbers are off, you can correct it because I showed my math.
    Last edited: Mar 30, 2014