Creating a high current LM317 regulator

Discussion in 'The Projects Forum' started by Patrick Joseph, Feb 25, 2015.

  1. Patrick Joseph

    Thread Starter New Member

    Feb 23, 2015
    11
    0
    Last year I built a bicycle powered generator as a demo at a kids science fair using a permanent magnet motor as the generator. The generator powered an inverter so that the kids could get an idea of how much more power is needed to power an incandescent light bulb than an LED or CFL bulb.

    To keep the voltage from rising above 13.8 volts I used an LM317 regulator circuit. I found the circuit design on a renewable energy web site in the U.K.
    This is the circuit that I used. It worked well. The darlington transistor carried most of the current but needed a substantial heat sink and fan when the power consumption increased.

    http://www.reuk.co.uk/LM317-High-Current-Voltage-Regulator.htm
    [​IMG]


    I found another circuit on the web site that they claim will handle much more current and uses 2N3055 transistors. These would be able to handle more current and dissipate the heat better.

    http://www.reuk.co.uk/LM317-Adjustable-Power-Supply.htm

    [​IMG]

    They suggest that more transistors can be added in parallel to increase the current capacity but the circuit is different between the LM317 out and the base of the transistors. I'm not sure why they are different.

    [​IMG]


    I did a circuit simulation using "Do Circuits" and neither one of these will work. The one at the top that I used last year works fine in the simulation but these do not. I haven't bread boarded the circuit yet, but maybe I should try.

    Does anyone think that these should work?
    If not, would you have any suggestions on something different?

    Thanks
     
  2. Papabravo

    Expert

    Feb 24, 2006
    10,135
    1,786
    A transistor can't work with one of it's terminals open circuited. They just don't work that way. Try starting by putting a typical load resistor in the simulation. The name of the connections is Common Collector, or Emitter Follower. The load resistor is essential.
     
  3. Patrick Joseph

    Thread Starter New Member

    Feb 23, 2015
    11
    0
    Understood. During the simulation a load is added. I didn't try running the circuit with an open circuit at the output. I also added a source of 16 VDC and a volt meter at the load to see what the output voltage would be. It wouldn't handle a large load (100 Ω or so) without exceeding the parameters.
     
  4. ronv

    AAC Fanatic!

    Nov 12, 2008
    3,281
    1,232
    The one on top should work - kind of. It is not a very good regulator, but you should get voltage out.
    The second one won't give much current because of where the 220 ohm resistor is located.
    You will still need a big heat sink even though you are using bigger transistors. The same power is being "burned" in the transistors.
     
  5. Patrick Joseph

    Thread Starter New Member

    Feb 23, 2015
    11
    0
    Thanks. I've tried it in the simulator, but not Real World. I'm almost leaning toward the circuit that I had used last year but use a couple of larger transistors connected in the darlington pair configuration.
     
  6. Reloadron

    Active Member

    Jan 15, 2015
    963
    232
    On another note when you parallel transistors like the 2N3055 in the configuration you have, you want about .1 to .5 Ohm resistors in each emitter leg. The problem as drawn is all of the 2N3055 transistors will not conduct equally some will conduct a little harder and faster than their counterparts, they will get hot faster and create a thermal runaway situation. To prevent this you place resistors in each emitter leg to balance the current load. Finally per ronv, not a very good design. If you use the design then I suggest emitter resistors.

    Ron
     
    pilko likes this.
  7. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,515
    1,246
    The first circuit is the classic, and will regulate better than either of the other two. If you check different LM317 data sheets, you'll find circuits with multiple power transistors instead of just one. The second circuit will work, but the design point for the LM317 has to be adjusted to compensate for the two diode drops between the 317's regulation point and the circuit output. Also, regulation will be fair to poor depending on the range of output currents needed. The third circuit probably will not work because messing with the currents around pin 1 seriously change how the 317 behaves.

    Bill Bowden has a circuit that is somewhere between your #2 and #3, and includes feedback to bring the power transistors inside the control loop for much better regulation. It is the 2nd circuit on this page: http://www.bowdenshobbycircuits.info/page12.htm#317pass.gif

    ak
     
  8. Dodgydave

    Distinguished Member

    Jun 22, 2012
    4,965
    744
    wayneh likes this.
  9. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    The voltage regulation on the emitter-follower circuits is horrible.
    The total power dissipation is the same unless you use a switch-mode supply.
    Here is what I have used, but leaving out the capacitors required for stability, and the reverse voltage protection diodes:

    Note the voltage regulation and power dissipation in each component plotted as a function of the load current.

    18c.gif
     
  10. WBahn

    Moderator

    Mar 31, 2012
    17,715
    4,788
    Patrick Joseph likes this.
  11. Patrick Joseph

    Thread Starter New Member

    Feb 23, 2015
    11
    0
    What is your opinion of the third circuit? If that can be utilized as a high current regulator with a DC source at the point where the bridge rectifier connects, then I could use the 2N3055 transistors that I have already purchased. And I think I can replace the 1K resistor with a pot to make it adjustable.
     
  12. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,515
    1,246
    The third circuit in your original post will not work. The Bowden circuit I mentioned uses 3055's and maintains regulation.

    ak
     
  13. Patrick Joseph

    Thread Starter New Member

    Feb 23, 2015
    11
    0
    I should have specified what I meant. I was talking about the third circuit on the bowdenshobbycircuits page that you suggested. It uses 2N3055 transistors.
     
  14. MikeML

    AAC Fanatic!

    Oct 2, 2009
    5,450
    1,066
    The third circuit in your posting is the one I said is Horrible.
    If the goal is to use a NPN for the bypass transistor, then turn the regulator upside down by using the negative complement to the LM317 called the LM337.

    The Bowden circuit requires a more positive input to the LM317 that you do not have.

    Here is the general idea of how to use the negative regulator with an NPN current-booster. Note that in this circuit, the circuit ground is a little different, and requires that the source be floating.

    18i.gif
     
    Last edited: Feb 26, 2015
  15. AnalogKid

    Distinguished Member

    Aug 1, 2013
    4,515
    1,246
    That's the one. I've use a variation with no problems. In your case you can tie the 317 input to the bulk DC source rather than a 2nd winding.

    ak
     
  16. Patrick Joseph

    Thread Starter New Member

    Feb 23, 2015
    11
    0
    Should the collectors of the 3055's be tied together and connected to the input of the 317? And that would be where the DC in connects. Is that right?
     
  17. Lestraveled

    Well-Known Member

    May 19, 2014
    1,957
    1,215
Loading...