Class AB BJT Darlington amplifier Explosion :(

Discussion in 'The Projects Forum' started by cariba, Apr 22, 2009.

  1. cariba

    Thread Starter Member

    Apr 22, 2009
    19
    0
    DEar all, just to ask for some advice. I am currently designing a class ab amp , to drive a 50 w speaker. I am using an 0-100v dc supply capable of supplying 10 amp. I already have done the va stages getting a p to p of 60 v . everything is fine with that . The severe problem i have is the darlington bjt complimentary pair. Every thing seems fine using a 1k resistor as load but as i try to drive an 8 ohm speaker everthing blows up:( since i am using a single voltage rail I am having a 50 v beetween the emitters and i am coupling the transistors to the speaker by a 1000uf cap. as for X over distortion I have used 4 diodes in series .

    Could some maybe help me solve my problem please. I tried using tip 137 and 132 . they should be able to handle the power.

    attached please find the final stage.

    Thanks in advance alistair
     
    • AB.JPG
      AB.JPG
      File size:
      32.3 KB
      Views:
      259
    Last edited: Apr 22, 2009
  2. StayatHomeElectronics

    Well-Known Member

    Sep 25, 2008
    864
    40
    Any additional information that you can provide with help us identify the problem. What transistor parts are you using? What is their power rating?

    How much power is being consumed by the transistors when no signal is going through them?
     
  3. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You DO know that those Darlington transistors have body diodes, right?

    Have a look at a datasheet. ST Microelectronics datasheet shows them.

    While the body diodes are great when switching other types of loads, they'll wreak havoc with your audio amp scheme.

    I suggest you have a look around Elliott Sound Products' pages; handy ideas for the DIY audiophile who wants to build their own amps and such:
    http://sound.westhost.com/index2.html
     
  4. PRS

    Well-Known Member

    Aug 24, 2008
    989
    35
    You can drive this amp with one input line through a single capacitor to the center node of the 4 diodes. Put .3 ohm resistors between the emitters and the output speaker (8 ohm resistor). This creates an idling current, which you lack.

    Lastly try reducing the 10k resistors experimentally.

    One more thing: Be sure your 1000 uF cap is rated for the voltage it is looking at. Put the short leg to ground.
     
  5. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    The darlington transistors will have thermal runaway. The load causes them to heat which causes them to conduct more which causes them to heat more which causes them to conduct more until they blow up.
    Emitter resistors help a little but what is needed is for the diodes to be mounted tightly to the heatsink of the transistors for thermal tracking. When the diodes heat then their voltage drops which reduces the increased current in the output transistors.
     
  6. cariba

    Thread Starter Member

    Apr 22, 2009
    19
    0
    Sorry forget to draw them but I also have included emitter resistors of 0.22R .

    Could it be because im using a single supply rail instead of a dual one, or maybe u have some suggestions of an appropriate class ab power stage.

    Attached please find the complete schematic which i have designed.

    It just a really basic VA amplifier , Im still an early amateur guys:)

    I really appreciate your help

    Thanks
     
  7. StayatHomeElectronics

    Well-Known Member

    Sep 25, 2008
    864
    40
    SgtWookie,

    Why do the body diodes cause problems with the audio amplifier? I have never done audio amplifier work myself, so I am just wondering.

    S@H Electronics
     
  8. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    Take a look at the datasheet for the Darlingtons our OP is using, and then sketch the body diodes into the circuit. You'll see that when the upper Darlington is conducting and the lower Darlington is OFF, the lower body diode creates a current path to ground. KAPOW!

    AudioGuru's mentioned the change in Vf of the Darlingtons over temperature, which is another big problem. Unless our OP has the diodes closely thermally coupled to the transistors, there will be a mismatch of Vf that will grow worse as the transistors heat up, leading to a thermal runaway situation.
     
  9. cariba

    Thread Starter Member

    Apr 22, 2009
    19
    0
    Errmm got your point, so thats the explosion im having:( now il try to build up a complimentary darlington pair myself.
     
  10. cariba

    Thread Starter Member

    Apr 22, 2009
    19
    0
    SO JUST A SMALL QUESTION, WHY DID THE TRANSISTORS DID NOT BLOW WHEN I USED A 1K LOAD INSTEAD OF A SPEAKER ? THE DIODES WHERE STILL IN ACTION AM I RIGHT??

    THANKS
     
  11. StayatHomeElectronics

    Well-Known Member

    Sep 25, 2008
    864
    40
    I can see how in the first diagram that the body diodes would be an issue if both NPN darlingtons are used. I think it is corrected in the second diagram where you show a NPN and a PNP being used. Am I right on this one?

    I looked at the MJ11012 part that you used. It has a V(BR)CEO, which is the breakdown voltage of the collector-emitter, of 60 V. You are running pretty close to the line here. The resistor voltage/current combination might not have sufficient to destroy the part right away but the speaker may just push you over the edge...

    The MJ11015 part has a breakdown voltage of 120V.
     
  12. PRS

    Well-Known Member

    Aug 24, 2008
    989
    35
    Here's a similar amp I recently designed. It works well, the diodes stay cool. The only heat is from the transistors, the load resistor and the two emmitter resistors. It runs steadily at maximum output indefinitely without stress. The two ended supply lets you directly couple the output to the speaker since the dc voltage there is zero.

    I'd like to see you reduce your input resistors by at least half and see what happens. I'm just curious. I think 10k is limiting the current too much. Will you give it a shot??? ;)
     
  13. Audioguru

    New Member

    Dec 20, 2007
    9,411
    896
    Hi PRS,
    The diodes do not get warm with a power of only 1mW in each one.
    The four diodes can be replaced by a single transistor used as a Vbe multiplier and can have a pot to adjust the idle current in the darlingtons for low power but also low crossover distortion.

    The 1.5 ohm emitter resistors reduce the output swing to only 24.4V p-p across the 8 ohm speaker for a max output of only 9.3W. If 0.22 ohm emitter resistors are used then the swing is 29V p-p and the power to the speaker is 12.3W.
     
  14. PRS

    Well-Known Member

    Aug 24, 2008
    989
    35
    Thanks, Audioguru. You're right, but all I have on hand are those 1.5 ohm resistors. They are rated at 1 watt and they get quite warm, I wouldn't say hot. I actually get only 22 volts peak to peak at the output before clipping. When experimenting, I found that by reducing the bias resistors I got a wider swing without clipping and when I found 8.2 k ohm resistors got me 22 volts I was happy. It's a nice amp, no distortion of any kind. Sine, triangle and square waves look perfect. I'm happy.
    This amp, by the way is for my workbench. I have no use for a 100 watt amp! Good grief! ;)
     
  15. cariba

    Thread Starter Member

    Apr 22, 2009
    19
    0
    Hi guys , first of all really tnx for the help:) i was doing a really big mistake with the capactitor bank , charging it to 100 v and then switching to the transistors .

    the only problem i have now , is the 50 hz humm of the supply on the speaker:(
     
  16. SgtWookie

    Expert

    Jul 17, 2007
    22,182
    1,728
    You either do not have enough capacitance across your supply, or you have (a) bad rectifier(s) in your supply.
     
  17. cariba

    Thread Starter Member

    Apr 22, 2009
    19
    0
    I an using a 100v 30000uF cap , maybe it could be the rectifier:(
     
  18. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    I don't think this circuit is correct. Re-design it.
     
  19. cariba

    Thread Starter Member

    Apr 22, 2009
    19
    0
    Oh my god, im just al college amateur, and the main concept was deisgning a va stage . I think this should satisfy the reuirments, since i got a va stage from 50mv up to 60V ending with a push pull driving a 4 ohm speaker, and i have no clipping or crossover distrotion at all
     
  20. mik3

    Senior Member

    Feb 4, 2008
    4,846
    63
    You don't need a push-pull pair, just use a transistor as an emitter follower and bias it at half the supply voltage.

    The circuit you have is not a push-pull pair.
     
Loading...