No announcement yet.

sunn 200S choke theory question

  • Filter
  • Time
  • Show
Clear All
new posts

  • #16
    Originally posted by danfrank View Post
    Hi, I know I'm replying years after this thread was started but I just joined the forum a few days ago. The Dynaco choke in question is rated 1.5 Henries @ 200ma. These were common in Dynaco tube amps like the ST-70 and others.
    It is my understanding that as the current goes up in a choke, the inductance decreases. As current goes down, inductance goes up. They used to make "swinging chokes" which had two ratings: One at a low current and another rating at a much higher current.
    diyAudio - Swinging Choke or Smoothing Choke
    Now, in push-pull amps, chokes aren't really required for the output stage because of common mode rejection. Any noise created in the output section cancels out in the output transformer as the two sides are dynamically opposed in currents. A choke in the power supply before the output transformer may help, but like stated above, isn't that cost effective vs benefit. IE. A really large and expensive choke would be needed to make a noticeable difference in sound. If the output transformer is Ultra-Linear, then the output transformer acts as a "choke filter" to the screens besides making them function as "pseudo-triode".
    Downwind of the output stage, everything works in class "A" so power supply noise rejection isn't super important as each stage pulls constant current from the PS. A choke will tighten up the sound by smoothing the supply ripple, so there is a benefit. Everyone has to remember that 50+ years ago, Electrolytic capacitor technology wasn't what it is today; power supply caps were pretty dismal and expensive. This is where chokes really came in handy in power supplies. Look at old Western Electric amplifiers from the 1920's and they had very few filter caps in them, but a lot of chokes. W.E. actually used a lot of neat tricks to get the job done. Worth taking a look at.
    Thanks for posting. There's some truth to a choke's change in value as small to larger currents are drawn but since current stays nearly even in a class A or near-A amp, you can expect the choke's value to be what it's expected to be. There were/are some swinging choke designs where inductance changes more radically. These aren't much found in audio amps, they're more suited to RF transmitters.

    I doubt David Hafler would have included the "pre filter" if he didn't think it necessary for the MkIII, MkIV & Stereo70 Dynaco amps which were copied into Sunn tube guitar/bass/PA amps. One fine day some 25 years ago, for a lark I tried running MkIII's from the raw B+ supply. It didn't work out well, sure I got another couple watts, barely more than stock. But I also got audible distortion. Running sine waves thru the MkIII's and watching the scope I easily saw the "blip" added = distortion. A bridge too far, retreated one step to stock and all's OK with the world & the Dyna MkIII's. Dam' I had 'em sold too, and in the time I wasted faffing around, the buyer found himself a pair of McIntosh C40 instead. Well no worries, he was happy with them and the C20 preamp he also scored, and had me work them over, so I ended up making a couple bucks & learning a lesson: my "improvement" to Dynacos was no improvement.
    This isn't the future I signed up for.


    • #17
      Having a CRC filter before the plates is still a good thing for Class AB amps, as the ripple-cancelling benefit of push-pull goes away as soon as one tube in the pair goes into cutoff.


      • #18
        Originally posted by danfrank View Post


        Downwind of the output stage, everything works in class "A" so power supply noise rejection isn't super important as each stage pulls constant current from the PS.
        Thanks for the info on Dynaco choke specs.

        In reality, the earlier stages are extremely sensitive to power supply noise. The average current is constant but it's signal current that matters. A typical 12AX7 common cathode stage has only about 14dB of power supply noise rejection i.e 1V of ripple on the power rail will produce 200mV at the plate. Also, the signal level is smaller so meaning you really have to pay attention to ripple in these stages. That is why there is often a RC filter ladder to successively attenuate it down the the lowest level input stage.
        Last edited by nickb; 01-22-2016, 09:14 PM. Reason: Typo
        Experience is something you get, just after you really needed it.


        • #19
          Originally posted by marku52 View Post
          I jsut see the amp getting hummy, because 500 ohms of inductive filter reactance went to about zero, but the 58 ohms remains as it was before. I don't see some runaway current to burn it up.

          I think I'm calling BS on Commons knowledge in this particular case, but welcome any correction.
          Just hummy? Try that on the bench and let us know how that works out for you...


          • #20
            Don't forget the screens are also sensitive to power supply noise. Not a problem with pentode connection, but it can be pretty noisy in UL - hence Fender's bias balance trick.

            I wonder how important the ripple might be to the overdriven sound. I do remember one experiment done where they used rock-solid supplies throughout a tweed bassman and supposedly it sounded awful, but I can't find the link.

            Nevermind, here it is:


            • #21
              As soon as the power stage clips, situation changes a lot: hum itself was cancelled at lower powers because of the circuit symmetry, still does if it clips symmetrically, but ripple still modulates the clipped sinewave and that can be heard.

              Compare it to a double sideband transmitter: carrier (in this case 100/120 hz) is cancelled, but the sidebands (audio modulated by it) are not.

              Imagine audio subject to low intensity Tremolo, with sawtooth oscillator set to 100/120Hz, and you'll be describing it exactly.
              Juan Manuel Fahey