Ad Widget

Collapse

Announcement

Collapse
No announcement yet.

Soft vs hard clipping - I feel such a fool

Collapse
X
 
  • Filter
  • Time
  • Show
Clear All
new posts

  • #46
    Some very strange notions in this thread. So output clipping is the same as preamp clipping? Every measure it? Unless using a single ended output, the push-pull configuration will cancel even harmonics that each side output generates. If you are not seeing 3rd harmonic stronger than 2nd from the power stage, you simply have a bum tube.
    Soft clipping is not voltage limited, it is nonlinearity in the active device at close to saturation and close to cut off and does have a compression effect. Look at the curves of a tube and see the exaggerated "s" where a given change in input signal has a smaller impact on the output. That is the region where the term "touch" comes into play in guitar amps, where slight difference in playing can intentionally move the amp from clean to compressed before the onset of clipping. It is like a compressor with the threshold set high and the ratio set to 10:1. But overplay it and hard clipping takes over.
    Preamp stages are usually single ended so a different harmonic content is expected. Stage gain is important in all forms of music production from console, to decks to guitar amps. Most home builders do not pay much attention to that but it separates the great sounding amps, when played well within its design parameters, and a clone that sound uninspired with essentially the same circuit. That natural non-linearity of active devices when near cut off and saturation should be determined for every stage so the signal levels between stages don't unintentionally overdrive the following stage well past that compression zone. Sure sometimes you want to hard clip but much of the time you don't and you can't help it but to push one of the stages into hard clipping if the gain of the preceding stage is too high.
    You can look at a spectrum of a power amp with signal being injected into the PI or PA-In jack and you will quickly see the harmonic content at all levels are dominated by odd harmonics. Do the same with the Pre-amp out or effects send and you will notice even in very clean mode, second harmonic is the dominate harmonic just as "nature" would predict. When using opamps, any induced moderate non-linearity has to be created with a nonlinear feedback response. A transistor in common base configuration can certainly act as a nonlinear feedback element for example. You can create soft clipping in solid state, take a look at the output of the classic Fuzz-Face, with just a couple Ge transistors. And it does sound musical as a result.

    Comment


    • #47
      I think many guitar amps (preamps at least) are designed so that they do not produce a "static" harmonic pattern. More likely the harmonic pattern of distortion will vary in interaction with overall envelope of the signal.

      e.g. Asymmetric clipping and capacitive coupling produces DC offset shift because the crossover point will always try to stabilise into the "middle" of the waveform. This means that in the following stages the symmetry vs. asymmetry of clipping will also gradually change. This characteristic is what introduces "touch sensitivity" and "feel", since static "waveshaping" just produces a single type of harmonic pattern where gain mainly affects the amplitude of those harmonics, not the pattern itself.

      I don't buy the idea about "musical" even vs. odd order harmonics. Nor the idea about "pleasantness" of hard vs. soft clipping. First of all, they are pretty much debunked by all real designs out there.

      Secondly, amplitude distortion from clipping also introduces intermodulation distortion. With that in mind you have to forget whjat is even and odd, because intermodulation of complex wave will always produce frequencies that are bound to be discordant with the fundamental frequency.


      Oh, and humans perceive intermodulation distortion much better than they do perceive amplitude distortion...

      ---

      Symmetric amplitude distortion produces only odd order harmonics, asymmetric produces both even and odd. Because of that, with asymmetric amplitude distortion the THD is most likely higher than with just symmetric.

      Soft clipping produces more amplitude distortion than hard clipping because distortion is introduced much earlier than when the waveform distinctly clips. This amplitude distortion will also intermodulate and in practice too much soft clipping will therefore ruin all note separation.

      There are no "good" or "bad" principles in this, although some dubious sources try hard to generate such image. Especially when distortion is deliberately produced as a "special effect" you can throw all the rules written by "HiFi" designers out of the window. Totally different application.
      Last edited by teemuk; 08-05-2015, 03:05 PM.

      Comment


      • #48
        Look at the curves of a tube and see the exaggerated "s" where a given change in input signal has a smaller impact on the output. That is the region where the term "touch" comes into play in guitar amps, where slight difference in playing can intentionally move the amp from clean to compressed before the onset of clipping. It is like a compressor with the threshold set high and the ratio set to 10:1. But overplay it and hard clipping takes over.
        There's this so-called "gain compression":

        Triode's gain decreases when its plate voltage approaches B+. Since gain gradually decreases before saturation it creates a "soft clipping" effect - but only to one half wave. The other half wave will not gain compress similarly because voltage swings to other direction.

        Bipolar transistors perfrom similarly: Gain decreases when collector voltage approaches B+.

        Pentode's gain decreases when its screen voltage decreases. In usual schemes the screen current is allowed to modulate screen voltage to some extent so it also produces gain compression effect. And similarly at one half wave only.

        Comment


        • #49
          Originally posted by Mark Hammer View Post
          Consider that the diodes in an op-amp feedback loop form a sort of variable conductor. When the signal voltage at the op-amp output reaches the forward voltage of those diodes, they conduct. But the gain of an op-amp is dictated by how much negative feedback from the output is employed. If there is more negative feedback, then the gain is reduced.

          So, the action of the diodes in the feedback loop modify the gain of the op-amp, temporarily reducing it. Of course, once the gain is reduced, via the diodes conducting, then that results in less output, which results in the feedback signal dropping below the forward voltage of the diodes. And so on.
          You are ignoring the reality of iep's plots. Your explanation needs some modification.

          The feedback in an op amp circuit of this kind establishes the condition that the same ac current that flows through the input resistance to virtual ground also flows through the elements in the feedback circuit, two diodes (oppositely directed) and a resistor in parallel, the resistor possibly of very high value. The instantaneous gain of such a circuit is the ratio of the feedback impedance to the input resistance. When neither diode conducts, the gain is the ratio of the resistors. This happens when the input voltage is very small. When a diode conducts (input voltage increases), its impedance drops to a very low value and the incremental gain drops to a very low value, meaning that the output voltage changes very little in response to further increases in input voltage. The signal does not drop below the forward conduction voltage of the diode; it stays high enough to keep the diode conducting until the input voltage is reduced back to where it was just before conduction occurred.

          As a result, the waveform is if anything a bit sharper than the other kind of circuit, depending on the value of the feedback resistor.

          Comment


          • #50
            Some op-amps use those back-to-back diodes right on their inputs.. Click image for larger version

Name:	2011123025911803.jpg
Views:	1
Size:	40.6 KB
ID:	839211
            Now Trending: China has found a way to turn stupidity into money!

            Comment


            • #51
              Originally posted by guitician View Post
              Some op-amps use those back-to-back diodes right on their inputs.. [ATTACH=CONFIG]35240[/ATTACH]
              Which is irrelevant when the feedback loop is operating normally since the ac voltage across the two input terminals is under a millivolt.

              Comment


              • #52
                Plug a guitar into that op-amp and it will distort at +/- .3v
                Now Trending: China has found a way to turn stupidity into money!

                Comment


                • #53
                  Originally posted by guitician View Post
                  Plug a guitar into that op-amp and it will distort at +/- .3v
                  No, not at all. We are discussing an inverting circuit where the negative feedback keeps the differential input to a very small level, less than a mv.

                  Comment


                  • #54
                    So an inverting buffer will not give me an equal output?

                    Oh, yeah the virtual earth ground...forgot my op amp stuff
                    Last edited by guitician; 08-13-2015, 12:09 AM.
                    Now Trending: China has found a way to turn stupidity into money!

                    Comment


                    • #55
                      https://en.wikipedia.org/wiki/Virtual_ground
                      My band:- http://www.youtube.com/user/RedwingBand

                      Comment

                      Working...
                      X