• EFFECT OF FEEDBACK ON HARMONICS. Updated 13-May-2006.
  It is well known that a square-law amplifier adds only second harmonics with no feedback, but many high order harmonics with feedback applied, but this is almost entirely misleading, as explained here.
• INPUT STAGE DISTORTION.
  Seven of the most common one and two transistor input stages are compared by calculating their levels of intermodulation distortion under identical stage output current levels.
• SLEW RATE AND T.I.D. (Updated 12 June 2006)
  The maximum slew rate requirement and the design techniques for low t.i.d. are investigated. Now includes measurement results for maximum CD slew rate.
• MEASURING DISTORTION.
  To test our low distortion designs an alternative to buying expensive distortion measuring instruments is to use the bridge-nulling method in which the input and output are directly compared.
• PHASE INTERMODULATION.
  Explanation, reduction and measurement of phase intermodulation distortion.
• COMMON-MODE DISTORTION.
  Common-mode distortion is examined, a test method described, and the performance of typical input stage transistors tested and compared.
• MEASURING CAPACITOR DISTORTION. (Updated 18-Aug-2008)
  Capacitors are known to add a little low order harmonic distortion, but does transient testing reveal something a lot worse than this?
• ERROR FEEDBACK AND NESTED LOOPS.
  An examination of how effective these methods can be compared to overall negative feedback.
• SYMMETRY. (Updated 01-May-2006)
  Symmetry can be good or bad, but just aiming for maximum circuit symmetry may add unnecessary complexity with no real benefit. It is possible in principle to achieve zero distortion in some circuits with correct use of symmetry.
• DISTORTION - HOW LOW CAN YOU GO?
  Is there any limit to how far power amplifier distortion can be reduced?
• FEEDBACK WORKS PART 1.
  The idea that high levels of overall negative feedback should be avoided is repeated often, but both theory and measurement lead to a different conclusion.
• FEEDBACK WORKS PART 2.
  Feedback is sometimes avoided to prevent the addition of high order harmonics, but this is based on analysis with a single frequency. An analysis based on music signals suggests that a high feedback approach may be more effective.
• FEEDBACK WORKS PART 3.
  The effects of phase shift and time delay on distortion reduction are investigated.
• OVERLOAD AND SHORT-CIRCUIT PROTECTION. Updated 19.Jan.2006.
  A small change to the usual load-line protection circuit may reduce distortion. Temperature compensated protection is also considered, with a simple method suggested.
• LOCAL FEEDBACK.
  Does local feedback avoid all the problems of overall feedback? Here the addition of high order harmonics is investigated.
• SQUARE WAVE STABILITY TESTING.
  Testing for stability using square waves and reactive loads may give misleading results not related to stability, while real stability problems may be missed. Examples are now included of problems revealed by clipping behaviour.
• ABX LISTENING TESTS.
  A few comments on the problems of conducting reliable listening tests, and the uses of ABX testing.
  
  For further coverage of audio power amplifier design there are a number of website sources. Two I suggest trying are:
  Elliot Sound Products. (Rod Elliot).
  and The Self Site. (Douglas Self).
  I disagree with some of the content in both cases, but I believe they include plenty of useful information. Self has a section in one of his books entitled 'FET output stages: why they are a bad idea.' while the Elliot site has an article saying that feedback does not add high order harmonics (now updated, saying that high harmonics are added, but only at significant levels for low feedback, which is more or less my own conclusion.)
  
  
• ARCHIVE.
  Older design notes and design ideas, Oct 2004.
• LINKS.
  An assortment of links to other sites.

HOME.