78d9cffc   
   From: adrian@poppyrecords.invalid.invalid   
      
   janneman  wrote:   
      
   > On Apr 17, 9:50 am, adr...@poppyrecords.invalid.invalid (Adrian   
   > Tuddenham) wrote:   
   > > janneman  wrote:   
   > > > On Apr 16, 5:14 am, TheOctavist  wrote:   
   > > > > Tube Rolling:-   
   > >   
   >> > Valves have certain specs, which all manufacturers of, say, an   
   ECC83 adhere   
   > > > > to in terms of gain, bias sensitivity, mutual conductance etc. All   
   >> > manufacturers have production spreads, so that a valve may have low   
   gain,   
   > > > > another of the same type higher gain, but still within the acceptable   
   > > > > spread.   
   > >   
    > > > Is there any objectove evidence that a say, Mullard, ECC83 is any   
   different   
   > > > > in spreads and performance than any other manufacturer's.   
   > >   
   > > > Is there any objective evidence that a say, Mullard, valve will sound any   
   > > > > different to another manufacturer's valve.   
   > >   
   > > > > Is there any objective evidence that valve circuits are so sensitive to   
   > > > > changing valves unless it's clear that it's due to the valves being at   
   > > > > opposite ends of their production spreads.   
   > >   
   > > > > Or is it yet another audiophile myth?   
   > >   
   > > > > and Hey Mr. Ian, you rock Sir   
   > >   
   > > > If by 'objective evidence' you include measurements, you can generally   
   > > > say that performance differences can be measured with different tubes.   
   > > > As tube equipment is often designed with little or no feedback they   
   > > > are much more sensitive to individual tube parameters than ss   
   > > > equipment where the feedback makes sure that individual active device   
   > > > performance differences do not lead to measured or otherwise objective   
   > > > differences.   
   > >   
   > > Triodes already include quite heavy internal feedback because the anode   
   > > potential affects the potential gradient between the cathode and grid.   
   > > The screening grid of the tetrode and pentode removes that feedback to   
   > > allow much higher voltage gain*.   
   > >   
   > > Because of this, triode circuits  give reasonably stable gain without   
   > > external feedback, whereas pentode circuits are more dependent on   
   > > individual valve characteristics.  A prudent audio designer always   
   > > arranges feedback around a circuit containing a pentode (things are not   
   > > as straightforward at R.F.).   
   > >   
   > > *The original purpose of the screen grid was to reduce the anode-grid   
   > > capacitance (and the Miller effect) which restricted the H.F.   
   > > amplification of triodes.   
      
      
   >   
   > Ahh yes, good points indeed. But wouldn't you agree that in tube   
   > equipment, even in pentode/tetrode circuits, feedback factors are   
   > generally much lower than in ss?   
      
   Yes, larger numbers are used when describing external feedback in SS   
   circuits, but they don't always accurately reflect what is really going   
   on.   Perhaps less external feedback is needed in a valve circuit   
   because it already has enough internal feedback - or perhaps the valve   
   circuit contains a transformer with phase shifts which limit the amount   
   of feedback which could be safely applied around that stage - or perhaps   
   the transistor circuit generates such a high distortion level that it   
   only becomes acceptable with masses of feedback.  In most cases the   
   designer knows the answer (if he is doing his job properly), but the   
   salesman is never going to let you find out.   
      
   You have to be very careful how you define the level of feedback.   
   Consider an  amplifier which, without feedback, has 10 dB more gain at 1   
   Kc/s than it does at 100 c/s and 10 Kc/s (very poor) and you decide to   
   flatten the frequency response by applying12 dB of feedback.  The   
   feedback at 1Kc/s would indeed be 12db, but at 100 c/s and 10 Kc/s it   
   would only be 2 dB.   
      
   Take another case of a Class-B transistor output stage where the gain   
   varies according to the current in each of the devices.  How much   
   feedback are you really using when the loop gain, bandwidth and   
   slew-rate all vary wildly throughout each cycle of the audio waveform?   
      
   In both these examples, the true feedback is greatest under the maximum   
   gain condition, which is not very helpful for counteracting the   
   shortcomings at the low-gain ends of the performance (which is what   
   feedback is often intended to do).   
      
   Furthermore, you need to consider how much feedback already exists   
   inside the device or its immediate circuitry, before you apply external   
   feedback.  You would find it difficult indeed to design a worthwhile   
   one-transistor audio amplifier without any feedback at all, the   
   distortion would be horrible at anything other than the smallest signal   
   levels.  In contrast, a single pentode without feedback gives quite   
   passable results for domestic equipment - and single triodes were used   
   in good quality professional equipment without external feedback for   
   many years.   
      
   One big advantage of external feedback is that it puts the overall   
   performance of the circuit under the designer's control.  He is not tied   
   to a certain characteristic determined by the valve manufacturers' grid   
   wire geometry.   Worse still, the tolerances on semiconductors are so   
   wide (and so temperature dependent) that it would be impossible to stay   
   within any kind of worthwhile specification without feedback in SS   
   circuits.   
      
      
   > I mean, the fact that over the limited life time of tubes you can   
   > measure (some say hear) changes does point to equipment parameter   
   > variation with device parameters.   
      
   If the circuit is well designed, you should not hear any change until   
   the valves degenerate to the point where they fall well outside their   
   design characteristics.  Even a small amount of feedback will take care   
   of that.   
      
   For triodes, which are used without external feedback, the grid geometry   
   determines the amount of internal feedback (until the end of life) and   
   that is unlikely to have any significant wear-out mechanism.   
      
   You may well hear changes if pentodes are used without feedback, as in   
   some domestic tape recorders and cheap P.A. amplifiers, but that is   
   because the design was driven by the need for the minimum number of   
   valves, not for a good specification and reliable long-term operation.   
      
      
   > And of course tube rolling would be senseless if device parameters   
   > were hidden by the circuitry!   
      
   I'm not sure what you mean by 'tube rolling', but if you mean using the   
   uncorrected large-signal characteristics of a valve to create distortion   
   effects, then I agree with you.  In that case you would definitely   
   expect to hear changes as the device aged (more-so for a pentode than a   
   triode).  However, that is the exact opposite of what a good valve   
   amplifier designer strives to achieve.   
      
   Valves can be used for stable high quality audio amplifiers, they can   
   also be used for effects units.  Some poorly-designed amplifiers have   
   unintentional effects units built into them by mistake and a few   
   specialist musical instrument amplifiers have them by design.  None of   
      
   [continued in next message]   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)