From: apogosso@tpg.com.au   
      
   "flipper"  wrote in message   
   news:a9j1a758kkoedvnr9qavc287sgvugadtn5@4ax.com...   
   > On Wed, 19 Oct 2011 01:02:54 -0700 (PDT), Patrick Turner   
   >  wrote:   
   >   
   >    
   > Unfortunately my tube models are not accurate at very low plate   
   > voltages (something I've run into before) but adjusting things to meet   
   > his criteria I have successfully modeled the 6SN7 'detector' so that   
   > it holds plate voltage within .4V of 80mA idle with peak currents   
   > through Rk corresponding to what you call 'Vdc' equivalent to a 190mA   
   > average. If, however, I put a 190mA idle, creating an *actual DC*   
   > "Vdc" of 2.87V (vs the 1.2V 80mA idle) plate shifts 8.8V, which is   
   > close to 20 times the shift when it's 'AC'.   
   >   
   > I also managed to model it using a 6DJ8.   
   >   
   > In short, it works if you properly bias the tube   
      
   Correct!   
      
   The 6SN7 trough detector is not just a common grid amplifier stage as   
   Partick is trying (in vain) to convince us. It is a detector close to   
   cutoff. For example, if the quiescent current through 15R is meant to be   
   80mA, then the cathode of 6SN7 would be at Vk=+1.2V. To keep the trough   
   detector close to cut-off, its plate voltage should be about Vk * mu. For   
   6SN7 mu=20, so plate voltage should be around 24V or even less (given   
   thermal emission potential and the fact that mu is smaller at low currents).   
   Plate current would be small, probably 50...150uA.   
      
   High gm sharp cut-off tubes like 6DJ8 will make a better trough detector,   
   but a transistor would be even better. Could be a n-JFET -- just select a   
   specimen (e.g., BFT46) with 1.2V cutoff voltage (or adjust 15R).   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)