From: info@turneraudio.com.au   
      
   On Monday, 26 November 2012 12:31:56 UTC+11, John L Stewart wrote:   
   > Tried some simple nfb amplifier simulations today based on 6V6 OP driven by   
   a 6SJ7. The amplifier uses the following settings- 6SJ7 G = 1.0 ma/volt rp =   
   One M Rl = 270 K Rg = 560 K Following grid resistor 6V6 G = 4.1 ma/Volt rp =   
   50 K Rl = 5 K The    
   first check was with a 560 K NFB path from the 6V6 plate to the 6SJ7 plate, a   
   simple form of local feedback.    
      
   Its shunt FB, with Ra of 6SJ7 and any other parallel R forming R1, with R2   
   being the 560k. Pages 333 and 334 of RDH4 explain it further.   
      
   NFB = 16.13 db DF = 4.54 Looks useful & easily done The second check tried NFB   
   from the output transformer secondary of an OPT of 5K to 8R. Results would be   
   similar for say something like a 35L6 driving a 4R speaker. NFB = 4.66 db DF =   
   0.95 (Unity    
   Damping?) Better than none! These numbers agree well with a simple mod I did   
   to a car radio about 50 years ago. A third check tried NFB from the OPT   
   secondary again, but this time the OPT was 5K to 200R (20%). NFB = 13.19 db DF   
   = 5.00 Looks good but    
   needs a special OPT All have the effect of lowering the apparent plate   
   resistance. More will result in sensitivity to power supply disturbances, just   
   as with a triode. My thoughts anyway, John    
      
   I tried various forms of the shunt FB as you mention in SE amps in old radios   
   where GNFB may have been difficult. I also tried it for a PP hi-fi amp, but I   
   found later the CFB from OPT worked better, with some GNFB.   
      
   But any PP UL amp with plain old UL OPT may be converted to local CFB while   
   using the existing PSU. One finds a CT choke with no air gap, able to take Ia   
   of each tube, and able to take about 100Vrms across winding without saturating   
   at above say 20Hz -    
   not hard to make, or could be a small old OPT primary. THEN, each cathode is   
   bypassed with 470uF to a screen tap, with screens taken to a fixed Eg2.   
   Phasing needs doing right, but then also sec load is reflected to tubes as a   
   higher RLa-a load. Many    
   things are possible with existing OPTs without needing any speacial OPT IF   
   there is a tap along a primary winding somewhere. Such things suit multigrid   
   tubes. Triodes don't need to be anything except a triode. So, in old radios, I   
   remove tube rectifier    
   and use Si diodes to boost B+, and then instal a 6L6 or EL34 in triode to   
   replace the 6V6. Voila, much better sound, and usually slightly more PO.   
       
   Patrick Turner.    
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)