From: info@turneraudio.com.au   
      
   > Now Mr Snook is an Englishman And Gentleman   
      
      
   ** Err  -  what does that make Pat ?   
      
       An Aussie bastard ?   
      
   But I have a splendid pedigree, I'm a well bred Auzzie Bah Stard.   
      
      
   > who Quad-II fanciers may find to be a rich source of DETAILED information   
   > on Quad-II.   
      
      
   ** Certainly long on trivial details and fetish stuff.   
      
     But very short on anything really worthwhile.   
      
   ....  Phil   
      
   Such a Philesque remark !   
      
   But meanwhile, today it rained well, and none of my cycling bunch turned up   
   where we usually start on a Sunday for a little 80km ride someplace.   
      
   OK, I get damp but not soaking wet, I go back home to shed where I do more   
   work on MkIV amps, and I dig out a Quad-II to measure it. I replace 680r,   
   180r, and both 0.1uF and 25uF. It seems Hunts 0.1uF had become 400k resistors.   
   Enough to turn on both    
   KT66 and fuse the Rk 180r. But just how 680r Rk for EF86 fused is unknown,   
   just some old junk fails, eh.     
   I set up safe mains lead. turn it on, and it came to life OK, and all Vdc were   
   correct, although I got about 1.5Vdc across 680k after about 20 minutes.   
   680k had gone up to 750k, but not a problem. Vac balance from EF86 are about   
   7% different, about normal and KT66 Iadc are within 25%, about right for   
   working tubes maybe 1/2 way through life.    
      
   After several hours I had a good looking graph of THD Vs Po for amp without   
   GNFB, ie, with R10 100r shunted, and with the GNFB. OPT was set for 16r0, and   
   load used was 16r0, so my test conditions were same as yours, and I found   
   average gain reduction    
   with NFB at Po less than 2 watts was -20dB, and that THD reduction at 9W,   
   which is the maximum class A point dropped 24dB with GNFB.   
   I have installed the graph into my page on Quad-II re-engineering and revised   
   the text so a decent record will exist for future.    
      
   Then I measured my MkIV Dynaco again but with load 4r0 on maximum Sec turns   
   and this gives same RLa-a as Quad-II. There is only 14dB GNFB, but THD below   
   2.2Watts was 6dB below Quad-II, so I put the MkIV curve as a dashed line under   
   the Quad curves.    
   I need ppl to know I seem to get less THD with less dependance on NFB than   
   Quad.   
      
   But the MkIV with KT88 run cooler, Ea 390V x Ia 47mA, in Quad, KT66 have +340V   
   x 65mA.    
      
   If I make Pda in KT88 = 30W, Ia = 70mA, and then I get same 9W class A as Quad   
   with 4k3 RLa-a. Also THD would be even lower than I have now. So now I wonder   
   just how much THD is in existing EF86, but whatever it is, maybe its a lot   
   lower than in KT66.   
      
   The MkIV has been set up to use 8r0 speakers at the highest outlet, and when   
   8r0 is used, the MkIV still makes more AB1 Po than Quad-II, and as it is with   
   low Pda, also makes 9W max class A. The higher RL makes KT88 have slightly   
   higher gain thus    
   increasing the dB of GNFB, so maybe much less THD than Quad-II.    
      
   Then it occurred to me you had a PP OPT used with sec as the input to drive   
   your Quad-II output stage. But did you think of iron generated 3H? Maybe much   
   more than your signal gene, and if a sig gene has Rout 600r, then the input Z   
   of tranny with GOSS    
   might need to be very high. The Iron makes 3H, and what phase is it?    
      
   If your sig gene has 0.002% THD, then which ever way you try to make a   
   balanced voltage source of +/- say 30Vrms, it is not as easy as using an OPT   
   you have laying about, but you must build a dedicated balanced amp complete   
   with bagfulls of NFB and even    
   then its THD may well be 0.005%.    
      
   Its easier to just measure THD of input / driver stage, and it should be   
   mainly 3H, then work out what phase it is, and see how that affects what you   
   measure with /without GNFB for the whole amp. Simple.       
      
   I should soon finish the editing with new graph and I'll let everyone know.    
      
   Amplifiers are only fully understood if they are compared.    
      
   Patrick Turner   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)