From: info@turneraudio.com.au   
      
   Big Bad Bob remarked cleverly.....    
      
   > I dunno how they'd do AM now, lots and lots and lots of utterly   
   incomprehensible didgets & bizibots I guess.    
      
   here's what _I_ would do, if it were solid state:    
      
   a) use a ring modulator with VERY heavy negative feedback    
      
   Yeah, I guess I could, but then its not simple with large amounts of FB.....   
      
   Have to a decent working schematic?   
      
   - or -    
      
   b) use a relatively linear MOSFET biased to its most linear zone with a    
   tube-like modulation circuit, i.e.  source-feed from osc for carrier,    
   gate modulation.    
      
   Yes, well, I tried for 3 weeks to get IRF610 to give a healthy bandwidth and   
   low THD of the envelope shape but their self capacitances seemed to defeat me.   
   BF469 worked better, PP for RF with balanced RF to bases, and voltage   
   controlled current source to    
   commoned emitters. Not bad even with untuned collector load of 33r to each   
   collector, with DC feed through CT choke If I wanted a tuned LC in collector   
   circuit, ( or drain circuit of mosfets, and variable F, then a whole other   
   approach is needed, and    
   then you realise the ordinary dual gang 5pF-480pF variable cap needs an easily   
   made coil, and quite low current, all like what is in HP6060A - Now that's a   
   good RF gene with AM modulation.   
   The tubes work with very low currents and highish voltages because tuned   
   circuit impedances are so high, so the trick is to use an anode coil with say   
   110 turns anode to anode with CT and on a ferrite rod, and have two secs in   
   parrallel of 4 turns each -    
   and I got 5.6Vrms easy, less THD than any of the SS I tried, and I'm in the   
   middle of fixing up the AF gain abd NFB, but results look promising using AF   
   drive to commoned screens, and an SS diff amp to supply Idc to commoned   
   cathodes so that Vdc from a    
   detector keeps the RF output level. Its not until you try these things you   
   realise how darn difficult it is to make everything work just right between   
   525kHz and 1.8MHz.   
      
   I'm concluding to myself that having a PP tube arrangement for RF probably is   
   not essential at all, because even with PP RF, the AF works the same way -   
   single ended - on both the tubes of PP pair. Screen mod is quite good, and it   
   means cathode can be    
   well bypassed to 0V, and RF delivered to g1 by low Z drive circuit. Getting   
   maximum pk-pk AM = twice carrier pk-pk IS difficult. Ia declines as % mod   
   increases, hence my use of Vdc controlled current source to cathodes. Even   
   then, still hard to get max    
   AM pk-pk to twice carrier pk-pk.   
   I heard plate mode works better, because modulator supplies power to the   
   output RF power - same principle applies for high pwr Xmitter or for tiny pwr   
   sig gene.   
   Basically, the Ia MUST increase and so must gm, to get the change of carrier   
   levels.   
   Its not very well explained everywhere I search. They say, do this, and that,   
   and no guaranteed to work schematic. Much BS with no details.   
       
      
   - or -    
      
   c) use a "digital exciter" circuit.  probably a ZILLION ways to make    
   THAT work.  I'd give PWM a try.  for 455khz IF modulator, you'd need    
   455Mhz with divide-by-1024 and 12-bit accuracy.  THEN, modulate the    
   455Mhz with 455khz digital signal, with amplitude calculated based on    
   digital sampling of the input.  Input is at 0.5V, output swing is +/-    
   256.  Input is at 1V, output swing is +/- 512 [max modulation].  Anyway,    
   that might be kind of interesting.    
      
   OMG. OMG twice!!!! Digi-pharkin-tal???    
      
   Now you've really pissed me off. Well, a little. Its OK, I'll recover.   
   Lord Noze how many gidajeticals are in the box, with virtualy no idea what   
   they are all doing........Private commercial knowledge......can't even find   
   one single easily understood description of digital radio transmission.   
   Not one single waveform anywhere, or any circuit. Oh yeah, all kept secret.   
      
   - and -    
      
   You'd use an IF freq. for the modulator, maybe 455khz like AM receivers,    
   then drive this into a freq. synthesis converter, similar to old-style    
   "crystal oven" freq. synth transmitters the Navy used back in the day.    
      
      
   it should work for just about the entire range, and you should be able    
   to use tuned SSB mechanical filters on the transmitter IF, with an "add    
   carrier" level adjust.    
      
      
   but yeah, BEST way to do AM modulation (In My Bombastic Opinion) is with    
   an IF carrier, so you can do SSB and 'carrier injection' and whatnot via    
   off-the-shelf mechanical filters and THEN drive the transmitter through    
   a freq. synth. thingy.  Super-accurate, good quality SSB, built-in    
   ACCURATE carrier generation for SSB reception, and so forth.    
      
   I don't want anything to do with SSB. I just want an easy done low THD AM,   
   DSB+c signal.    
      
   [that worked really nice with TUBES back in the day, it's how the Navy    
   transmitters (that I studied for ET 'A' school) worked, though they used    
   solid state switching diodes to generate harmonics for synthesis]    
      
   Maybe all that worked just fine. But maybe nothing is able to easily be   
   applied using low pwr tubes in a simple home made box, now riddled with drill   
   holes after so many attempts at doing things I did before which didn't work as   
   well as stuff I tried    
   later....and maybe somethings I've tried 5 times before I stop frowning and   
   swearing at the darn gear.   
      
   The shall work.    
   It shall work over the proposed F range.   
   Performance shall be consistent in all behavior over the entire RF range.   
   It shall have low THD in envelope shape up to 97% mod.   
   It shall be able to be modulated by AF between 15Hz and 10kHz and up to 97%.   
   Ringing of envelope shape when using 1kHz square wave shall be minimal.   
   There shall be minimal production of spurious behaviors or parasitic F.   
      
   Unless you have goals, nothing is realised. I'm a horrid man, because that's   
   how my mind works. :-)   
      
   Anyway, soldering on to nowhere is fine for me, I don't mind being totally   
   useless, the world won't miss me later, so its the journey and the wonderment   
   for me.   
   Patrick Turner   
      
   --- SoupGate-Win32 v1.05   
    * Origin: you cannot sedate... all the things you hate (1:229/2)