[Amps] new amp race

[Jim Thomson]

Date: Tue, 23 May 2017 23:26:55 -0500
From: "Jeff AC0C" <keepwalking188 at ac0c.com>
To: "Bill Turner" <dezrat at outlook.com>, "Amps group"
<amps at contesting.com>
Subject: Re: [Amps] new amp race


<A single band amp, 3-500z glass tube  - that's not even a beginners 
challenge by comparison.  the Cout is about 7x higher plus the monster anode 
on the YC156 makes a hell of a good capacitor with the enclosure where the 
effect on a glass tube is token.

When you hang the L on the YC156, you get releif on 10m (for example) but it 
pushes the Cin zero point on the pi-net down in frequency so that you may 
get it working on 10m but now it's marginal on 12m and you have a monster Q 
on 17m.

The bottom line is a YC156 makes a great beginners project for the low 
bands.  But if you want to make it work on the top end, my advice is to pick 
another tube and save yourself a mountain of hours.  Despite the somewhat 
higher costs, it's actually pretty hard to beat an 8877 with a grid 
protection circuit as a limit-capable all band tube when you consider the 
entire project.

73/jeff/ac0c

##  On a buddys  YC-156, I designed the 80-10m , band switched PI net output
tank using GM3SEKs   PI  spread sheet.  That took 2 of us a trillion hrs b4 we nailed it.

##  That YC-156 tube is a bitch to work right on 10m.   The tube, on a wooden test bench,
measures 36 pf from anode to grid..... which then rises to 53-55 pf, when grid flange is bolted
to the chassis.  That additional C  is caused by the  proximity of the lower anode fins to the
chassis below them.   The ceramic tune cap we used was a 10-375 pf @ 15 kv type. 

##  So now you have 55 + 10 = 65 pf....with the tune cap fully UN meshed !  The fix
is to introduce a TINY bit of uh...between anode and C1  tune cap.  We used silver plated
wide cu strap for both that tiny coil..and also the main 10M tank coil.   If you make the tiny coil
too big, the  resulting plate load Z  ends up low....which then requires a real small main 10M
tank coil.   Make the  tiny coil b4 main PI net.... too small, then the Q  will rise on 10M, to perhaps
something you dont want.  You want the tiny coil to be big enough to also work on 15M band.    

##  we call that type of network an... L- PI.  It works on the principle that the tubes 55 pf, combined 
with the tiny uh, forms a ..step down L network, which drops the plate load Z of the tube WAY down
on 10M.....down low enough that the following  PI net can handle..with a semi normal loaded Q. 
Which of course will  require higher tune + load  cap values....and a real small  main 10M  tank coil
between em. 

##  As jeff found out, the optimum size of that tiny L  coil, between anode and C1  tune cap is different
between  10M..and  15M.   So a compromise has to made, and here we are talking about miniscule
amounts of uh, like .23 uh, .25 uh  etc.   So then you have to wind flat strap coils..used since 1.125 inch OD tubing
is impossible to work with... using different IDs... for both the  tiny L coil..and also different IDs... for the
main 10M coil....such that you end up with the correct amount of uh on both of em...and that their
overall lengths are also correct, to make the connections.  Flat strap coils have virtually zero C between turns,
since they appear like 2 x knife edges  facing each other. 

##  Even with all of that, and several versions tried, and more software employed, and also moving some of the
C2 from the 80-10m, bandswitched, and manually tuned...PI tuned input...to the cathode.... the best we could get
on 10M was  7.5 kw.   And  9 kw on all the lower bands.  The high power PI  used a triple wafer model 85
bandswitch, with all 3 wafers in parallel, to handle the current.     Tuned input consisted of a pair of broadcast
air variables and  a tapped  solid 8 gauge coil..which was tapped with a small single wafer bandswitch. 

##  That was extremely difficult to get right on 10M.   If it was a monoband 10M, it would have been a lot
easier.   The YC-156  is a 80-40-30-20-15-10m affair.  Not used on 12m or 17m.  The loaded Q rises on 12M,
if the 10M position is re-used. 

##  I have seen one  YC-156  monoband amp on 6M....and that was a trick and a half...with eff  not that great. 

##  On my hb  3CX-3000A7,  I used a roller  for  160-80-60-40-30 M..then  .375 tubing  taxcked on the end of the
roller for   20M.   15M  tap on the 20m coil done with a modified HV solenoid contactor.    The L-PI  trick was also
used...to get the 15M  loaded Q way down....down to about 8.   Then when 15M position re-used on 17M band, 
Q rises to 12.   In this one off case, I had to use a .66 uh coil between  anode...and main PI net.   It would also
work without the .66 uh coil, but the Q rose a lot higher, and could not be re-used on 17M band....  Q was way too high on 17M.   

##  3CX-3000A7  is  24 pf between anode + grid.....which rose to 33 pf, when inserted into socket, same deal with proximity of lower
fins to the chassis.   The 3CX-6000A7 measures 24.5 pf  on the bench...and it rises to 41 pf  when in the socket.

##  If I had to do this again,  I would leave  10 + 12m out of the amp...and instead build a monoband 10+12m amp.  Since the
fil xfmr is external, and ditto  with the B+  supply, no big deal to re-use both on a monoband  10-12 amp. 

##  latest amp project is  160-80-40-20-17-15m.. band switched.  I would not use a roller again..pita, buys you nothing really.    
RF parts used to supply a metal chimney for the 3x3 tube.....but it added exactly 10 pf when installed.  Useless chimney.  
Their latest version uses clear acylic +  teflon ring  on top..then no additional stray C. 

##  8877 and  3-500Z  is doable..and fairly straight forward.  But these YC-156 metal tubes are no fun.  Interesting to note, 
the YC-156 is 4.94 diam..and the 3x6 is  6.125 diam..yet the 3x6 has way less stray C.    The ...ceramic stem on both the 
3x3 and also 3x6 is way smaller diam than the YC-156.  3x3 and 3x6  both use the same socket. 

Jim   VE7RF