Amps
[Top] [All Lists]

Re: [Amps] Plate Cap Quandary

To: "Jim Thomson" <jim.thom@telus.net>, <amps@contesting.com>
Subject: Re: [Amps] Plate Cap Quandary
From: "Carl" <km1h@jeremy.mv.com>
Date: Sun, 7 Aug 2011 09:11:03 -0400
List-post: <amps@contesting.com">mailto:amps@contesting.com>
----- Original Message ----- 
From: "Jim Thomson" <jim.thom@telus.net>
To: <amps@contesting.com>
Sent: Sunday, August 07, 2011 7:54 AM
Subject: [Amps] Plate Cap Quandary


> Date: Sat, 06 Aug 2011 11:39:58 -0700
> From: Jim Barber <audioguy@q.com>
> Subject: Re: [Amps] Plate Cap Quandary
> To: Manfred Mornhinweg <manfred@ludens.cl>
> Cc: amps@contesting.com
> Message-ID: <4E3D8A7E.10106@q.com>
> Content-Type: text/plain; charset=ISO-8859-1; format=flowed
>
>
> I wasn't paying attention when the beginning of the thread marched by,
> thus the bad timing.
>
> So: center-tap the 10m coil with the Tune C, then just leave the
> "left-hand" end open?
>
> The YC156 suffers from high Cout as well, but theoretically has gain up
> to around 110mhz assuming I remember correctly. Perhaps this trick would
> work there as well... ?
>
> Thanks,
> Jim, N7CXI
>
> ##  No the "left hand"  end is NOT open!    The "left hand" end goes 
> between the plate block cap
> and the C1 tune cap.     This  puts  1/2 the uh   BEFORE  the main PI 
> net... and the other 1/2
> BETWEEN the  tune and load cap.    The theory is.. the tube C between 
> anode to grnded grid +
> the "left hand" end  makes a STEP DOWN  L  network.     The plate load Z 
> of the tube is now
> stepped down via the L network.. to a much lower value.     A lower value 
> plate load Z  will of course
> require a  high C /low L  main PI net.   It's amounts to a cheap and quick 
> method to transform the tubes
> sky high plate load Z.... down to something the main PI net  can actually 
> handle.
>
> ## the scheme works.. and it works very good too.  I used it on my 
> 3CX-3000A7 amp to transform the
> plate load Z  way down on 15m.  It's low enough that I can run a loaded Q 
> = 8   on 15M.     The same
> coil  tap is also used on 17M.    On  17M, the Q rises to 12.
>
> ##  when using the mfj set up to test all this, the MFJ is placed on the 
> output of the amp,   A resistor, who's
> value is equal to the plate load Z of the tube... is wired  between anode 
> to chassis, [ or output side of  plate block
> cap and chassis]  to simulate the high plate load Z of the tube  [ BEFORE 
> it get's transformed to a lower value].
> Then tweak the tune + load caps for min swr on the mfj.   OK, now leave 
> the tune+ load alone... and increase/decrease the
> freq on the MFJ,,until  the swr on the mfj  rises to 2:1  then  note the 
> freq  spread between these 2 x freqs.  When you add
> the extra uh  BEFORE the main PI net.,.. and  now transform the plate load 
> Z  way down.... then  repeat the MFJ test, you
> will find that the 2:1 swr points on the MFJ are now a LOT wider.   It's 
> just a simple way to see how you loaded Q has dropped
> simply by adding the L  network.    The L network is just the tube C + the 
> extra uh  b4 the pi-net.   The tube C is now 'free C'
>
> ##  Like Carl sez.... the L net will have a slight effect on the lower 
> bands...[ and also transform the plate load Z down a bit... but it's
> not a huge  amount]... but it's moot point. If u lower the plate load Z on 
> 10m.... by a huge amount... it will lower  quite a bit on
> 12m as well [ good].... and not quite as much on 15m [also good].... and 
> barely any lowering on 20M.  [ good too].
>
> ##  you can simulate ALL of this  using the GM3SEK  PI net calculator. 
> There is a line on there for the tube C.  There is also a line
> on there for  "stray L"  between  anode  and up to the  C1  tune cap.  You 
> can see right away, by inserting a tiny bit of stray L, that
> the resulting plate load Z  is now transformed to a lower value.   The 
> higher the value of stray L.. the lower the tube Z gets transformed to.
> You can also change the loaded Q  on his excellent PI spreadsheet.    By 
> adding just a tiny bit of stray L, you will see that you can now lower the
> loaded  Q.   Add some more stray L... then change the  loaded Q to a lower 
> value, and keep going.   I usually juggle the values, such that  the
> stray L value is the same as the Pi net coil value.... but not always.
>
> ##  you can achieve the transformation effect  by tapping the pi net  coil 
> with the  hot side of the tune cap.... OR
> you can use  2 x separate coils.   In the 3CX-3000A7 amp... I used 2 x 
> separate coils..at right angles...and different values too.
> I wound a .66 uh coil..consisting of 7 x turns of  3/8" OD tubing..with a 
> 1.5"  ID... an placed it  between the plate block
> caps and the  Tune cap.  Coil between tune and load cap was also 3/8" 
> tubing... but a lot bigger diam..like 3".
>
> ##  Point is  either method works just fine.   The gm3sek spread sheet is 
> dead on  when adding the stray L.  It seems  to
> me that the plate load Z  did not  drop to  1/4 of it's original value, 
> when  using Carl's technique.  It was WAY  down.. but not
> that low.  However,   by shifting the C1  tap point over a bit ,  you end 
> up with even more stray L... and  LESS  coil between
> tune and load cap.    Z  is now even lower... which of course requires 
> more tune + load C..and less coil.
>
> Later... Jim  VE7RF


The C1 tap can be most anyplace on the coil, and the ratio will increase the 
more you move the tap towards the output side.

Ive found the midpoint or a bit less to be a good compromise between a 
useable C1 on 10M and excessive current when heading for the low impedance 
side. That requires a heavier inductor and switch contacts but since you 
dont operate 10M its a moot point.

My interests were primarily 2x 4-1000A, 4x 4-400A and an occassional 
4CX5000A as those were popular BC pulls during my custom building days. The 
output C is also a lot less than a YC-156.

Carl
KM1H 

_______________________________________________
Amps mailing list
Amps@contesting.com
http://lists.contesting.com/mailman/listinfo/amps

<Prev in Thread] Current Thread [Next in Thread>