[Amps] 200-ohm tank and "un-un" in HF amp design?

[Jim W7RY]

See below.
73
Jim W7RY


On 10/20/2013 10:01 PM, Jim Thomson wrote:
> Date: Sun, 20 Oct 2013 00:51:44 -0700
> From: Jim Barber <audioguy at q.com>
> To: "Amps at contesting.com" <Amps at contesting.com>
> Subject: [Amps] 200-ohm tank and "un-un" in HF amp design?
>
> Assuming adequate core/wire sizing and airflow, can anyone think of a
> solid reason *not* to design a legal-limit HF amp output tank for 200
> ohms, then use a 4:1 "un-un" transformer to step it down to 50?
>
> I'm curious because doing so would bring the loading cap requirement for
> a 4000-ohm plate impedance down to under 1000pF at 1.8 mHz, which can be
> done with a large air variable alone.
>
> Note: Yes, I know it *has* been done already, but I'm asking for
> opinions as to whether you think it should be.
>
> Thanks and 73,
> Jim N7CXI
>
> ##  years ago,  QST  tested a whole bunch of linear amps in one issue of the magazine.
> It may have been the QRO or Commander brand amp that had the 4:1  un-un in it.  Guess what, it overheated
> badly and cracked in half.   The idea of course was with a PI net based on a 200 ohm output
> instead of a 50 ohm output, the load cap  value is way less.  Then the broadband 4:1 un-un
> takes it from 200 ohms...down to 50 ohms.
>
> ##  I think it was beefed up after that, and is still used.   With a sky high 4000 ohm plate load Z
> your tune + load cap requirements  will be a LOT lower vs  say a 1800 ohm plate load Z.
>
> ##  If it was me,  I would use a conventional PI net...and just pad the load cap.  Padding a load
> cap is easy, piece of cake really.  IF a NON shorting bandswitch is used, simply use the 160m
> contact to switch in the 160M padder.  The 160m contact isnt used anyway... since the exact amount
> of inductance is used for 160m.   IE: add up all the various coils and you have the correct amount for 160m.
> The band switch is only used to tap the 80-10m bands.  We used this scheme on the 3CX-6000A7 amp.
>
> ## IF a shorting switch is used, the 160m contact cant be used to switch in a 160m padder.... cuz of the
> multiple pole rotor used on a shorting switch.  Its kinda a waste to add an additional wafer  just so you
> can utilize one contact on 160m for padding.   Plan B is to simply use a SPST relay or a DPST relay....
> with all contacts in parallel...and use the relays contacts to switch in 160m NPO padder capS.   We
> used the relay scheme on the 3CX-3000A7 amp.  SPST- 40A  contactor was used.   A standard
> 30A  relay, like a spst or dpst has also been used in the past.   You can also use a small, shorting
> style bandswitch to progressively add  160m padderS.   IE:  3 x 170pf padders.  170- 240-310 pf.
>
> ##  as noted, the 200 ohm un-un scheme means the peak V across the 200 ohm loading cap will be
> exactly DOUBLEe that of  a load cap that’s at a 50 ohm point.   The plate spacing will have to be at least DOUBLE,
> and the capacitance will  now drop to just ONE QUARTER.  You just lost it right there.   Cuz now you have to
> Double or triple the size of the cap to get the C back up.
>
> ##  BTW, the op said the plate load Z was 4000 ohms.   He didn’t say it was 4 kv.   With a 4k ohm load, vs
> a 2 k ohm load, the tune + load C  will be one half anyway.....and the required uh in the tank coil will be double...
> this is for a normal PI net..with a 50 ohm output...and no un-un used.



XXXXXXXXXX HE most certanlly did! Please see the email where I asked him 
what tube, plate voltage and current.


>
> ##  another method that uses less load C is to use a PI-L.... but I don’t recommend it.   Any PI –L  will  typ
> put the load cap at somewhere between 200-400 ohms. The main tank coil ends up being even more uh than a PI,
> then you also end up with the hassle of bandswitching the 2nd coil in the PI-L....pita.   The touted higher harmonic suppression
> of a PI-L is mainly a myth.....a plane jane PI  is more than ample for harmonic suppression.
>
> ##  Building a 200:50 ohm un-un that will handle 1.5 kw CCS  on all bands  160-10M..with SWR thrown in on top of that
> is doable though.  Array solutions has several  200:50 ohm un-uns that will do the job..and rated for  3-5-10-20 kw if you
> need it.
>
> ###  Another possible solution is to use a PI-L on 160m only...so a smaller load cap can be used.   Ameritron does that on
> some of their amps.
>
> ##  IMO... stick with a conventional PI-NET.....and just pad the load cap  with a bunch of surplus NPO 5 kv ceramic doorknob caps,
> then use a relay to switch em in.     Allen bond at max gain systems and also fair radio in lima ohio have them.   You cant blow em up,
> and they don’t drift either.   N470 caps  ( N470 has a slightly worse TC  than a NPO..and will also work as padders..and still
> wont drift..with the provisio that you use 2-4 of em in parallel.)
>
> ##  I would  avoid torroids in the main tank coil too...and ditto with the use of torroids in the L portion of any PI-L.
>
> later.... Jim   VE7RF
>   
>
> _______________________________________________
> Amps mailing list
> Amps at contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
>
>