[AMPS] CT filament secondary

[John T. M. Lyles]

Arden, Bill,
Adding a hum bucking circuit is plausable, it's done in some older
commercial b'dcast transmitters. But it must be tweaked for each tube,since
it is not exactly feedback, but a method of 'feedforward' noise correction.
I did it in a commercial product, had a small rheostat to control the hum
injection. It lowered AM noise of an FM rig from -48 to about -60 dB,
depending on the care to which it was set up.

I made a mistake last summer, when my filament pigs showed up for the 100
KW tetrode stages I built. No center tap - Argh! I scrambled to see if the
manufacturer could rewind the secondaries, but it was very costly. The turn
around for new iron from that company was on the order of 12 weeks at the
time, so I was stuck. Figured I could live with the hum modulation during
testing, and if I reversed the phase on one filament, the overall ripple
would cancel in the push pull output when they are combined in a balanced
load. But it meant each tetrode would have this AC modulation. I figured it
was as simple as taking the filament AC voltage, and superimposing that on
the grid conduction bias point, which was -120 VDC. It results in maybe 10%
AM worse
case. Its this 60 Hz component that is a pain, as it will ride on the
filament, end to end will have a potential difference. No matter what you
do with chokes and RF caps, to get the RF current to ground (that should be
close to the socket anyhow, not at the secondary), you have to deal with
the AC component of the heater voltage over the cathode current ground
return. One side of your transformer will be grounded to get the DC cathode
current back to the power supply.  Thats why center tap is needed on the
secondary of the transformer. Or you may be lucky and have a tube with a
center tap of the filament, but those were typically in receiving tubes.
You can also make up a resistor network to ground, and make a fake center
tap. It will dissipate power. I haven't calculated this, but the best way
is to bite the bullet and use the CT transformer to do this right. I did,
found another source of transformers, custom, in about 3 weeks delivery,
with < 1/2 the cost of the original ones. ($500 versus over
$1500/transformer). The tranformer has tapped 440/480 volt primary, 60 Hz.
So I have a pile of non centertapped 15 VAC, 300 Amp transformers, that are
useless except maybe to use for welding or charging batteries - 440 VAC
primaries make them obsolete for home use!

73

John
K5PRO


>Maybe a little filament transformer derived AC could be bled into the grid
>return to offset the effect of the unbalanced filament.  Sorta like putting
>the grid-cathode potential diffrence mid-span between the filament
>endpoints?  Is this the hair-brained idea of a whacko?
>
>Arden  Allen   KB6NAX   Vallejo, CA   gumbear@pacbell.net
>
>From: "Mike Feher" <n4fs@monmouth.com>
>To: Old Tube Radios <boatanchors@theporch.com>
>Subject: Re: Fil. Xformer w/o center tap
>Date: Tue, 8 Sep 1998 12:47:11 -0400
>MIME-Version: 1.0
>Content-Type: text/plain;
>	charset="iso-8859-1"
>Content-Transfer-Encoding: 7bit
>
>How about a switch? For one month one side is grounded and then for the next
>the other and so on. 73 - Mike
>
>Mike Feher, N4FS
>89 Arnold Blvd.
>Howell, NJ, 07731
>732-901-9193
>
>Message-Id: <199809081934.MAA11901@mail-gw5.pacbell.net>
>From: "Arden Allen" <gumbear@pacbell.net>
>To: Old Tube Radios <boatanchors@theporch.com>
>Subject: Re: Fil. Xformer w/o center tap
>Date: Tue, 8 Sep 1998 12:36:55 -0700
>MIME-Version: 1.0
>Content-Type: text/plain; charset=ISO-8859-1
>Content-Transfer-Encoding: 7bit
>
>Hi Ken;
>
>> One solution to a "no center tap" filament transformer is to put
>> r.f. chokes in series with the filament leads, then a pair of r.f.
>> coupling capacitors in series across the leads between the
>> chokes and the filament. The r.f. return connects between the
>> two r.f. caps.
>
>It's not the cathode circuit RF impedance Bill is concerned with, it's the
>effect of 60 cycle modulation of the RF signal envelope due to *imperfect
>linearity* (power gain shift) with bias.  With a center tapped filament
>transformer the hum modulation is cancelled to a great extent due to part
>of the filament going positive while part is going negative.


--
FAQ on WWW:               http://www.contesting.com/ampfaq.html
Submissions:              amps@contesting.com
Administrative requests:  amps-REQUEST@contesting.com
Problems:                 owner-amps@contesting.com
Search:                   http://www.contesting.com/km9p/search.htm