[Amps] high frequency filament excitation TSPA

Mon Jul 17 09:29:47 EDT 2006

Thanks Steve,
   Definitely no offence taken but much gratitude to your response is at
hand. I'm definitely not an electronic engineer, thus my bucket is pretty
empty in that respect and any and all help is appreciated.
   Yes, I believe the diode on the isolated end isn't necessary unless the
tube has a directly heated cathode. Although the diodes on the isolated end
are totally unnecessary in the 5000's case, I don't think enough ac could
slip through them to be detected, but I can see where a small amount would
pass. I'm not sure that the ZSAC would go up if I clipped the diodes on the
isolated end. When time allows I will do that and give you a report. I must
admit that at the time of construction I didn't know the 5000 had an
indirectly heated cathode and oxide coated filament, certainly something I
should have been aware of.
Regards,
Larry N5BIP

-----Original Message-----
From: Steve Thompson [mailto:g8gsq at eltac.co.uk] 
Sent: Monday, July 17, 2006 2:33 AM
To: Larry Carman
Subject: Re: [Amps] high frequency filament excitation TSPA

Hi, Larry.

 From what you describe, I think the diode from the isolated end of the 
heater isn't needed. I think it will be injecting half wave rectified 
heater volts into the bias chain - the amount of hum will probably be 
small as it's a low voltage compared with the 40V of bias you need.

I mean no offence by offering these comments and I apologise in advance 
if I'm telling you things you already know.

In a directly heated tube, the filament is the cathode and we have to 
heat it to make it work - but all the tube current flows through the 
filament as well. If we put dc on the filament to heat it, the voltage 
to the grid is different at one end than the other, so the current tends 
to flow more at one end - that end gets worn out faster, and you can't 
get the full current capacity. By using ac, the polarity reverses 60x 
per second, so the current concentration is swapped from one end to the 
other 60x per second and the filament gets equal use at both ends. By 
having a ct transformer, the + and - at opposite ends of the filament 
cancel out as far as the *average* voltage to the grid is concerned, so 
there's no hum.

With an indirectly heated cathode, such as in your '5000, the heater is 
just there to make the cathode hot. It's not in the tube's current path 
and the heater voltage doesn't affect the cathode-grid bias voltage so 
we can use either ac or dc for the heater supply. If we use ac there's 
no need for a ct transformer, the ct is only needed to make the average 
voltage zero between a directly heated filament and grid. The isolated 
end of the heater is meant to be just that and the biassing etc. should 
come only off the cathode connection.

If you remove the diode from the isolated end of the heater, you will 
probably find that the standing current goes up and you will need to add 
some extra diodes to get back to the right value.

73, Steve

Larry Carman wrote:
> I'll probably get some rebuttals from this but here goes.  I constructed a
> 3cpx5000a7 amp a couple years ago. The 15 VAC transformer I used for
warming
> the filament was not center tapped. Although the 3cpx5000 has it's cathode
> tied to one side of the filament and should be biased accordingly, I used
> diodes with there cathodes attached to each side of the filament windings
> then tied all there anodes to a common point which served as an artificial
> center tap ( for lack of another term ). To regulate bias voltage to the
> cathode of the 3cpx5000 I used two HV diode modules in series, attaching
the
> anode of these to the artificial center tap. With the HV module I obtained
> 40V DC bias via the B- and artificial center tap. At about 5KVDC plate
> voltage I get a little over 300 mils ZSAC with this arrangement. Thus far
I
> haven't had any complaints of hum and all seems to work well. I didn't
have
> an oscilloscope to verify at the time of construction. Maybe everyone's
just
> being nice to me.
> Regards,
> Larry N5BIP  
> 
> -----Original Message-----
> From: amps-bounces at contesting.com [mailto:amps-bounces at contesting.com] On
> Behalf Of R L Measures
> Sent: Thursday, July 13, 2006 4:25 PM
> To: Tom W8JI
> Cc: amps at contesting.com; John T. M. Lyles
> Subject: Re: [Amps] high frequency filament excitation TSPA
> 
> 
> On Jul 13, 2006, at 12:10 PM, Tom W8JI wrote:
> 
> 
>>Another problem  people commonly miss is voltage drop in the
>>filament choke.
>>
>>A normal properly sized filament choke has a few tenths of a
>>volt AC drop across the windings from filament current.
>>
>>This is meaningless in a center tapped filament transformer
>>feeding a directly heated tube, but when one side of a
>>heater is connected to the cathode you either need a
>>hum-compensating pot at the DC return for the cathode, you
>>need a third winding only for dc that ties to the filament
>>pin common th the cathode (a trifilar choke), or an entirely
>>separate dc return choke that does NOT carry any filament
>>current.
>>
>>It's surprising how little ripple voltage in bias at the
>>cathode can cause hum on the signal.
>>
>>By the way, I see quite a few Internet pages on constructing
>>amps with Russian tubes that ignore this problem. It's also
>>a problem with the 3CX5000, 3CPX5000 series of
>>heater/cathode tubes that have the cathode tied to one side
>>of the heater. You can't run the cathode DC back through the
>>same choke winding as the heter uses without inducing some
>>unwanted hum. The only exception is when the transformer end
>>has a hum-balancing pot.
> 
> 
> Another solution is to use a separate RFC to carry the DC HV negative  
> directly to the cathode so that there is no DC in the heater's  
> bifilar RFC.   Since the cathode Z is 50-ohms, only c. 30uH is needed  
> to cover 160m - 10m.
> 
>>I know a fellow in Mississippi who worked on his 3CPX5000
>>for weeks and never figured that out. His solution was to
>>run a filtered DC supply!
>>
>>73 Tom ...
> 
> 
> R L MEASURES, AG6K. 805-386-3734
> r at somis.org
> 
> 
> 
> _______________________________________________
> Amps mailing list
> Amps at contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
> 
> _______________________________________________
> Amps mailing list
> Amps at contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
> 
> 