Cathode connections on directly heated power tubes...

[K8DI]

Well, I've gone and started another build.  I am starting with iron from an old National Radio VHF Aviation band transmitter. I discovered that was where the transformers came from after I found and obtained them for a great price. The best I can tell, the unit used gas or mercury rectifiers (filament trans has a pair of 2.5v/10a windings). I believe it used a pair of 4-65A tetrodes, modulated by a pair of 809 triodes.  Those all have 6.3v filaments, and there's a pair of 6.3v/10a windings as well.  There were two plate transformers, each 1500VCT at 350mA, hence four rectifiers, each pair on a filament winding. It also had four 10Hy chokes, which supports the idea of choke input filters and vapor/gas rectifiers, but I wasn't able to get them/they were long gone. I sourced a pair of 8Hy/80 ohm/300mA chokes to take their place, along with a few 866As and 812As (which I will use in place of the 809s), from a generous ham friend.

The thing is, 4-65A's, 809's and the expected 866A/3B28/etc. all have the filament as cathode. I am used to seeing in such cases, that the HV return, the "cathode", that you would connect to ground through a meter, you'd use the filament center tap....but none of those filament windings have a center tap.  So...I can just tie one end to ground through a meter/etc...  But, how well will that work/what's the downside?  Which end of the filament goes to ground, is there a preference?  any comments or experiences with this particular dilemma?

Other design notes:  transformer audio in to differential amplification: 6N7 driving one of 6K6/6F6/6V6, then the modulators though a transformer PP plates to PP grids. I have a multi-match Thordarson on hand for that. As mentioned, the 812As for the modulators.  The mod iron is a standard ratio unit.  The RF side is looking like an undetermined small signal tube or possibly solid state amp driven by a 0dBm source, then into a 1614, to drive the pair of 4-65As. I'll parallel them, because I don't know what I am doing around PP RF circuits...Then standard pi net output.  The goal is 20m-15m-10m, if everything plays nice and I get lucky with layout, I'll try to get 6m out of the thing too.

So, let me know what you think about grounding one side of a 4-65A filament...

Ed

[KA3EKH]

Think the thing is that you have to use a filament transformer with a center tap or have a couple resistors on each leg of the filament of equal value or bad things will happen. Saw one that had a filament transformer without a center tap but had a low value pot between both legs and the center tap grounded, something like a hum balance pot.
Don’t know if I would put a plate current meter in the filament side being you will also see screen current at the same time. Would want to see screen and plate current separate myself.

[WQ9E]

Another approach is to use a small 12.6V filament transformer with a center tapped secondary and wire its secondary in parallel with your amplifier transformer.  This provides the needed center tap for metering (and tube operation since the cathode will need a ground reference for tube conduction to happen).

I added an external screen current meter to my Johnson Desk KW in order to directly read screen current.  Johnson provides metering of cathode and control grid current so the difference between the two is divided between the screen and plate current and they assume it will end up at the desired level by following the rough output network settings in the manual.

Screen current is an excellent indicator of proper tuning of the amp; I put a low value resistor in series with the screen and use that to drive the external screen current meter.

Rodger WQ9E

[MikeKE0ZUinkcmo]

Both the 4-65 and 809 have 6V filaments, so you would need a 6.3V center tapped transformer secondary, and a 1A current rating would be fine since the transformer is there just to make the center tap, not supply any power to the tubes.   The only current through the little transformer's windings would be equal to the plate current plus screen current.   Nothing more.   Just insulate the primary winding wires and tie them out of the way.

As Rodger indicated you just wire the small transformer in parallel with the 6V-10A secondary and now you have a center tap ground return/meter connection.  If you use both secondaries, one for the 809s and one for the 4-65s you'll need a small transformer transformer for each.

The old Dentron amplifiers (schematics below) had non-center tap filament transformers with one side "grounded", and they had about 15% AC hum with no audio input.  



The "fix" was as Rodger indicated in his post.

[K8DI]

Both the 4-65 and 809 have 6V filaments, so you would need a 6.3V center tapped transformer secondary, and a 1A current rating would be fine since the transformer is there just to make the center tap, not supply any power to the tubes.   The only current through the little transformer's windings would be equal to the plate current plus screen current.   Nothing more.   Just insulate the primary winding wires and tie them out of the way.

As Rodger indicated you just wire the small transformer in parallel with the 6V-10A secondary and now you have a center tap ground return/meter connection.  If you use both secondaries, one for the 809s and one for the 4-65s you'll need a small transformer transformer for each.

I’ve realised this method has a second value, by not insulating the primaries…but deriving bias from them.  My junk box had a 12.6v and an 18v unit; giving about 60vac and 40vac to do bias supplies with. Bleeders will be chosen to keep the secondary current and the actual filament source winding current within spec.  Before this I was already planning on having to add a bias supply…two birds with one transformer!

Ed

[KD6VXI]

Let me add here:

You don't have to match voltage on the CT transformer.
If you have, say, 5 volt filaments. (4-xxxx tubes say) and a 9 volt transformer with a CT, you can wire the secondary of the 9 volt in parallel with the 5 volt fils xformer and it functions as a CT just fine.

I've not gone wildly mismatched, but was able to do 2 to 1 voltage with no issue.

--Shane
WP2ASS / ex KD6VXI