Low vs High aka Series vs Parallel Filaments

[ak_47_boy]

I am designing with a tube that has two 6.3V filaments. 12.6V vs 6.3V. What way should i go for best sound? What way is easiest?

thanks

[Tweeker]

Easiest depends on your filament supply. Lower currents induce currents into other circuits less.

[andrew_whitham]

Hey there,

AC, DC or regulated?

12.6v makes the DC / regulating easier, - less current smaller components and so on.

Individually controlling 6.3v to each triode section sometimes helps with the matching between sections. Although I've never tested across an individual tube, I've seen a couple of 6.3v tubes in series (oh dear) where the second tube had only 5v on the heater pins...

BTW AK_47_boy. is this a power output tube or a pre-amp tube? how much current are we talking about here?

Andy

[Wavebourn]

Originally posted by andrew_whitham
I've seen a couple of 6.3v tubes in series (oh dear) where the second tube had only 5v on the heater pins...

You are lucky Andy, I've never seen that.

If the tube have separate filaments like 12AX7 there is no harm in connecting them in series, the tube is designed for such operations. I even use 6P3S-E tubes with filaments in series to power them from the same 12.6V source I use to rectify and regulate for small tubes. However, I have to use similar tubes, for example don't mix 6P3S-E and Chinese 6L6GC, but it is not a problem since anyway it is better to use matched tubes in the output stage.

[andrew_whitham]

Originally posted by Wavebourn


You are lucky Andy, I've never seen that.

Yeah I felt lucky too

Maybe I wasn't clear these were separate tubes - both JAN 6922's for the record. Not a combined tube,

However it got me (at first) thinking that the same problem might exist in a shared envelope. I know it shouldnt, but. On the tubes I've seen in bits they have 2 separate heaters linked by a bar under the mica. I guess that the 2 heaters are not necesarilly forced to be *identical* but they must be from roughly the same production run - close enough then?

Incidentally wavebourn, do you current or voltage regulate? current regulation of multiple tubes in series seems to be logical, although I've never tried it. (and I cant find that reference either!)

Cheers

Andy

[ak_47_boy]

Its a directly heated thoriated filament output tube.

[Wavebourn]

Originally posted by andrew_whitham


However it got me (at first) thinking that the same problem might exist in a shared envelope. I know it shouldnt, but. On the tubes I've seen in bits they have 2 separate heaters linked by a bar under the mica. I guess that the 2 heaters are not necesarilly forced to be *identical* but they must be from roughly the same production run - close enough then?

Factory rejects out of specs?

Incidentally wavebourn, do you current or voltage regulate? current regulation of multiple tubes in series seems to be logical, although I've never tried it. (and I cant find that reference either!)

Andy

I regulate a voltage, but limit a current on the level just above needed for hot tubes though they start slower.

[Dave Cigna]

Originally posted by ak_47_boy
Its a directly heated thoriated filament output tube.

Then I would definitely parallel them. With DC heating a series connection would put the two 'cathodes' 6.3V apart making it difficult or impossible to get them both to bias up the same.

With AC heating, a series connection makes it impossible to null heater supply hum. The best you could do is to ground the center connection between the two filaments. In that case, one end of each would be grounded (obviously) while the other end would be swinging 6.3V. That's the same as applying a 3.15V 60Hz signal between grid and cathode. It makes the amp hum.

What you want to do is create a virtual ground at the center point of each filament so that either end is swinging 3.15V and the two ends are out of phase. The grid to cathode voltage (averaged across the entire filament) remains fairly constant thus reducing hum. This is typically accomplished by grounding the CT of the filament transformer, or putting a pair of equal resistors across the filament with the center connection grounded, or using a hum pot with the wiper grounded.

-- Dave

[andrew_whitham]

Originally posted by Wavebourn

Factory rejects out of specs?

I wouldnt like to comment they were fitted as 'standard' in a particularly well known brand... Probably just my dumb luck

I never got a pair with "matched" heaters though in the end I gave up when I got close enough

Andy

[Tweeker]

5687 is an example of a tube with heaters allowing series or parallel heater connections.