24-10-2023, 09:28 AM
For the 1j29b:
Though, this seems like a convenient convention only. Internally, the production difficulties for having tapered rods, or non-parallel spacings, would be significant! To prove it, I tried swapping over the filament leads with a DPDT switch, and the anode current changed by at most 2%.
I also tried it with the (single filament) 1j24b, same result.
Yes I agree - I used mine in parallel.
For a directly-heated filament, there must of necessity be a potential gradient along the filament to get it hot. So no way can it be a unipotential cathode. Ideally, the reference point would be midpoint, which of course isn't accessible, but the equivalent is the mid-point of the filament heating supply.
If this is a transformer, it's easy.
The other point is that the filament voltage 'eats up' some of the grid voltage swing - as the grid goes more negative, the positive end of the filament cuts off first while the rest of it still passes current. Unless there is compensation, such as grid rods tapered along their length, or assembled on a slant (which is already determined not to be the case), the electron beam gets progressively shut off and it takes a further 1.2V to cut off the negative end of the 1.2V filament. So even if the characteristic is completely linear up to commencement of cutoff, there'll be curvature over the last 1.2V and the cut-off can't be any sharper than this.
It's a distortion mechanism.
With filaments in series this happens over 2.4V, it's a good reason to not use this option!
I used 0.6V - 0 - 0.6V filament supply, referencing everything to the centre tap, and there's practically zero change in current as the filament voltage swaps polarity. It's square wave, which has the best possible peak:RMS ratio, so there's never any instant when the total filament voltage is greater than 1.2V.
Why? I've used g3 as a control voltage for AVC, also g2 (offset in level), and I've used g3 in a frequency changer for oscillator injection, going between 0 to -20V, and it works well. Each case, there's more than 'a few volts' on g3!
(23-10-2023, 11:46 AM)Mike Watterson Wrote: The top common connection of the two filamentary cathodes is regarded as LT+ and the separate bottoms are LT-
Though, this seems like a convenient convention only. Internally, the production difficulties for having tapered rods, or non-parallel spacings, would be significant! To prove it, I tried swapping over the filament leads with a DPDT switch, and the anode current changed by at most 2%.
I also tried it with the (single filament) 1j24b, same result.
(23-10-2023, 11:46 AM)Mike Watterson Wrote: ...series operation is slightly nonsymmetrical and not as good. They always recommend decoupling the centre tap on series operation.
Yes I agree - I used mine in parallel.
For a directly-heated filament, there must of necessity be a potential gradient along the filament to get it hot. So no way can it be a unipotential cathode. Ideally, the reference point would be midpoint, which of course isn't accessible, but the equivalent is the mid-point of the filament heating supply.
If this is a transformer, it's easy.
The other point is that the filament voltage 'eats up' some of the grid voltage swing - as the grid goes more negative, the positive end of the filament cuts off first while the rest of it still passes current. Unless there is compensation, such as grid rods tapered along their length, or assembled on a slant (which is already determined not to be the case), the electron beam gets progressively shut off and it takes a further 1.2V to cut off the negative end of the 1.2V filament. So even if the characteristic is completely linear up to commencement of cutoff, there'll be curvature over the last 1.2V and the cut-off can't be any sharper than this.
It's a distortion mechanism.
With filaments in series this happens over 2.4V, it's a good reason to not use this option!
I used 0.6V - 0 - 0.6V filament supply, referencing everything to the centre tap, and there's practically zero change in current as the filament voltage swaps polarity. It's square wave, which has the best possible peak:RMS ratio, so there's never any instant when the total filament voltage is greater than 1.2V.
(23-10-2023, 11:46 AM)Mike Watterson Wrote: Note in Pentode OR Triode mode the g3 must go to a low voltage, either f-, f+ or within + or - a few volts. So with isolated filaments the g3 ought to go to common of filaments. F3->f1, F1->f2 and F2 and F2 in the photo are shield rods.
Why? I've used g3 as a control voltage for AVC, also g2 (offset in level), and I've used g3 in a frequency changer for oscillator injection, going between 0 to -20V, and it works well. Each case, there's more than 'a few volts' on g3!
