I've seen a Zener in Blencowe writing as a method of producing DC to elevate filament ac by the DC level. Funny I don't remember filtering for it...but maybe it's not necessary, considering it's usually fully ac. I just remembered this after thinking about halfwave rectification producing significant DC flux. That always made me think adding a second diode (or even 3 more!) to a HWR has more benefits...lower ripple, higher ripple freq., and might result in a cooler transformer...but that's all hunch and no data to back that up.Off topic a bit but I recently saw, a tube heater scheme iirc, and it had a diode hanging off one of the transformer terminals. Said something about helping to keep the transformer from magnetizing. I can't find where I saw it. Any insight on this?