Hello, I hope the schematic makes it clear. Pairs of 6,3V heaters connected in series and referenced to ground or DC elevation. From upper to lower, the first shows both centers referenced, the second only one and the third leaves that to a pair of resistors. What do you think would be the right thing to do?
Attachments
I've been using third option, but with much higher resistance, from 10kohms to as high as 100kohms. Then I include a cap to GND. If the output tube is cathode biased, I just use the cathode voltage as bias, if not I usually voltage divide from b+ to about 80V.
Has been faultless for me...
Really don't think it's too critical, any of those will do the job.
Has been faultless for me...
Really don't think it's too critical, any of those will do the job.
To be assanine I would use separate resistors for each channel, then decouple to each channel's local ref node ('GND'). But that may be overdoing it.
The trick works b/c any free electrons escaping the heater will be at a potential higher than the cathode (which the stray electrons are closest to) and thus will not be atracted to it. Having the heater at a potential equal to, or lower than, the cathodes, will often atract some to the cathode, and you get mains hum into your sensitive tube. So the point is simply to keep the heaters at a potential above the cathodes, and nit picking about how that is done is fun but perhaps not critical.
The trick works b/c any free electrons escaping the heater will be at a potential higher than the cathode (which the stray electrons are closest to) and thus will not be atracted to it. Having the heater at a potential equal to, or lower than, the cathodes, will often atract some to the cathode, and you get mains hum into your sensitive tube. So the point is simply to keep the heaters at a potential above the cathodes, and nit picking about how that is done is fun but perhaps not critical.
One concern is your use of 12.6V, rather than 6.3V. Any stray capacitance, either through the cathode heater interface, or to the grid from heater circuit path, will cause more hum current to flow when the AC voltage difference is higher. Sometimes that hum current can be minimised by using a tuned humdinger, especially if you have asymmetric hum current coupling.
Heater elevation has no influence on stray capacitance (only on stray heater-cathode resistance).
Heater elevation has no influence on stray capacitance (only on stray heater-cathode resistance).
Last edited:
In my small experience I have found that just grounding the heaters almost always cures hum. I use DC elevation only when the maximum cathode to heater voltage for a given tube is exceeded. In one case where Vkh max is 200V and the cathodes sit at 240V it was very convenient for me to refer heaters at a nearby 200V. You suggest this is not a good idea?
oh absolutely, you want to keep heater to cathode ratings within specs.
I'm sure grounding heaters is usually the best solution. Simplest is best. I have more or less fallen into a habit after having success with elevation. And after reading about it recommended in the RCA manual, well it's stuck with me. Whatever works for you works.
I'm sure grounding heaters is usually the best solution. Simplest is best. I have more or less fallen into a habit after having success with elevation. And after reading about it recommended in the RCA manual, well it's stuck with me. Whatever works for you works.
I tend to use one resistor pair pseudo-tap per winding, right at the tube socket, then run a small wire to either ground or the DC reference. Using Toroidal transformers like those from Antek gives you two windings, so it's pretty convenient to simply use one per channel in a stereo amplifier. it's tidy, too
I tend to split the difference between the ground and cathode voltages to keep it all about the same potential, and I find I end up anywhere from 50~80 volts in most builds. This assumes concertina, cathode followers, etc, stuff with a steady DC point on the elevated cathode. If no elevation is required I tend to reference the tap to ground, except in low level stuff like phono stages or DAC I/V stages with low-value I/V resistors and high-mu tubes (similar to a phono setup)
How exactly can it cause problems when done right? I'm not saying that you're implicitly wrong, but source materials and prior experience tends to disagree. Elevated heater reference reverse-biases the internal filament-to-cathode diode relationship, and can reduce hum in many cases versus simply tying the filament to ground (through a center tapped winding or through resistor pseudo tap) without causing any issues in performance or longevity.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.