For a SE amplifier OPT pair, avoid shipping Edcor from USA to Italy, you could check don-audio.com (has some Edcor on stock) or ogonowski.eu (great products and service).
I think a litte then yes I could ride this way...
The tube voltage regulator for Vg2 seems a clever solution...I also have seen somwhere the use of Zd but I don't know much about it. I'm not familiar with full pentodes and how do they play with a speaker system.🙂
For Cbeuche: I know DonAudio and it has a good deal of OT. Also here there are some DIY builders I could get in touch. Novarria of Milano (Enovaz) instead is good for power transformers, not much for output transformers.
Full pentode mode exhibits a high O/P impedance (poor damping factor) and, therefore, loop voltage NFB of some kind is essential. Distortion concerns are another reason for NFB, but, even if linearity was not an issue, the matter of damping factor mandates its use.
Useful info. regarding gas discharge regulator tubes can be found here. While the 0A2 is not described in GE's writeup, it is quite close to the described 0D3/VR150 in behavior.
Useful info. regarding gas discharge regulator tubes can be found here. While the 0A2 is not described in GE's writeup, it is quite close to the described 0D3/VR150 in behavior.
Ok!
Just for sake of information I found this document comparing different spectra of armonics for some kinds of tubes, operating in a test amplifier. It seemed interesting to me.
Alternatively you can just use 5W 120V (or whatever you need for dropping to 150V from your B+) zener diodes instead of R7 (post #19).
Cheaper, simpler, quite stable (depends just about only on the B+ stability) and actually less noisy than gas regulators. Definitely less pretty though 🙂
Cheaper, simpler, quite stable (depends just about only on the B+ stability) and actually less noisy than gas regulators. Definitely less pretty though 🙂
Dropping from B+ through a series Zener does not work well. It increases relative instability of G2 voltage, instead of decreasing it by a shunt regulator that we need.
However, it is still a good idea to try both and compare, nothing is better than own experience.
However, it is still a good idea to try both and compare, nothing is better than own experience.
I read it. It's interesting but I dont think you can draw much of a conclusion from it.
The 5k OPT is pretty much perfect for best THD performance when 6P1P is used, for example, but not the best load for 6P3S.
Which may have more to do with the author's conclusions, than anything else.
It would've been nice to see the distortion figures compared with the gains, then you'd see which valve has the best intrinsic/inherent 'best' harmonic spectra per unit gain.
In other words I'd guess that means the best geometry/alignment of that geometry.
Also as Wavebourn has just mentioned (cross posting), a proper zener regulator shunting to ground is better, you can bypass it, and then no worries about zener noise.
Last edited:
Something like 1N5380B has the voltage variation of about 1V in the 0-10mA range, which is more than enough for 6P15P operation. So like I said, G2 voltage stability depends mostly on B+ stability - while it's not perfect, it's OK for a simple project.
I read yesterday an article that say instead the use a 6.3V row of zeners, because this value has the more linear behaviour between the 2 regions (zener effect vs avalanche effect), so the minimum noise and highest stability. I haven't gone further into this and I cannot say, one should experiment and see...
The two other possibilities are just as good, shunt or 0D3 tube regulation. The latter as Eli Duttman suggested make also sense to me, in a "all-tube" circuit. I must only consider which one is better on the practical and economical aspect 🙂
Noise is absolutely no issue since you feed the screen grid of an output tube, not a control grid. By the way, gas discharge regulators are more noisy than Zener diodes.
There's a reason why these TV HD finals never had any audio applications mentioned. The most linear part of the plate characteristics are well within red plate territory. For the 6CD6, that would mean around VPK= 250V and IPq= 0.2A for 50W. The 6CD6 is specced for PD= 20W. While these types are rated conservatively, considering the "brick on the key" operation, you can spec bust by about 50%. Going 250% over spec is a bit much. It would probably sound quite good, but not for very long. You could get the screen voltage down, but you'd lose output as well. TV HD types need to be used as PP pairs that can hit the most linear parts of the characteristic without needing red plate bias.
That's true, but I can hardly imagine 10% B+ sag in a class A amp of any decent design...
And if you mean mains voltage variation, regulating (truly) just the G2 voltage won't help much.
I generally use 1N53**, in 24V variety (I forget the exact p/n) to make a ladder to ground, with appropriate resistance, and tap off where I want, then bypass.
I think it works better, but I cant prove it.
And counter to Wavebourns notion, although yes there are tolerances WRT the regulated Voltage vs diode current, at least at the smaller end of the scale, where my amps have been focused, I see less than a volt sag, with HT at 300V, at maximum output (SET)
The sag would likely be worse in some kind of PP amp, but I'm not sure how much.
Caveat being, you design the zener shunt to operate at X current, assuming HT is fixed and constant.
Current isn't quite constant, but close enough to know your stabilisation voltage after warm up, will be pretty much unchanged each time.
In my experience, small Watts, the change with supply voltage variation, start up, warm up to temp co stabilisation, is greater than the sag.
I.e. that is to say, start up to fully temperature stable, each time, is maybe +/-2V (pretty insignificant) and sag at full output, even less than that, at 1V.
Sag at clipping, heavier than I would listen to, is more severe.
Last edited:
Well; there are many different ways to skin the same cat. ;-)
Try different currents and voltages, but do not forget that nested feedback loops do exist. What we need, neither very low THD and harsh clipping with too much feedback, nor huge intermods with no feedback. We need as low as possible distortions that in dynamics mimic the distortion profile of our own ears.
I could not find the video of Edelweiss-2 with 6CD6 tubes, but here is one with 4P10S. The same approach. The guy who posted videos removed them due to copyright for the music he played.
Wavebourn Edelweiss-2
The reason why they were not used (except by such innovators as Philips, for example), because of copy-paste design approach.
Last edited: