Seems that you don't need a preamp at all, that's good and not my case BTW.
As I use compound gain stages, mainly cascode and SRPP I must elevate heaters, no choice.
As I use compound gain stages, mainly cascode and SRPP I must elevate heaters, no choice.
I'm not out to persuade you to use the 431, . . . . . and this IS your thread after all. Many I respect and learn from say the same as you, but in defense of the little guy . . . . . well. . . . on the ones I use I work it out to about 2uV RMS, and that level of noise just doesn't seem significant in a tube amp.
And if your English isn't good I sure didn't notice it. 😉
I have evaluated a TL431 based shunt regulated power supply vs a stack of 9V batteries in a low noise amplifier with a very high transconductance triode. In both cases, the 1/f noise of the triode dominated the measurable noise.
So let's say you get the worst case scenario of 8uV p-p of noise in your HV supply. Now you have the output stage which acts as an AC voltage divider of sorts for that noise, and you step it down through the output transformer.
How quiet is quiet enough for you?
So let's say you get the worst case scenario of 8uV p-p of noise in your HV supply. Now you have the output stage which acts as an AC voltage divider of sorts for that noise, and you step it down through the output transformer.
How quiet is quiet enough for you?
Thank you very much! Nonetheless sometimes I have the feeling that people need a cryptologist to read my posts.
Here the equivalent input noise for some valves
The Sound of Silence: Lowest-Noise RIAA Phono-Amps: Designer's Guide - Burkhard Vogel - Google Libros
The TL431 is two orders of magnitude higher. Good for TV's SMPS.
i) How do you obtain, lets say, 400V 150mA from a TL431?
ii) A power amplifier is a bit more complex than its power output stage, what about a cascode or SRPP input stage on the same amp?
As low as possible, I cannot give you a precise number, but this number decreases as you decrease the noise in every section, so discarding noisy solutions can work, isn't it?
Just out of curiosity, do you have an answer for my original question?
The datasheet covers how to deal with higher current, others here have mentioned putting the TL431 under the cathode of a tube that sits across your regulated B+ as a method for dealing with the voltage limitations.
Again, you're having a nervous breakdown about few uV of noise that you can't hear and won't be able to measure.
If you want to discard sources of noise, then the tubes themselves need to go! You can make a very quiet amplifier with transistors.
What was your original question?
This is really hilarious, you don't have an answer, and I don't like the sound of electrolytic capacitors. 😀
Putting a 10000μF cap from EL34 cathodes to ground is indeed changing the design, and it does not work at all, it ruins the original design 🙄
If you are so confident about your regulator design, why ask the question in the first place?
And yes, "ruining" a design in Spice is educational.
And yes, "ruining" a design in Spice is educational.
The original design should work quite well with DC heaters, I have no idea with AC heaters, so I did ask.
If you don't know the answer, please stop trolling.
If you don't know the answer, please stop trolling.
Popilin,
'Just took a look at your first post again. For the AC heater hum , just try putting a positive bias voltage on the heater.
(I usually use something around 30VDC on a 6.3V heater. As low as 15V can be enough though on one tube I recall it took almost 50V to completely stop it.)
The idea is that there is a virtual diode between the heater and cathode and as it's not there by design, it's not a very good one and easy to saturate so that it can't conduct the hum. My experience is that it often eliminates any audible hum.
'Just took a look at your first post again. For the AC heater hum , just try putting a positive bias voltage on the heater.
(I usually use something around 30VDC on a 6.3V heater. As low as 15V can be enough though on one tube I recall it took almost 50V to completely stop it.)
The idea is that there is a virtual diode between the heater and cathode and as it's not there by design, it's not a very good one and easy to saturate so that it can't conduct the hum. My experience is that it often eliminates any audible hum.
Cathodes are at about 25V above ground, so 50V should be enough, but in addition to the "diode" between cathode and heater you have a capacitance, EL34 datasheet says 10pF, which is in the original simulation, the time varying magnetic field produced by the filament was ignored
Bac = (Uac x 10⁸) / [√2 π fo S N]
If the filament forms a single spire (N=1), the area into the loop (S) could be really small and AC magnetic field could be very high, the cathode acts as a shield, and grids are grounded anyway, but this transmision mechanism cannot be ignored.
Regardless all countermeasures, you always have an AC voltage on the cathode, but this does not appears into the simulation, maybe by the shunt action of the bottom MOSFET, which motivates this thread.
Bac = (Uac x 10⁸) / [√2 π fo S N]
If the filament forms a single spire (N=1), the area into the loop (S) could be really small and AC magnetic field could be very high, the cathode acts as a shield, and grids are grounded anyway, but this transmision mechanism cannot be ignored.
Regardless all countermeasures, you always have an AC voltage on the cathode, but this does not appears into the simulation, maybe by the shunt action of the bottom MOSFET, which motivates this thread.
Well, you obviously know a lot more than I do, but regarding your last comment above I would ask then if choosing a MOSFET with high capacitance would be a good idea.
No, not really, I did study physics, in electronics issues I play by ear. 😀
Empirically, if such a thing exists in simulations, I found that capacitance of lower MOSFET does matters a little, and the IRF510 behaves better than the FQP2N60C, I don't know why.
I think that I got it, at EL34 cathodes (Drain of M3) the simulation shows an impedance of about 10mΩ, and the drain current of M3 is about 21nA RMS, then the AC voltage at EL34 cathodes is
Vac ≈ 2 x 10⁻¹⁰ V
The gain of the grounded grid stage is
G = (µ+1) Ra / (Ra + Rp) ≈ (µ+1)
Then, on the regulator output
Vac ≈ 2(µ+1) x 10⁻¹⁰ V
Whatever is µ, it is a very small number.
I hope the calculation is right
Vac ≈ 2 x 10⁻¹⁰ V
The gain of the grounded grid stage is
G = (µ+1) Ra / (Ra + Rp) ≈ (µ+1)
Then, on the regulator output
Vac ≈ 2(µ+1) x 10⁻¹⁰ V
Whatever is µ, it is a very small number.
I hope the calculation is right
- Status
- Not open for further replies.