Building parafeed EL84 DC 2A3 amp
- Thread starter Bas Horneman
- Start date
Around 400 dollar per watt. Assuming 3 watt power.
Caught an error. a 75uF too much. Not that it matters.
Things I have learned/deduced during the build: a 2a3 is very suitable for parafeed:
If you return the the lower connection of the OPT to the cathode you minimize the audio frequency current loop and eliminate the cathode bias cap and the power supply cap from the audio loop.
3.3uF at 20Hz has a capacitive reactance of about 2500 Ohms (2411 Ohms).
If the output transformer primary is 2500 Ohms, the combination of the output transformer primary and 3.3uF cap will be -3dB at 20Hz
And, that is also -1dB at 40Hz.
Suppose the output transformer primary was 5000 Ohms, and the cap is 3.3 uF.
Then, it will be -3dB at 10Hz, and -1dB at 20Hz.
See how that works?
That is why somebody picked 30uF as the coupling cap, plenty big enough (even a bit too much, won't hurt), and probably he had a 30uF cap in his capacitor bin.
If the output transformer primary is 2500 Ohms, the combination of the output transformer primary and 3.3uF cap will be -3dB at 20Hz
And, that is also -1dB at 40Hz.
Suppose the output transformer primary was 5000 Ohms, and the cap is 3.3 uF.
Then, it will be -3dB at 10Hz, and -1dB at 20Hz.
See how that works?
That is why somebody picked 30uF as the coupling cap, plenty big enough (even a bit too much, won't hurt), and probably he had a 30uF cap in his capacitor bin.
Thank you!
This is the post by Paul Joppa that convinced me to go for it (15 years ago):
"Posted by Paul Joppa (P) on February 6, 2009 at 18:41:15
In Reply to: Last call: FS-100 plate chokes posted by mqracing on February 6, 2009 at 08:18:55:
Just a thought for anyone interested ...
In my experience, a larger impedance plate load always makes for better sound in parafeed. I've tried, or at least listened to, a wide variety of chokes as well as current sources. I once built a preamp with switched plate loads of 20, 24, 40, or 48 henries, for example (using 6BL7s parafeed to B7-5K) plus a current source - believe me, every inductance increase was an improvement.
Extrapolating that experience, here's a really wild idea - get a pair of these gapped for 60mA and couple to a TFA-2004Ni with a 27uF capacitor. If you can afford the nickel core TFA (still the best sounding OPT I've ever heard), this choke is neither inappropriate nor unreasonably expensive. Feed it with a 2A3 running 275v at 55mA - cathode resistor 910 ohms - or if you're feeling flush, one of those 2.5v 300Bs at 300v, 60mA (same cathode resistor).
Of course the greatest benefit will be in the bass, so use this to drive some ginormous efficient speakers with killer bass.
I predict you will never hear sound as good with a 2A3."
In lieu of the the 27uF. I used a 30uF Continental polyprop in oil that had a good reputation (anecdotal) That I had bought years ago.
Yes. Will try it soon. I kept it off because John Broskie ridiculed it. But will try for myself. Probably this weekend.
Bas Horneman,
Thanks for the post.
I am sure you caught the "Yea, Sure" negative affirmation.
Regarding the electrolytic bypass cap that is across the self bias resistor in the schematic Post # 128 . . .
"Many argue that this arrangement is the purer path, as it no longer places the icky bypass capacitor in the signal path."
I caught the comment: "Yea, Sure". It means the statement is Un-True.
If it is "not in the signal path", that means it does not have any signal current. Really?
Start with the cap in the circuit, apply a 1kHz test tone at the input, and a load resistor across the output transformer secondary, and measure the
voltage across the load resistor.
Then remove the bypass cap, and re-measure the voltage across the load resistor.
If there is any difference in the two measured voltages, you need the bypass cap. Well it Does make a big difference.
The self bias resistor "takes" the DC quiescent current.
The bypass capacitor "takes" the signal current.
When you remove the bypass cap, the self bias resistor "takes" Both the quiescent DC current, and the signal current.
The plate impedance, rp, of a 2A3 is 800 Ohms. The 2A3 u = 4.2
With an un-bypassed 910 Ohm self bias resistor, the plate impedance is Increased by 910 Ohms x u = 3822 Ohms (Rk x u).
With that, the gain will go down, and the damping factor will go down.
Just my opinion, please prove me wrong, measure with, then without, the bypass cap.
Thanks for the post.
I am sure you caught the "Yea, Sure" negative affirmation.
Regarding the electrolytic bypass cap that is across the self bias resistor in the schematic Post # 128 . . .
"Many argue that this arrangement is the purer path, as it no longer places the icky bypass capacitor in the signal path."
I caught the comment: "Yea, Sure". It means the statement is Un-True.
If it is "not in the signal path", that means it does not have any signal current. Really?
Start with the cap in the circuit, apply a 1kHz test tone at the input, and a load resistor across the output transformer secondary, and measure the
voltage across the load resistor.
Then remove the bypass cap, and re-measure the voltage across the load resistor.
If there is any difference in the two measured voltages, you need the bypass cap. Well it Does make a big difference.
The self bias resistor "takes" the DC quiescent current.
The bypass capacitor "takes" the signal current.
When you remove the bypass cap, the self bias resistor "takes" Both the quiescent DC current, and the signal current.
The plate impedance, rp, of a 2A3 is 800 Ohms. The 2A3 u = 4.2
With an un-bypassed 910 Ohm self bias resistor, the plate impedance is Increased by 910 Ohms x u = 3822 Ohms (Rk x u).
With that, the gain will go down, and the damping factor will go down.
Just my opinion, please prove me wrong, measure with, then without, the bypass cap.
Last edited:
I'm confused though. My post in 128 is not to say the bypass cap should be removed. But that Broskie vaguely seems to say that there is no benefit to connecting the ground of the output transformer to the cathode instead of ground.
I have one of my 2A3 parafeed amps wired to the cathode, the other to ground. Didn’t hear any real difference when I made the change.
Where I hear a real difference is with the driver….the Stereomour has a 3B7 driver. It is the one I listen to…more detail and deeper sound stage.
I will have to play with the parafeed cap…move up slowly towards 30uf and see how it changes.
I have a TFA 2004 JR that I plan to pair with a 10Y driver.
Where I hear a real difference is with the driver….the Stereomour has a 3B7 driver. It is the one I listen to…more detail and deeper sound stage.
I will have to play with the parafeed cap…move up slowly towards 30uf and see how it changes.
I have a TFA 2004 JR that I plan to pair with a 10Y driver.
Oooh a direct heated tube, nice.
Somehow my 801a's that I have are sensitive like hell...I had to remove power transformer from the chassis. Kinda soured me to the 801a. Perhaps I should try it again now that I have a bit more experience.
How do you like the BeePre? I am thinking about building a 300b preamp.
Aluminium.