It is worth mentioning here that dielectric thickness is a factor: high-voltage polyester caps are OK for audio, provided the audio signal is always a small fraction of the capacitor voltage rating. This may not be true for a typical parafeed cap, I admit, although in a valve amp you don't often need to worry about polyester distortion anyway because it is grossly swamped by all the other sources of distortion in a valve amp!
Parafeed circuits add an extra layer of engineering that is required versus traditional Air Gapped output transformer single ended circuits.
First the same as the traditional issues of Air Gapped output transformers:
The Antek 230:5 has a primary impedance transformation of (230/5)Squared = 2,116.
At mid frequencies:
8 Ohms on the 5V secondary will appear as 16.9k Ohms on the primary
4 Ohms on the 5V secondary will appear as 8.45k Ohms on the primary
At low frequencies, the primary will be inductive, and that inductive reactance is in parallel with that 16.9k or 8.45k Ohms that is reflected from the 8 or 4 Ohm load. So far, so good.
Next, the unique Parafeed factor:
What is the inductance of the Antek primary?
Suppose that primary is 10 Henries. Series Resonance of the primary inductance and the 1 uF cap is:
F = 1/(Root(L * C)) * 2 * pi = 50Hz
4uF will put that same 10 Henry primary at a series resonance of 25Hz.
You might try one of those "motor start" or similar capacitors of 4uF @ 660VAC. The modern ones are oil filled but with modern plates and dielectrics (metalized poly?). That should not 'over-voltage' even if the signal is at resonance, and even if the loudspeaker load comes off (but the tube or the output transformer might 'over-voltage'). If the primary is 40 Henries, and there is a 1uF capacitor, the series resonance will be at 25Hz.
First the same as the traditional issues of Air Gapped output transformers:
The Antek 230:5 has a primary impedance transformation of (230/5)Squared = 2,116.
At mid frequencies:
8 Ohms on the 5V secondary will appear as 16.9k Ohms on the primary
4 Ohms on the 5V secondary will appear as 8.45k Ohms on the primary
At low frequencies, the primary will be inductive, and that inductive reactance is in parallel with that 16.9k or 8.45k Ohms that is reflected from the 8 or 4 Ohm load. So far, so good.
Next, the unique Parafeed factor:
What is the inductance of the Antek primary?
Suppose that primary is 10 Henries. Series Resonance of the primary inductance and the 1 uF cap is:
F = 1/(Root(L * C)) * 2 * pi = 50Hz
4uF will put that same 10 Henry primary at a series resonance of 25Hz.
You might try one of those "motor start" or similar capacitors of 4uF @ 660VAC. The modern ones are oil filled but with modern plates and dielectrics (metalized poly?). That should not 'over-voltage' even if the signal is at resonance, and even if the loudspeaker load comes off (but the tube or the output transformer might 'over-voltage'). If the primary is 40 Henries, and there is a 1uF capacitor, the series resonance will be at 25Hz.
Last edited:
The speakers only go down to about 100hz if I remember correctly. My theory is, why feed the speaker more bandwidth than it can reproduce? I tried a 4.7uf oiler on this amp before but didn't notice much of an increase in bass response (if any), but I also had one of the speakers out of phase. I'll try it again, at least for the entertainment of the forum.
Last edited:
Nectarpuke,
How about posting a schematic of the amplifier, please.
Suppose there is a bypass cap on the input stage that is -3dB at 100 Hz, and an RC network that is -3dB at 100Hz, and the rp of the output tube plate is equal to the inductance of the output transformer at 100Hz. Now the amplifier is More than -6dB at 100Hz (and a few dB down at 200 or even 400Hz).
Even if the amplifier is -3dB at 100Hz due to a single dominant pole, it is -1 dB at 200Hz.
How about posting a schematic of the amplifier, please.
Suppose there is a bypass cap on the input stage that is -3dB at 100 Hz, and an RC network that is -3dB at 100Hz, and the rp of the output tube plate is equal to the inductance of the output transformer at 100Hz. Now the amplifier is More than -6dB at 100Hz (and a few dB down at 200 or even 400Hz).
Even if the amplifier is -3dB at 100Hz due to a single dominant pole, it is -1 dB at 200Hz.
I've tried posting a schematic, but the file size is too large. I haven't really put any effort into finding out how to shrink it down to size. The output stage is a 6N6P battery biased differential stage (Vgk -6 volts roughly, Vak 180 volts, Ik 22mA). The plates are loaded with a cascode ccs (10m45s on top of a dn2540 with a trim pot set to about 250 ohms... output is taken off the lower devices source) I did try the 4.7uf coupling capacitor last night after reversing the primaries on one side to get the phase correct. Major improvement in bass response that was not there before the phase reversal. (I double checked to make sure I wasn't losing my mind by swapping the secondaries... the low end disappeared).
The driver stage is a 6sn7 phase splitter with an lm317 on the tail set for 16mA. The plates are gyrator loaded using the circuit found on Bartola Valves' website. Plate voltage is 230 volts if I remember correctly. The voltage was adjusted for equal current draw on both triodes. This stage is capacitively coupled to the output with .22uf Russian teflons. I am also using Silk grid chokes on the 6N6P.
The driver stage is a 6sn7 phase splitter with an lm317 on the tail set for 16mA. The plates are gyrator loaded using the circuit found on Bartola Valves' website. Plate voltage is 230 volts if I remember correctly. The voltage was adjusted for equal current draw on both triodes. This stage is capacitively coupled to the output with .22uf Russian teflons. I am also using Silk grid chokes on the 6N6P.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.