This is an amp that I am currently bread boarding, it is a project suggestion on the Electra Print web site for a SE 6BX7 amp....
My question concerns the 450uf el cap that connects from B+ to the cathode of the 6BX7. Is it possible that this is here by design to inject some of the power supply ripple into the cathode, so that it can cancel out with the ripple on the plate, for better PSRR?
John Broskie does something similar in this article:
But in the Broskie article the value of the capacitor is a carefully chosen ratio of mu, at 1/mu (+ 1 I think). But on the Electra Print schematic they just seem to have jammed in a big capacitor
Parafeed-Amplifier Design
John Broskie does it again in this recent 300B amp:
Two-Tube Single-Ended Amplifier & Speakers (halfway down the article)
All for PSRR benefit?
Last edited:
This is called an "ultrapath" capacitor. The idea is that the signal current circulates in the loop:
tube plate -> primary -> ultrapath capacitor -> tube cathode.
Without the capacitor, the signal current goes in the more complex loop:
tube plate -> primary -> power supply capacitor -> ground common -> cathode resistor -> tube cathode.
tube plate -> primary -> ultrapath capacitor -> tube cathode.
Without the capacitor, the signal current goes in the more complex loop:
tube plate -> primary -> power supply capacitor -> ground common -> cathode resistor -> tube cathode.
Without that capacitor there, you would have to connect a bypass cap across the 180 Ohm resistor.
Otherwise, much of the signal voltage on the output tube grids would appear on the 180 Ohm resistor, and less signal voltage would appear on the output tube plates.
That would result in 3 things . . .
lower the gain,
lower the power out,
and lower damping factor.
If you really want to have a better balance of the two 6BX7 triode currents, then replace the 180 Ohm cathode resistor with two 360 Ohm resistors, one for each cathode to ground.
Then you need either two capacitors, one from B+ to one cathode, and the other from B+ to the other cathode;
Or, one cap across each 360 Ohm self bias resistor.
The individual self bias resistors will tend to better balance the cathode currents, and you can easily see if one triode section is way to strong, or way to weak (when new, and as time goes on).
If you want to check the current balance in the two 6J6 triodes, then carefully measure the voltage across each plate resistor. If the balance seems badly mis-balanced, then check the wiring, and then you may try another 6J6.
Otherwise, much of the signal voltage on the output tube grids would appear on the 180 Ohm resistor, and less signal voltage would appear on the output tube plates.
That would result in 3 things . . .
lower the gain,
lower the power out,
and lower damping factor.
If you really want to have a better balance of the two 6BX7 triode currents, then replace the 180 Ohm cathode resistor with two 360 Ohm resistors, one for each cathode to ground.
Then you need either two capacitors, one from B+ to one cathode, and the other from B+ to the other cathode;
Or, one cap across each 360 Ohm self bias resistor.
The individual self bias resistors will tend to better balance the cathode currents, and you can easily see if one triode section is way to strong, or way to weak (when new, and as time goes on).
If you want to check the current balance in the two 6J6 triodes, then carefully measure the voltage across each plate resistor. If the balance seems badly mis-balanced, then check the wiring, and then you may try another 6J6.
Last edited:
So On the Electraprint schematic shown the cathodes are in parallel, no cathode bypass to ground, and ultra path capacitor? That's all it takes? Is the cap value critical in this example? In John Broskies 300B amp he also has a cathode bypass to ground and the ultra path cap value is determined as 1/mu+1 fraction of whatever value the cathode to ground cap is, correct? So if the mu is 4 and the ground bypass cap is 500uf then the ultra path cap should be 100uf, 1/mu+1 of 500? I'm trying to understand if when not using a ground cathode bypass, like the electraprint amp, then the ultra path cap value isn't very critical? But if you use both cathode caps then the ultra path one should be that mu fraction of the grounded cap?
Also if I used two separate cathode resistors on the electraprint amp of twice the ohms, I assume I would have to ultra path each separately? Probably better to just use one cathode resistor when paralleling a tube where you also want to use ultra path too?
Also if I used two separate cathode resistors on the electraprint amp of twice the ohms, I assume I would have to ultra path each separately? Probably better to just use one cathode resistor when paralleling a tube where you also want to use ultra path too?
Last edited:
You need to bypass both the parallel impedance of the two cathodes, which is also in parallel with the 'impedance' of the 180 Ohm resistor.
Suppose the capacitance is too low . . .
Then an example of what you may get is frequency response that is -1 dB at 80 Hz, and -3dB at 40 Hz.
If you do not individually bias the parallel 6BX7 triodes, you may get badly unbalanced currents.
Checking the balance of two triodes on a tube checker is not good, unless you test it at the exact voltages that the amplifier presents to the triodes.
Your amplifier, when it is re-designed to give individual self bias to the two triodes, is the very best way to check the balance of the two triodes.
Some triodes with a common cathode balance very well. Others balance very badly. Some fall between the two limits.
The 6BX7 triodes were designed for TV sets, but in different voltages and currents, one as a vertical oscillator, and one as a vertical amplifier. The two triodes are built the same, but not tested for real close balance.
Your mileage may vary.
Suppose the capacitance is too low . . .
Then an example of what you may get is frequency response that is -1 dB at 80 Hz, and -3dB at 40 Hz.
If you do not individually bias the parallel 6BX7 triodes, you may get badly unbalanced currents.
Checking the balance of two triodes on a tube checker is not good, unless you test it at the exact voltages that the amplifier presents to the triodes.
Your amplifier, when it is re-designed to give individual self bias to the two triodes, is the very best way to check the balance of the two triodes.
Some triodes with a common cathode balance very well. Others balance very badly. Some fall between the two limits.
The 6BX7 triodes were designed for TV sets, but in different voltages and currents, one as a vertical oscillator, and one as a vertical amplifier. The two triodes are built the same, but not tested for real close balance.
Your mileage may vary.
Last edited:
Thanks guys, this is getting clearer. I've been breadboarding this without that ultra path cap because I didn't know what it was for. I split the cathodes to two 360R bypassed to ground caps. Later I'll try it with two ultra path caps and no ground bypass instead. And I'll also try it with both ground caps and ultra path caps using that mu fraction Broskie explained to ratio the ultra path caps. And what the hell I'll also try it like the schematic.
Trying all those combinations can be a nice learning experience.
But ultimately, the low frequency response of that output stage is dependent on three things:
1. The parallel plate impedances, rp, of the triodes (which is dependent on the plate currents).
2. The inductance of the output transformer (and enough laminations to not saturate).
3. Properly bypassing: The impedance of the two cathodes that ultimately are in parallel, and that is in parallel with either the impedance of the 180 Ohm resistor, or the the impedance of the two 360 Ohm resistors that are in parallel.
There is no getting around that, unless you apply negative feedback from the output transformer secondary back to the input tube (please do not do that).
And if you do apply negative feedback, pay attention to the output transformer, it has to have enough laminations to not saturate, negative feedback will make it worse.
And even if there are enough laminations, I still say, do not apply negative feedback to that amplifier. It uses triodes for all the stages, and it should have reasonable distortion and reasonable damping factor; . . . and reasonable frequency response and reasonable square wave response . . . if the output transformer is relatively good.
Happy Listening!
But ultimately, the low frequency response of that output stage is dependent on three things:
1. The parallel plate impedances, rp, of the triodes (which is dependent on the plate currents).
2. The inductance of the output transformer (and enough laminations to not saturate).
3. Properly bypassing: The impedance of the two cathodes that ultimately are in parallel, and that is in parallel with either the impedance of the 180 Ohm resistor, or the the impedance of the two 360 Ohm resistors that are in parallel.
There is no getting around that, unless you apply negative feedback from the output transformer secondary back to the input tube (please do not do that).
And if you do apply negative feedback, pay attention to the output transformer, it has to have enough laminations to not saturate, negative feedback will make it worse.
And even if there are enough laminations, I still say, do not apply negative feedback to that amplifier. It uses triodes for all the stages, and it should have reasonable distortion and reasonable damping factor; . . . and reasonable frequency response and reasonable square wave response . . . if the output transformer is relatively good.
Happy Listening!
Last edited:
Thanks 6A3sUMMER I wasn't going to start learning NFB using this amp. I'm using one electron UBT1 transformers 1.6k 15W I think they have around 150ma DC limit.