[Goobley]

Well the torpedo runs its LEDs at about 0.5mA (low efficiency types like yours). If you're happy with voltage you get In practice using these LEDs at low current thn that's fine, just don't necessarily expect it to match what your theory suggests... When it comes down to it it's reality that counts

 

[KimLaroux]

Quote:

 

 

 
I just finished building this circuit. The only difference is I biased the right-hand tube using two diodes powered directly from B+.
 
Interestingly enough, it worked first run.
 
I noticed two problems though. The first is the gain of the amp is quite low. So low in fact I can turn the volume knob all the way up and it still doesn't hurt my ears... The second is this design rejects no PSU noise, at all. Bypassing the plate resistor using a cap cuts the audio, but not the noise. I'm guessing this means the noise comes directly from the supply and is being injected directly to the output without any attenuation.
 
Any ideas?

 

[holland]

Quote:

 
12AU7. Can you explain to me how different value plate resistors would improve or degrade the operation of the circuit? I'm just curious, really. The only thing I can think of is higher value would improve PSRR, true?

 
 
The plate resistor or anode resistor is allow voltage swing and customize gain.  At a minimum you should pick 2*Rp where Rp is about 7700 for the 12AU7 which is a good point to maximize gain and voltage swing.  15K is a good start, though I would look at shifting the operating point to 80V on the plate.  Broskie had it at 1/2 B+.  For you, that would mean a grid of about -2.2V, or a bias of 5V.  You don't have to do that, but playing with it may be fine.  You'll probably want some deterministic tests to go with it.  Lower noise, etc., through measuring equipment.
 
The divider circuit does 2 things, when from tube output.  Bias *and* negative feedback to reduce gain and linearize the local stage.  Biasing from that point is advantageous, as opposed to B+.  Do not bias from B+.  You will inject PSU noise directly into the grid.  Bad, bad, bad.  If you do use B+, you will need a transistor in there, IMO, for more isolation.
 
For a hybrid you want to reduce gain anyway, feedback helps a bunch there.  Local or global, up to you.  Even the SSMH had feedback, though non-tube people and newbies would never know it.
 
Anyhow.  I would play with this after you get the circuit operational and functioning.  Since I see you started already, on LED biasing, does it work?  Do you get sound, are the DC operating points where you want?   What ever you do, bias from the 106V.  I would not worry too much about plate voltage.  It doesn't need to be precise, just close enough.  Tubes and load lines are not 100% accurate anyhow.  For 80V plate, -2.2V grid is a bit low, but that is about where the SOHA 2 ran (48V plate, 2V cathode @ 1mA, just about).  I don't think you'll clip, but am not sure.  Any how, for a quick test, you can just use a 5V zener.  Yes, it's noiser, but for testing it's quick and dirty.
 
On a side note, I would probably feed the voltage through a 300 ohm grid stopper resistor as well.  Minor tweaks that you can do after you get up and going.  The grid stopper is for stability, not really needed, but I would do it anyway based on common practices.

 

[holland]

Quote:

 
I just finished building this circuit. The only difference is I biased the right-hand tube using two diodes powered directly from B+.
 
Interestingly enough, it worked first run.
 
I noticed two problems though. The first is the gain of the amp is quite low. So low in fact I can turn the volume knob all the way up and it still doesn't hurt my ears... The second is this design rejects no PSU noise, at all. Bypassing the plate resistor using a cap cuts the audio, but not the noise. I'm guessing this means the noise comes directly from the supply and is being injected directly to the output without any attenuation.
 
Any ideas?

 
Do not use B+, at all.
 
You injected the PSU noise via your use of B+ for the LED.  Bad, bad, bad.  Don't do it..  I'm not sure about the gain, but you can bump up the plate load, I would look at 80V plate voltage instead of 106V.
 
What is your SS section of your hybrid?
 
Edit:  honestly, stick with resistors for now.  Get a baseline, and then add the LED bias, but do it from the plate output.  You want to make sure you get the values correct, because it's negative feedback.
 
Edit2: add a grid stopper of 300 ohm too, on the rightmost triode.  It's common practice to prevent oscillation, I would just do it, it doesn't cost much.

 

[wakibaki]

Quote:

Congratulations for solving a problem... 

 
Thanks.
 
Try to be a bit less hostile. Just because we disagreed in the past, that's not good reason to try to perpetuate the disagreement, especially on such a meagre basis.
 
w

 
@holland
 
If you read a bit further in Broskie, I'm pretty sure I remember where he discusses transformer coupled amplifiers with CCSs.
 
@KimLaroux
 
Oh well, looks like no-one is prepared to simulate your circuit for you...

 

[holland]

Kim, took a closer look at the circuit.  Without feedback and perfect bias through something like a battery, you should be getting a voltage gain of around 6.x.  Any feedback will diminish that.
 
What is your SS stage, you may have negative gain there?

 

[holland]

Quote:

 
@holland
 
If you read a bit further in Broskie, I'm pretty sure I remember where he discusses transformer coupled amplifiers with CCSs.
 

 
Will do.  Until lately, I had 0 interest in transformers.  I'm designing a parafeed amp now, and will definitely be doing research.

 

[wakibaki]

Quote:

 
I just finished building this circuit. The only difference is I biased the right-hand tube using two diodes powered directly from B+.

Interestingly enough, it worked first run.
 
I noticed two problems though. The first is the gain of the amp is quite low. So low in fact I can turn the volume knob all the way up and it still doesn't hurt my ears... The second is this design rejects no PSU noise, at all. Bypassing the plate resistor using a cap cuts the audio, but not the noise. I'm guessing this means the noise comes directly from the supply and is being injected directly to the output without any attenuation.
 
Any ideas?

 
If you read the Broskie link posted by holland you will see that the PSRR of the stage is practically non-existent. Biassing the CCA stage from the B+ throws away the feature that gives the Broskie circuit it's intrinsic beauty, which is that with correctly chosen (feedback) bias resistors connected to the anode, the design is self-biassing for a wide range of voltages. It's easy to arrange, the upper resistor should be mu-1 times the lower.
 
You tinker with Broskie's circuits at your peril, they are subtle and ingenious. Better to look at them for a long time and read his notes carefully, because there is a lot to learn from them.
 
w