By popular demand, and because....

[SoulFetish]

Never is a long time daz, check this out:

Interesting. Q8 & Q9 are darlingtons right? What is their purpose in the power supply? I know it's off topic, I was just curious.

[Enzo]

Nope, TIP 31 and TIP 32 are plain old bipolars.

regardless, single or darlington, they are pass transistors, the zener at the base of each sets a voltage, which the emitter follows. Et voila, voltage regulator.

[SoulFetish]

Nope, TIP 31 and TIP 32 are plain old bipolars.

regardless, single or darlington, they are pass transistors, the zener at the base of each sets a voltage, which the emitter follows. Et voila, voltage regulator.

Damn that's useful. I like it. Thanks!

[loudthud]

What I find interesting about the schematic Enzo posted is the dual N-channel output with a bootstrap (C6) so Q4 cab saturate.

[SoulFetish]

What I find interesting about the schematic Enzo posted is the dual N-channel output with a bootstrap (C6) so Q4 cab saturate.

Yeah that's weird. Does that cap have no effect on Q5?

[loudthud]

Yeah that's weird. Does that cap have no effect on Q5?

Q5 is not used (NU) in this particular amp. It was probably included on the schematic just for the PCB layout. Another amp might use Q4 and Q5 to achieve more output and use the same PCB. C6 pushes the supply for Q1 and Q3 above the V+ rail so that the extra 5V or so is available to turn Q4 fully on. Q1 and Q3 don't mind the little nudge they get form C6 because their Collector impedance is quite high. If the Gate Voltage of Q4 could only go to the rail, the positive output could only go up to about V+ minus the threshold of Q4 (when driving a resistive load). It's really just a trick to eliminate the need for a separate power supply. Q6 doesn't have this problem because it's Source is tied to the V- rail. Some MOSFET complementary source follower designs just throw away the extra output swing (and power) they could get and let the output clip softly.

[Enzo]

I hadn't thought about the circuit, I just remembered a lot of CArvin schematics used color. SO I grabbed one.

[Chuck H]

After two cathode-bypassed gain stages, I think blocking would indeed normally be an issue at the V2A stage.

The partial cathode-bypassing of stages V1A and V1B, and the low value coupling cap (0.0047uF) to V2A, all give some bass cut (I think - without doing any calculations) which is good when blocking might be an issue later in the circuit.

With the large 470k grid stopper on V2A, combined with the 220k to make up the grid leak, you get a good balance between the charging and discharging time constants for the 0.0047uF coupling cap – again a good measure against blocking.

If we take the output impedance of the V1B stage as (say) 40k, the charging time constant when the grid is conducting is approx.
( 40k + 100k + 470k||220k ) x 0.0047 e-6 = 1.36mS

And for discharging (when the grid is not conducting):
( 40k + 100K + 220k ) x 0.0047 e-6 = 1.69mS

It’s a nice design, in my opinion.

Without doing the math... You could actually omit the voltage divider. Just use a 58k load against the tubes internal resistance. Because of the lower impedance you'd need to bump the coupling cap to .017uf to maintain the same knee frequency. But this should still reduce the time constant even further. Then...

You could reduce the grid stop value. Which would add a little gain so you could then reduce the load resistance even further, again necessitating a change of the coupling cap value. And so on...

I'd bet that a simple .02uf coupling cap, a 47k load and a 10k grid stop would give better results with a small increase in gain.

[daz]

Without doing the math... You could actually omit the voltage divider. Just use a 58k load against the tubes internal resistance. Because of the lower impedance you'd need to bump the coupling cap to .017uf to maintain the same knee frequency. But this should still reduce the time constant even further. Then...

You could reduce the grid stop value. Which would add a little gain so you could then reduce the load resistance even further, again necessitating a change of the coupling cap value. And so on...

I'd bet that a simple .02uf coupling cap, a 47k load and a 10k grid stop would give better results with no reduction in gain.

Ya had to temp me didn't ya? I just went thru another fiasco after i posted the schematic. I was going to leave it with the final results that i really liked on the schematic but the parts were just tacked in so i had to redo them in permanent fashion. Can you guess what happened? Yep....spent about the last day and 1/2 going crazy trying to figure out why the tone went to $hit after i soldered them in permanently. Long story short it was the gain pot. And now you're tempting me with more ideas? What are ya, satan? Well, i probably will have to now or i will die of curiosity. Dammit Chuck !

[Chuck H]

Well... I like eloquence in design. Playing the tubes internal resistance against the load rather than employing a component resistor voltage divider omits one part AND reduces impedance (and therefor time constant) on a circuit that could use a lower time constant anyway. Eloquence. Fewer parts and more goal oriented

Sorry?

[daz]

Well... I like eloquence in design. Playing the tubes internal resistance against the load rather than employing a component resistor voltage divider omits one part AND reduces impedance (and therefor time constant) on a circuit that could use a lower time constant anyway. Eloquence. Fewer parts and more goal oriented

Sorry?

Just ribbin ya. I tried it but i'm not sure if i'm doing exact what u said. I replaced the 220k with a 47k and use a 10k grid stopper in place of the 470k. Tried a .022 and the .0047. But i seemed to get the beginning feel of blocking and the gain was reduced a bit more than i'd like.I know blocking+less gain doesn't make sense but thats just a description of what it sounded and felt like to me. So i tried a 22k in place of the 10k and raised the 47k to 100k. That worsened the tone. In some ways it sounded good with the 10k and 47k but the way it is on the schematic sounds better and considerably different. But i think if i played with your values and tried to make up for things i don't like about it elsewhere it might end up being worthwhile. But then i'm back to redesigning the whole thing again so not sure i wanna go there. But it did seem like it had potential. The one thing i did like is it was a bit more percussive when i roll back the guitar volume.

[daz]

I've been screwing with it since i last posted. Yeah....i'm hopeless. It's like a giant jigsaw puzzled and i can't stop till theres nothing more to gain. Anyways, i have been playing around with your idea and i now think i may like it. Gotta wait till tomorrow because it's late but i think i like the articulation more than the schematic, and thats real important because it has a huge effect on my playing. Just may need to change some other things to accommodate the different sound like treble peaker and gain pot bleed values. We shall see...

[Chuck H]

Did you remove the 100k series resistor?

[Chuck H]

I've been screwing with it since i last posted. Yeah....i'm hopeless. It's like a giant jigsaw puzzled and i can't stop till theres nothing more to gain. Anyways, i have been playing around with your idea and i now think i may like it. Gotta wait till tomorrow because it's late but i think i like the articulation more than the schematic, and thats real important because it has a huge effect on my playing. Just may need to change some other things to accommodate the different sound like treble peaker and gain pot bleed values. We shall see...

The lower impedance accounts for the greater articulation. The higher the impedance, the greater the effect of impedance rising with frequency. With the lower impedance circuit there will be more balance across the audible range. That is, less loss at higher frequencies. I expected this affect and wondered if you might like it or not. Sometimes, Especially WRT guitar amps, it's actually good to have a circuit induced HF roll off. It avoids the un-aesthetic application of top end "bleeder" type circuits. So, the circuit I posted doesn't have as much roll off. For better or worse.

[daz]

The lower impedance accounts for the greater articulation. The higher the impedance, the greater the effect of impedance rising with frequency. With the lower impedance circuit there will be more balance across the audible range. That is, less loss at higher frequencies. I expected this affect and wondered if you might like it or not. Sometimes, Especially WRT guitar amps, it's actually good to have a circuit induced HF roll off. It avoids the un-aesthetic application of top end "bleeder" type circuits. So, the circuit I posted doesn't have as much roll off. For better or worse.

I did jump the 100k and A/B'dit both ways and i'm undecided on that but it seems to sound more smooth and refined with it but i'll be A/B'ing it further. Top end wise i can't really say it rolls of less, in fact it seems like maybe more but in a good way that can be compensated for via EQ. By the way, I DID use values close to these at times on my other amp but then it had more gain for some reason even with the same values. I never could figure that out.