Finding the proper load for a pentode push-pull output stage

I think the metal can might get rid of heat a little more efficiently, probably why it can take a little more abuse. The near-identical power output is also interesting.
 
I think the metal can might get rid of heat a little more efficiently, probably why it can take a little more abuse. The near-identical power output is also interesting.

Just another reason to pick up those metal can 6L6s, I guess :)
 
I can't imagine the procedure is fundamentally any different, but you just have to consider the left-ward shift in the operating point as the current increases. For example, the RCA 6L6GB data sheet does show a difference like you say... 6.6K primary for 360V/270V fixed bias vs 9K for 360V/270V cathode bias. I think this is to account for the pivoting of the class B load line as the cathode current increases (and thus the bias voltage gets dragged more negative), although I'd defer to @dcgillespie on that one...
That's correct, there is no difference in how the load lines are drawn for either self-bias or fixed bias, just remember that - with self-bias, all the voltages are referenced to the cathode instead of the 0V. Here is a quick refresher on the output stage bias.
 
Cathode bias effectively raises the impedance of the tubes, and thus requires a higher plate-to-plate load impedance to account for the increase. Under dynamic conditions, the value of the cathode resistance is quite small relative to the plate impedance, and so has little impact in that regard -- as previously mentioned. The real bug-a-boo with cathode bias and Class AB1 operation is the effect that the average increase in current draw during Class B conditions has on the effective screen grid voltage. Any increase in average current draw across the cathode bias resistor represents a decrease in screen grid voltage, which then raises the internal impedance of the tubes, and ultimately shifts the Class B load line from that which would otherwise be similar to fixed bias operation in that class with regards to loading conditions.

Dave
 
Even more food for thought... In most cases, the operating conditions from the datasheet show that the output power with self-bias is lower than for fixed bias - this is mainly due to the voltage drop across the cathode resistor, i.e., there is less voltage across the tube. As Dave briefly touched upon, self-bias doesn't handle being driven hard very well, this means that for a given tube and B+, you always get higher power with lower distortion using fixed bias. In addition, self-bias output stage also has issue with trasient recovery compared with the fixed bias, BUT, there is always a but... the reason that you see self-bias in many amplifiers is due to the fact, it adapts to tube aging and requires no adjustment unlike the fixed bias case.

Anyway, whether the output stage is self-bias or fixed-bias, there is always an ideal load impedance that results in the highest power for a certain distortion level - if the load resistance is lowered, the distortion rapidly increases; or if the load resistance is increased, the maximum power decreases, but at least the distortion also decreases. For much more on this, pleaser refer to RDH4.
 
Okay, given the method described on this thread (composite load line class A transitioning to class B), I've been able to do a bit of analysis and I think I have a solution that will work and meet the requirements I've got. So that leads me to the following:
  • 8K plate to plate primary
  • 30 Watts output (per channel, both channels driven)
  • 430VDC to the center taps (this provided from an Edcor XPWR188 power transformer with silicon bridge rectification)
  • EL34 output tubes in push-pull
  • 220VDC on the screens
  • EFB regulated output stage
Power output, on paper, was actually more towards 40 watts with 250 VDC screens, so I needed to reduce the screen voltage down to 220VDC to shrink max power output to something that can be handled by the off-the-shelf new stock output transformers. I intend to use Transcendar 30W models, so I will finally get to try these out for mee self.

I was thinking 8K primary was too high for EL34s, but it actually it works out okay (on paper), and gives max current demand of the output stages (both channels driven) of 370 mA (185mA per channel), which can easily be sourced by the Edcor XPWR188, which supports 500 mA on the HV winding. With this setup, The first 6 watts or so are (supposedly) class A push-pull, while the remaining wattage is delivered through class B operation.

I could easily squeak out more power, up to maybe 45 watts/channel with the same power transformer if I used 4K output transformers, but my son who requested this amp, says it must have a smaaaaall footprint (I'm thinking 16 x 12 inches), so the smaller the transformers are, the better, while still delivering a respectable 30 watts output.
 
I could easily squeak out more power, up to maybe 45 watts/channel with the same power transformer if I used 4K output transformers, but my son who requested this amp, says it must have a smaaaaall footprint (I'm thinking 16 x 12 inches), so the smaller the transformers are, the better, while still delivering a respectable 30 watts output.
30W from a pair of EL34's is quite low, but I guess if he does not need the power and wants a small footprint then it will have to do... or, you can try a pair of EL84's and push them a bit like Modjeski did with his Music Reference amp.
 
Thats really not *that* low. The ST70 only made 35 from a pair, the Pilot SA-260 made 30 from a pair. I've never actually bench tested mine, but I have no special reason to believe it makes more than it's claimed power.
 
Thats really not *that* low. The ST70 only made 35 from a pair, the Pilot SA-260 made 30 from a pair. I've never actually bench tested mine, but I have no special reason to believe it makes more than it's claimed power.
Take a look at the datasheets, the typical output power is 40W ~ 50+ Watts, so 30 Watts is basically just cruising.:)
 
Perfect. Who likes to buy tubes more than twice a century?

The datasheets also don't factor in transformer loss, so that 40w is probably closer to 30-35 at the speakers.
 
Plus I already have the EL34 tubes. That's what drove the whole thing--tubes were already purchased and available. Funny things drive design requirements sometimes. Kind of bass-ackwards probably.
 
Funny you should ask. I'm drawing up the schematic right now in fact.

I've decided to use Triode Electronics A470 outputs, with 355V to the center taps. Should deliver 35 watts/channel at the speaker terminals. It has power output like a Dynaco ST-70, but topology resembles an Eico HF-89, which was my inspiration for it.

Input sensitivity should be 0.7V to deliver that 35 watts out. I'm trying to decide what tube to use for the frontend. Either 6SN7 or 12AU7. Can't decide.

See attached for design in process.
 

Attachments

  • EL34_Mk2.pdf
    128.2 KB · Views: 26
Hey Kevin,
What is that AC Balance trimmer (20K) on the PI? I am not sure I understand how it works. Or is it drawn incorrectly? Would appreciate the explanation.
Cheers, Paul.
 
I think that AC balance pot is to dial in the gain for the cathode-driven half of the LTP phase inverter (they're not perfectly balanced unless you use a CCS in the tail, so you compensate by having a slightly higher load resister on the cathode-driven side).
 
The trimmer pot in conjunction with its companion 15K fixed resistor adjust the plate load and hence the gain of the second triode. When the 20K trimmer is full open, it increases the plate load by about 8.5K. When at zero it of course adds no additional plate load. It's in parallel with the 15K resistor to limit current through the trimmer. The values I chose (20K trimmer and 15K fixed resistor) are somewhat arbitrary. Really just needed about 20% adjust so I can dial in AC balance, while keeping heat dissipation down in the trimmer itself.

The nice thing about this kind of manual adjust is that if there are any imbalances down stream such as unmatched gains on the power tubes or slightly unmatched windings on the OPT primary, those can be corrected. These would not be able to be corrected with just a CCS in the tail of the splitter.

I'm not a rocket scientist on this subject but it seems to me that what really matters in "balancing" your output stage is to get the signals perfectly balanced at the OPT primary where they are combined back into one signal that's fed to the speaker. To support that, it could mean that the signals coming out of the splitter need to be slightly imbalanced.
 
Last edited:
The only amp I own with an AC balance pot also needs to be ever so slightly off center to come out to minimal distortion. Initially I balanced it using the scope to get equal but opposite output on the inverter, but with the THD meter hooked up it wanted to be a bit off from that point.
 
Back
Top Bottom