Here's a schematic to illustrate the topology.
I use IRF820 for the MOSFETs; R2 and R5 are set so that they drop about 10 volts.
Apply a voltage reference from B+ to junction of R4 and R7 - this determines plate voltage for tubes.
R1 determines current thru both tubes; if not substantial, I use a trimmer pot here.
C3, the parafeed cap, is paramount to sound quality. As I said, I recommend soviet caps for best quality. 2µF is ok for 10k OT primary impedance, if you need more power and a lower primary impedance, you'll need more capacitance. I recommend using the smallest cap you can get away with, for quality reasons. Smaller caps tend to be much better quality.
I do the PI with a centertapped transformer for simple construction. It might be better to put a same kind of stage in front of this; I haven't tried it yet.
This topology provides surprisingly large power.
I use IRF820 for the MOSFETs; R2 and R5 are set so that they drop about 10 volts.
Apply a voltage reference from B+ to junction of R4 and R7 - this determines plate voltage for tubes.
R1 determines current thru both tubes; if not substantial, I use a trimmer pot here.
C3, the parafeed cap, is paramount to sound quality. As I said, I recommend soviet caps for best quality. 2µF is ok for 10k OT primary impedance, if you need more power and a lower primary impedance, you'll need more capacitance. I recommend using the smallest cap you can get away with, for quality reasons. Smaller caps tend to be much better quality.
I do the PI with a centertapped transformer for simple construction. It might be better to put a same kind of stage in front of this; I haven't tried it yet.
This topology provides surprisingly large power.
Attachments
I'm kicking around a 6C4C (2A3) with a 5843 (or 6c45p) driver. Parafeed looks out of the question as headphone secondary OPT's don't exist.
The big question is Electra-Print the only source for a custom 2.5k/40 OPT ?
I really want a quality OPT that takes advantage of the low winding ratio. To me with an SET the tubes are really secondary, it all starts with OPT then the power supply.
The big question is Electra-Print the only source for a custom 2.5k/40 OPT ?
I really want a quality OPT that takes advantage of the low winding ratio. To me with an SET the tubes are really secondary, it all starts with OPT then the power supply.
Sounds like a great plan.
Though, using tubes that require 6.3V DC heaters complicates things because now you'll need a 9VAC transformer. 2.5V and 5V heaters can use the existing 6.3VAC tap from the power transformer.
Electra-Print makes the best transformers anyway, and besides, you are in the USA. The price just depends on the DC bias current handling ability, and whether you want it made of copper, partial silver, or pure silver. I also get my transformers from Transcendar, which is another option.
Though, using tubes that require 6.3V DC heaters complicates things because now you'll need a 9VAC transformer. 2.5V and 5V heaters can use the existing 6.3VAC tap from the power transformer.
Electra-Print makes the best transformers anyway, and besides, you are in the USA. The price just depends on the DC bias current handling ability, and whether you want it made of copper, partial silver, or pure silver. I also get my transformers from Transcendar, which is another option.
I power my 120 ohm headphones (and my 60 ohm headphones) from the same OT winding as my 6 ohm speakers (10k : 4R). Works great! 40 ohm phones would probably work even better.
My schematic wasn't single ended, but since the OP said he needed a lot of power... Anything over a few watts is a tall order on SE, like a 1000 volts supply tall order.
My schematic wasn't single ended, but since the OP said he needed a lot of power... Anything over a few watts is a tall order on SE, like a 1000 volts supply tall order.
10W or more is attainable with the 300B at plate voltages of 400V from experience. (Distortion levels at a saner 1 - 2W are suitable for headphone drive)
The issue isn't power, these new style headphones only need about 2W headroom, the issue is the voltage swing required to pull that off across a 38ohm coil. Its asking a lot. But finding the right output transformer is the key.
I am trying to picture how two 16 ohm secondaries tried together to give balance output would work. The calculation for voltage swing required across the primary is less than a single 32 ohm secondary because its a squared relation.
With balanced out I could use a input transformer for balanced in will give give the amp enough overall gain to make it all DHT, ie :
input transformer-4p1L-6C4c-3k:16+16.
Balanced is the way to go, its easier to find a transformer winder who can make two 16 ohm secondaries than one who can make a 32 ohm secondary correctly (they almost always get it wrong as in two much DCR.)
Thanks for talking us thru this.
Now I just have to build a dam balanced out DAC
Last edited:
Actually the issue is power, if you use a transformer with the appropriate secondary impedance the voltage issue will take care of itself. Noting that impedance scales as the square of the turns ratio two 8 ohm windings in series would give you 32 ohms, two 9.5 ohm windings in series would give you 38 ohms and two 10 ohm windings in series would give you 40 ohms. Getting a custom output transformer built by Electra-Print would provide the required output voltage to achieve 2W or more into a 32 - 40 ohm load as I mentioned previously, an off the shelf transformer with a 16 ohm secondary would be about 3dB worse than a 32 ohm secondary from the perspective of linear voltage swing, but a 300B would give you some margin to spare.
Note that two 16 ohm secondaries in series gives you a 64 ohms and could be a pretty severe mismatch unless you boost the primary impedance to compensated..
I can't get it why it should be so hard to develop 2W over 38 ohms. Quick back of the envelope calculations show that you need 9VRMS for that. Using a 5k:38R transformer implies in a transformation rate of about 12. That means that 9VRMS x 12 = 108VRMS should be developed on the primary. That is something like 304V peak to peak. I think the 300B is quite capable of that? Output impedance would be around 800R / 144 = 5,55 ohms. High load of 5k on 300B = low distortion. With 300B one could even use something as 3k:38R, than even less swing is needed, but output impedance rises.
Or am I making a wrong step in my (back of the envelope) calculations?
Or am I making a wrong step in my (back of the envelope) calculations?
Just stumbled across this:
Audio Note Kits -HE-6 Headphone Amplifier
Coolest output transformers ever?
Thanks,
Chris
Audio Note Kits -HE-6 Headphone Amplifier
Coolest output transformers ever?
Thanks,
Chris
I don't see why. Reason would dictate the use of single double C loop, or a double one if symmetry is important, but other than that, as long as the available copper window is properly filled with iron, I see no particular problem.
More complicated (uselessly), more eye-catching, but technically no other problem -IMHO-
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.