Hi all 😀
For my next project I'd like to build a DHT headphone amplifier. My primary requirement at this point is that it use only currently-manufactured tubes - reliably acquiring NOS tubes here is becoming harder and harder. So unfortunately the likes of the wonderful 45 are not an option. Plus I intend to tinker with the project a fair bit (trying various CCSes, filament regulators, etc) so tube casualties may occur. I've thus chosen the 2A3 - I adore the sound and plenty of companies still make them. I'm open to alternatives though.
I've spent many many hours looking over various schematics, including the TSEII, the Free Lunch, and many others. Most call for tubes that have been out of manufacture for ages.
So: what recommendations do the wise have for a core circuit that uses new tubes?
At first, I'll use a standard OPT with load resistors and once I'm happy with the circuit I'll have an appropriate transformer custom wound.
For my next project I'd like to build a DHT headphone amplifier. My primary requirement at this point is that it use only currently-manufactured tubes - reliably acquiring NOS tubes here is becoming harder and harder. So unfortunately the likes of the wonderful 45 are not an option. Plus I intend to tinker with the project a fair bit (trying various CCSes, filament regulators, etc) so tube casualties may occur. I've thus chosen the 2A3 - I adore the sound and plenty of companies still make them. I'm open to alternatives though.
I've spent many many hours looking over various schematics, including the TSEII, the Free Lunch, and many others. Most call for tubes that have been out of manufacture for ages.
So: what recommendations do the wise have for a core circuit that uses new tubes?
At first, I'll use a standard OPT with load resistors and once I'm happy with the circuit I'll have an appropriate transformer custom wound.
FWIW, I've seen a 3 stage design that uses the 2 sections of a 6SN7, for voltage amplification, and a 2A3 "final". That sort of setup would surely work for you.
Decide if you want currently manufactured dual plate 2A3 tubes,
Or, if you want currently manufactured mono-plate 2A3 tubes.
I have used both.
And, it is almost time for me to build again, this time with DC filaments, not AC filaments.
I will almost certainly use the currently manufactured mono-plate 2A3 tubes.
I have JJ 2A3 mono-plate tubes (the glass is larger, and the plates are rated at 40W dissipation).
I also have Sovtek and Electro Harmonix mono-plate tubes.
I only have a single Valve Art 2A3 left (dual plate).
I would use any of the above, but might have an opinion of my favorite when I use DC filaments this time.
Output transformers have tradeoffs:
Generally, higher primary impedance has less output power, but lower distortion, and higher damping factor.
Generally, lower primary impedance has more output power, but higher distortion, and lower damping factor.
I will use a 2 stage circuit. I use CD players, and Computer audio output, and both have enough signal voltage to only need 2 stages. The driver plate will either use a current source, or a resistor as the plate load. I will not use negative feedback, this is a 2A3 triode, and with 3k or up to 5k output transformer, will have reasonable damping and reasonably low distortion at low power output (mostly used for near-field listening).
Straightforward simple designs usually work well for me.
Or, if you want currently manufactured mono-plate 2A3 tubes.
I have used both.
And, it is almost time for me to build again, this time with DC filaments, not AC filaments.
I will almost certainly use the currently manufactured mono-plate 2A3 tubes.
I have JJ 2A3 mono-plate tubes (the glass is larger, and the plates are rated at 40W dissipation).
I also have Sovtek and Electro Harmonix mono-plate tubes.
I only have a single Valve Art 2A3 left (dual plate).
I would use any of the above, but might have an opinion of my favorite when I use DC filaments this time.
Output transformers have tradeoffs:
Generally, higher primary impedance has less output power, but lower distortion, and higher damping factor.
Generally, lower primary impedance has more output power, but higher distortion, and lower damping factor.
I will use a 2 stage circuit. I use CD players, and Computer audio output, and both have enough signal voltage to only need 2 stages. The driver plate will either use a current source, or a resistor as the plate load. I will not use negative feedback, this is a 2A3 triode, and with 3k or up to 5k output transformer, will have reasonable damping and reasonably low distortion at low power output (mostly used for near-field listening).
Straightforward simple designs usually work well for me.
Last edited:
Thank you for the input.
I have Electro Harmonix 2A3s, with some JJs on the way from a local distributor. A 2 stage circuit would likely do the trick for me as well - my DAC has 3 selectable output levels. Since this will be for headphones, a higher primary will be very suitable.
6A3sUMMER, what circuit are you planning on using?
I have Electro Harmonix 2A3s, with some JJs on the way from a local distributor. A 2 stage circuit would likely do the trick for me as well - my DAC has 3 selectable output levels. Since this will be for headphones, a higher primary will be very suitable.
6A3sUMMER, what circuit are you planning on using?
I will probably use a 12AY7 in parallel, with a 900V IXYS current source for the plate load.
I will have separate self bias resistors and separate bypass caps for each cathode.
I will use separate grid stopper resistors.
I sometimes needed to use individual 1k resistors from the plates to the IXYS part, when using the 12AY7 in parallel.
About 3 mA per triode section (6mA from the IXYS).
RC coupling from the 12AY7 to the 2A3.
The 2A3 will have DC for the filaments.
It will have a self bias resistor and bypass cap, and a series pair of resistors across the filaments, with the junction going to the top of the self bias network.
The plate load will be either 3k or 5k, I have transformers of both those impedances.
I will not use any negative feedback of any form.
I can get about 300V B+ with solid state diodes, using a choke input filter.
I will use Schottky diodes for the 2A3 filament supply, with a secondary voltage high enough that I can use a CRC filter, and have the R large enough to get good filtering. All that is Brute Force.
There will be lots of filtering of B+ and of the 2A3 DC filament supply.
My target is 500uV ripple or less, my other amps have less than 100uV ripple,
I will have to see if I can get the ripple that low.
I expect it to be that simple, and start with the circuit as described above.
Do not forget to allow extra room around and above the JJ 2A3 tubes.
I will have separate self bias resistors and separate bypass caps for each cathode.
I will use separate grid stopper resistors.
I sometimes needed to use individual 1k resistors from the plates to the IXYS part, when using the 12AY7 in parallel.
About 3 mA per triode section (6mA from the IXYS).
RC coupling from the 12AY7 to the 2A3.
The 2A3 will have DC for the filaments.
It will have a self bias resistor and bypass cap, and a series pair of resistors across the filaments, with the junction going to the top of the self bias network.
The plate load will be either 3k or 5k, I have transformers of both those impedances.
I will not use any negative feedback of any form.
I can get about 300V B+ with solid state diodes, using a choke input filter.
I will use Schottky diodes for the 2A3 filament supply, with a secondary voltage high enough that I can use a CRC filter, and have the R large enough to get good filtering. All that is Brute Force.
There will be lots of filtering of B+ and of the 2A3 DC filament supply.
My target is 500uV ripple or less, my other amps have less than 100uV ripple,
I will have to see if I can get the ripple that low.
I expect it to be that simple, and start with the circuit as described above.
Do not forget to allow extra room around and above the JJ 2A3 tubes.
Last edited:
I'd be interested in your decision tree for cathode resistor biasing of the output stage. Always a personal choice - tradition vs. modernism, etc.
Thanks, as always,
Chris
Thanks, as always,
Chris