I've been experimenting with an all-tube headphone amplifier for a bit and have more or less settled on the 6gc5 pentode for power out. It prefers B+under 200vdc and I'm particularly fond of the punchy sound. Triodes using the same B+ supply wind up with 70-120v on their anodes because of my plate resistor choices. (depends on tube, bias and resistor value) I'm really satisfied with the sound as-is, but it's leaving me wondering if it would be even better if I could be giving the tubes more of what they want, you know? Are there any good twin-triodes out there that perform their best (or thereabout) within that voltage range? I've tried 6FQ7, 6EU7, 12AT7 and 5963 so far. Some give specs for my voltage range but the tube specifications state that such voltages are far from optimal. Admittedly I may be reading them wrong though. I'm also picking up a boat load of RF from the audio source cable. I can easily hear everything that my bench PC (audio source) is doing. Hi-mu triodes seem to exacerbate that but I'm going to have to take measures to stop it in addition to using a shielded enclosure and cables. Maybe bias the front-end cold to lower gain and/or use medium mu triodes? I'm not sure. As-is, it has a pretty stupid amount of headroom so lowering output isn't an issue whatsoever.
Last edited:
How are you loading the tube output and matching the HP impedance? What specs?
I'm going to have a look at the 6DJ8. Thank you for the suggestion.
I've loaded the 6GC5 three ways. I initially used an Edcor SE OPT with a 5k primary and 8ohm (or similar) secondary. Sounded good in both ultra linear and triode modes. Then I moved on to a Dukane 710-3092 line matching transformer. Initially I hooked it up just like any other SE OPT and got decent results. Application of a large amount of negative feedback got better results. After that, I learned about non-air gapped transformers tending to saturate when they had DC applied to them. I then used the same transformer like one would do inter-stage coupling between tubes. 20k 5w or thereabout anode resistor. Higher prevented the tube from drawing appropriate current. Screen grid tied to the anode with a 1k resistor. I used a 7.5 uF motor run film capacitor to couple the anode to the primary of the transformer. Other end of the transformer primary to ground. (orange and black wires) The secondary of said transformer is either 4 or 8 ohms. No negative feedback applied because it honestly wasn't necessary. My headphones are nothing special. Probably in the neighborhood of 20 ohms but I'm not sure. My goal through all of this was and is to make something cheap (if non-standard) that produces good sound. Aside from the aforementioned RF, I have been exceptionally happy with the sound. The OPT in particular has really surprised me with how well it performs. They cost $2.50 at parts express and are rated to four watts. I highly recommend others try them even if they want nothing to do with any other aspect of my design. They're just so cheap.
Sorry for all the edits... I really need to start to proofreading. Some of my typos can cause eye damage...
I've loaded the 6GC5 three ways. I initially used an Edcor SE OPT with a 5k primary and 8ohm (or similar) secondary. Sounded good in both ultra linear and triode modes. Then I moved on to a Dukane 710-3092 line matching transformer. Initially I hooked it up just like any other SE OPT and got decent results. Application of a large amount of negative feedback got better results. After that, I learned about non-air gapped transformers tending to saturate when they had DC applied to them. I then used the same transformer like one would do inter-stage coupling between tubes. 20k 5w or thereabout anode resistor. Higher prevented the tube from drawing appropriate current. Screen grid tied to the anode with a 1k resistor. I used a 7.5 uF motor run film capacitor to couple the anode to the primary of the transformer. Other end of the transformer primary to ground. (orange and black wires) The secondary of said transformer is either 4 or 8 ohms. No negative feedback applied because it honestly wasn't necessary. My headphones are nothing special. Probably in the neighborhood of 20 ohms but I'm not sure. My goal through all of this was and is to make something cheap (if non-standard) that produces good sound. Aside from the aforementioned RF, I have been exceptionally happy with the sound. The OPT in particular has really surprised me with how well it performs. They cost $2.50 at parts express and are rated to four watts. I highly recommend others try them even if they want nothing to do with any other aspect of my design. They're just so cheap.
Sorry for all the edits... I really need to start to proofreading. Some of my typos can cause eye damage...
Last edited:
I got a pair of 6dj8's off of Fleabay. Looked up the datasheet and the voltages are bang-on for what my power supply delivers. Thank you for the recommendation.
With an easy load, 12AX7 works fine at low voltage. It descends from 1st Audio tubes for domestic radios, often worked at 100V-150V.
The 25L6 (even if called 6GC5) is able to deliver about 50 times as much power as most headphones need. If _I_ were to "optimize", I'd design-out that huge heater demand, at least for use in hot weather. My last tube headphone amp used 1/4 the heater power and was still loud and clean enough for any use.
The 25L6 (even if called 6GC5) is able to deliver about 50 times as much power as most headphones need. If _I_ were to "optimize", I'd design-out that huge heater demand, at least for use in hot weather. My last tube headphone amp used 1/4 the heater power and was still loud and clean enough for any use.
B+ is 170v.
Yes, they do deliver much more power than is technically necessary. I have them biased on the cold side and am limiting the volume on the front end though. Filament draw isn't much of an issue for me. I already had an edcor transformer laying around. Even then, that transformer is cheap. $35 or thereabouts. I've tried several other configurations with lower gain, but it's the sound that's attracted me to the 6GC5. I don't claim to be an audiophile, but they just sound better to my ear. Others would have great and clear highs (if only a bit harsh without feedback) but the low end would lack definition. (Both with line transformer and actual SE OPT.) They still sounded good, don't get me wrong. I just kept trying different tubes from my stash and these really impressed me. Bass is extremely well defined and has a nice thump to it. I don't really like bass heavy music per se, but some songs just call for it here and there and it really comes out and says hello in this config without the volume having to be turned up really high. It pulls that off while retaining the crisp and detailed (but not harsh) high end. I dunno. I just really like it.
Yes, they do deliver much more power than is technically necessary. I have them biased on the cold side and am limiting the volume on the front end though. Filament draw isn't much of an issue for me. I already had an edcor transformer laying around. Even then, that transformer is cheap. $35 or thereabouts. I've tried several other configurations with lower gain, but it's the sound that's attracted me to the 6GC5. I don't claim to be an audiophile, but they just sound better to my ear. Others would have great and clear highs (if only a bit harsh without feedback) but the low end would lack definition. (Both with line transformer and actual SE OPT.) They still sounded good, don't get me wrong. I just kept trying different tubes from my stash and these really impressed me. Bass is extremely well defined and has a nice thump to it. I don't really like bass heavy music per se, but some songs just call for it here and there and it really comes out and says hello in this config without the volume having to be turned up really high. It pulls that off while retaining the crisp and detailed (but not harsh) high end. I dunno. I just really like it.
Throwing my suggestions into the pot
Medium mu gain:
6N16B, or some of the similar wire ended triodes seems to work well at the HT you are using (say 6111, 6112, 6021, 6N16B, 6S6B) are all pretty linear with anodes between 100V and 150V, at reasonable currents. Though like most, I have found that some exhibit even lower THD at more elevated HT and anode V.
For an output, maybe 6S19P? Triode wired 6P30BR? (I'm aware these are almost two ends of the current spectrum) either way, I suspect they are "man enough" for headphones. (Though the 6P30BR would like a 10k load far more than a 5k one, and the 6S19P probably the reverse, 2.5k may suit better)
I've used triode wired DHP (1j29B/1j24B) to good effect in a headphone amp up to about 50mW (300R headphones), and the larger 1P24B, with a larger dissipation, Inhave no doubt would work well for Lowe Z phones
That's my experience exhausted lol.
Medium mu gain:
6N16B, or some of the similar wire ended triodes seems to work well at the HT you are using (say 6111, 6112, 6021, 6N16B, 6S6B) are all pretty linear with anodes between 100V and 150V, at reasonable currents. Though like most, I have found that some exhibit even lower THD at more elevated HT and anode V.
For an output, maybe 6S19P? Triode wired 6P30BR? (I'm aware these are almost two ends of the current spectrum) either way, I suspect they are "man enough" for headphones. (Though the 6P30BR would like a 10k load far more than a 5k one, and the 6S19P probably the reverse, 2.5k may suit better)
I've used triode wired DHP (1j29B/1j24B) to good effect in a headphone amp up to about 50mW (300R headphones), and the larger 1P24B, with a larger dissipation, Inhave no doubt would work well for Lowe Z phones
That's my experience exhausted lol.
Last edited:
Look for something designed for old battery operated sets for output Z closer to your tranny Z. Something like a 3Q5-GT has a 110v plate and 8K primary and can deliver about .5W. You could maybe do a parallel setup for more power and put the heaters in series for a 6v heater supply. It's not a triode but you could strap it or just keep it pentode, why not?
Last edited:
Thank you for all the recommendations. I've got some research to do. This all got me thinking more about output impedance. I am now considering buying some dedicated Edcor OPTs to see how they sound compared to what I'm currently running. I've included a portion of the 6GC5 datasheet which seems to indicate an ideal transformer impedance for this particular tube of ~2k ohms for lowest distortion and highest power out (even if it's not needed per se). Is that right? They have several over there rated for both 1.7k and 2.5k. I've also tried a 6GF7A dissimilar duo-triode in a "one-tube-per-channel" situation. It's a High mu/Low-mu triode tube. It doesn't have the bass that the 6GC5 does, but that may also be an impedance issue as well. Highs are wonderful but the low end is just kind-of jumbled and mushy. Power for days though.
I was previously running a voltage doubler to get higher B+ for the tubes that prefer it but started running into noise and I assumed that it may have been some sort of grounding deficiency with my design. I built a new 350v power supply using a vintage tranny from an old radio and applied 5u4 rectification in lieu of a doubler but the interference remained. And I mean a LOT of it too. Buzzing, popping, whining. Every little bit of anything that the PC did came through the headphones and I didn't have another audio source to try. I think I have that figured out now though. I think I'm getting a ground loop through the source cable to my PC. Up until about a week ago I got minimal interference, which I just attributed to the unshielded rats nest of wires all over my breadboards. Something changed though. That PC came over on the Mayflower though so who knows what it could have been. Loads of people in other forums describing the exact problem I'm having though, so I'm going to attempt to either build or just purchase a "humbucker" audio source isolation board.
I was previously running a voltage doubler to get higher B+ for the tubes that prefer it but started running into noise and I assumed that it may have been some sort of grounding deficiency with my design. I built a new 350v power supply using a vintage tranny from an old radio and applied 5u4 rectification in lieu of a doubler but the interference remained. And I mean a LOT of it too. Buzzing, popping, whining. Every little bit of anything that the PC did came through the headphones and I didn't have another audio source to try. I think I have that figured out now though. I think I'm getting a ground loop through the source cable to my PC. Up until about a week ago I got minimal interference, which I just attributed to the unshielded rats nest of wires all over my breadboards. Something changed though. That PC came over on the Mayflower though so who knows what it could have been. Loads of people in other forums describing the exact problem I'm having though, so I'm going to attempt to either build or just purchase a "humbucker" audio source isolation board.
Attachments
Last edited:
Something else to be aware of is the affect of running a @24 ohm load on 8 ohm taps. This 3x resistance will create 3x primary impedance for the tube to drive so if you start with 5K ohms you can estimate 15K ohms the tube will be pushing into. That cuts the power the tube can create. So you should look for a 16 ohm tap, and match your tube's load impedance and tranny primary Z, close as you can get. Maybe undershoot the primary Z by half if you can't get a 16 ohm tap. Doesn't have to be perfect by any means just get the dart on the board. If the tube you are using now wants 2K and you are giving it 15K, that's not on the board. Small signal triodes you are interested in usually want 50K-100K to drive into.
Last edited:
I've been using dual dissimilar vertical oscillator/amplifier triodes with good results, choosing ones with higher filament voltages so I can run them with a switching adapter. 13DR7 works quite nicely. I've been running with a 150V supply generated by a separate DC-DC converter of my own design. I've been using Transcendar output transformers - somewhat ugly, but serviceable.
Ground the input of your amp and see if the noise persists. Floating heater supplies can cause hum.
The 6GC5 is a 6W6 in a fat 9 pin bottle. The 25L6 is a 6W6 with a 25 volt heater. All of these make nice audio amps. I tend to use the 6W6 since it is cheap and shares the same pinout as most of the common audio tubes. If the amp uses a 6W6, then you could try 6K6's, 6V6's and even 6L6GC's if your transformer had the heater current capability. A triode wired 6V6 sounds nice too.
Using a tube capable of several watts to drive a pair of headphones is overkill, but it works pretty nice. I have my 32 ohm Sennheisers connected directly to the 8 ohm speaker output of my 2A3 TSE-II. This loads the 2A3 with about 12K, so I only get a watt......still too much power. I tried adding resistors in parallel with the phones to present a lower impedance to the tube, but for most music, I prefer no resistor.
Been there, fought that. I could even hear the data flying about every time I moved the mouse. I built a portable PC with 4 inch full range speakers and a 20 WPC class D amp......same stuff. It was real bad with the motherboard audio and the audio amp running on PC power. A USB sound box and a separate laptop brick to power the class D amp broke the loop and killed the noise.