[L0rdGwyn]

What's the advantage of the EL83 over the EL84?

It has a little more gain than the EL84 with similar plate resistance. But if I am being honest, I think the high gm Russian tetrodes probably sound better, will do one more A-B, but more likely to go that direction.

 

[tubebuyer2020]

My dac is much smaller than your dac. I could live in Tokyo!

I don't know, Tokyo guys are on a whole other level:

Normal electricity just wouldn’t do anymore. To tap into what Mr. Morita calls “pure” power, he paid $10,000 to plant a 40-foot-tall concrete pole in his front yard. On it perches his own personal transformer—that thing shaped like a cylindrical metal garbage can—which feeds power more directly from the grid.
...
“It’s completely beyond my understanding,” says his wife, Reiko, 57. “But if I take it away from him, he will lose the motivation to live.”

https://www.wsj.com/articles/a-gift...ave-it-all-a-personal-utility-pole-1471189463

 

[Mr Trev]

A "spud" is a tube amp that only uses one tube per channel, typically transformer coupled, either in a SET output stage or parafeed. Given there is only a single tube, these are usually low power designs, but work very well as headphone amplifiers. Tubes for spud designs need to have both high gain and low plate resistance, so are usually high transconductance types.



A spud is really a single stage SET or parafeed amplifier, there is basically just a high-gain transformer-coupled output stage without an input stage. Since headphones don't need a lot of power, spud designs make a lot of sense - keeps the signal path short with few components which lends itself well to big soundstage and high degree of clarity. Won't be huge on power, likely around 500mW (max rating for the Sowter 8665 transformers), but that is plenty for the vast majority of headphones. I don't know exactly how compact yet, I actually started drafting a chassis design the other day - going to start with 8" L x 10" W x 2.25" H and see where it goes.

I tried the EL83 - both sound good, not sure which direction I will go. The EL83 is a good looking tube.



I have a bunch of E55L tubes which are a popular spud choice, but I am saving those for a LCR phono stage

I take it then it wouldn't come close to enough current to drive planars?

 

[L0rdGwyn]

I take it then it wouldn't come close to enough current to drive planars?

Probably depends on the planar and its sensitivity, output impedance of around 10ohms on the low-impedance tap if using 6E6P or 6E5P. I don't own any real power hungry planars so I cannot test subjectively, but I've picked up around Head-Fi people like to use high-wattage amplifiers with their planars, even if they are using less than 1% of the total power! Seems like overkill to me, but lots of people are doing it so maybe there is something to it. In theory, even 250mW would drive any planar to unlistenable volumes. 250mW should get the HE-6 to 108dB peaks (going purely by the spec sheet). 500mW should be enough for any headphone, but like I said, I hear people like to drive their planars with massive amounts of unused power on tap and feel there is some advantage there, but I haven't tried it myself.

If high power is needed, a more traditional two-stage SET design is probably the way to go over a spud.

 

[bcowen]

Probably depends on the planar and its sensitivity, output impedance of around 10ohms on the low-impedance tap if using 6E6P or 6E5P. I don't own any real power hungry planars so I cannot test subjectively, but I've picked up around Head-Fi people like to use high-wattage amplifiers with their planars, even if they are using less than 1% of the total power! Seems like overkill to me, but lots of people are doing it so maybe there is something to it. In theory, even 250mW would drive any planar to unlistenable volumes. 250mW should get the HE-6 to 108dB peaks (going purely by the spec sheet). 500mW should be enough for any headphone, but like I said, I hear people like to drive their planars with massive amounts of unused power on tap and feel there is some advantage there, but I haven't tried it myself.

If high power is need, a more traditional two-stage SET design is probably the way to go over a spud.

The Schiit Lyr 3 puts out a little over 9 watts into the DCA Aeon 2's. I felt it was pretty sufficient.

 

[L0rdGwyn]

I actually feel pretty confident this spud is done, 6E6P it is. It's basically the input stage of my cap-coupled 45 SET build, except instead of a 45 output stage, there is just an output transformer! Bigger coupling cap too, for parafeed.

Around 9-10 o'clock on the spud is pretty darn loud for a 97dB/mW 300ohm headphone. Full tilt on the pot hits over 500mW in the low, medium, and high impedance settings into 32ohms, 120ohms, 300ohms, respectively with a 2Vrms input.

Square waves look good, THD at 1mW into 300ohms is 0.048% H2 dominant.

 

[Xcalibur255]

I'm going to bring my annual tradition of mentioning that somebody should do an EML 30A or 20B spud over to this thread. Because there's a chance Keenan just might do it someday. I suspect 30A is the harder reach because of the higher resistance but it's the only one that really seems like it has enough gain too.

 

[Xcalibur255]

Probably depends on the planar and its sensitivity, output impedance of around 10ohms on the low-impedance tap if using 6E6P or 6E5P. I don't own any real power hungry planars so I cannot test subjectively, but I've picked up around Head-Fi people like to use high-wattage amplifiers with their planars, even if they are using less than 1% of the total power! Seems like overkill to me, but lots of people are doing it so maybe there is something to it. In theory, even 250mW would drive any planar to unlistenable volumes. 250mW should get the HE-6 to 108dB peaks (going purely by the spec sheet). 500mW should be enough for any headphone, but like I said, I hear people like to drive their planars with massive amounts of unused power on tap and feel there is some advantage there, but I haven't tried it myself.

If high power is needed, a more traditional two-stage SET design is probably the way to go over a spud.

My personal theory about this has always been that the high power design ends up offering some secondary benefits such as improved dynamics (due to the robust power supply) and possibly good compatibility with difficult loads because it has ample reserves. I agree with you though, the groupthink around "needing" lots of power just doesn't jive with the science behind how all this stuff works. 500mW is adequate for any headphone on Earth if the entire design is good, unless the goal is to give yourself hearing damage quickly.

 

[Mr Trev]

Probably depends on the planar and its sensitivity, output impedance of around 10ohms on the low-impedance tap if using 6E6P or 6E5P. I don't own any real power hungry planars so I cannot test subjectively, but I've picked up around Head-Fi people like to use high-wattage amplifiers with their planars, even if they are using less than 1% of the total power! Seems like overkill to me, but lots of people are doing it so maybe there is something to it. In theory, even 250mW would drive any planar to unlistenable volumes. 250mW should get the HE-6 to 108dB peaks (going purely by the spec sheet). 500mW should be enough for any headphone, but like I said, I hear people like to drive their planars with massive amounts of unused power on tap and feel there is some advantage there, but I haven't tried it myself.

If high power is needed, a more traditional two-stage SET design is probably the way to go over a spud.

Fair enough.
Send a spud my way and I'll let you know how it works out

 

[L0rdGwyn]

I'm going to bring my annual tradition of mentioning that somebody should do an EML 30A or 20B spud over to this thread. Because there's a chance Keenan just might do it someday. I suspect 30A is the harder reach because of the higher resistance but it's the only one that really seems like it has enough gain too.

That could be fun, only tricky part - those tubes have pretty high plate resistance, lowest on the 20B at 3K3. That means you are looking at a output transformer with a primary impedance of 10K or more. Transformer parasitics and bandwidth can become issues with a high primary impedance. Transformers would probably have to be custom and still unlikely to get a nice flat response. But if you can sacrifice all that, it could be done! I actually have some LL9202 50mA transformers sitting around here somewhere, can be wired for 11K, could work but the HF would probably roll off around 15kHz based on what I've seen. They are good sounding transformers, even so. Gain also is an issue with an amplification factor of 20 using a high impedance OPT, for all that probably would get less than 250mW with a 11K load - probably would necessitate a step-up transformer to swing some voltage on the 20B grid. Could use some local negative feedback to improve the bandwidth and output impedance, but again, already limited on available gain. Then there's the filament supply, and with a SUT, another place to potentially couple magnetic noise. Things start to get complicated pretty fast lol it hurts the brain. Could definitely be done, but wouldn't be easy peasy.

 

[L0rdGwyn]

Fair enough.
Send a spud my way and I'll let you know how it works out

If / when I build one, will probably send on a short tour

 

[triod750]

If / when I build one, will probably send on a short tour

Since when was Sweden a short tour?
Ah, yes, from Norway it is. Stupid me.

 

[Xcalibur255]

That could be fun, only tricky part - those tubes have pretty high plate resistance, lowest on the 20B at 3K3. That means you are looking at a output transformer with a primary impedance of 10K or more. Transformer parasitics and bandwidth can become issues with a high primary impedance. Transformers would probably have to be custom and still unlikely to get a nice flat response. But if you can sacrifice all that, it could be done! I actually have some LL9202 50mA transformers sitting around here somewhere, can be wired for 11K, could work but the HF would probably roll off around 15kHz based on what I've seen. They are good sounding transformers, even so. Gain also is an issue with an amplification factor of 20 using a high impedance OPT, for all that probably would get less than 250mW with a 11K load - probably would necessitate a step-up transformer to swing some voltage on the 20B grid. Could use some local negative feedback to improve the bandwidth and output impedance, but again, already limited on available gain. Then there's the filament supply, and with a SUT, another place to potentially couple magnetic noise. Things start to get complicated pretty fast lol it hurts the brain. Could definitely be done, but wouldn't be easy peasy.

Yeah I think most people who do it use a 2:1 or 4:1 step-up. I'm sure it would be an expensive build to get right. A guy can dream though.

 

[Monsterzero]

I've picked up around Head-Fi people like to use high-wattage amplifiers with their planars, even if they are using less than 1% of the total power!

My personal theory about this has always been that the high power design ends up offering some secondary benefits such as improved dynamics (due to the robust power supply) and possibly good compatibility with difficult loads because it has ample reserves. I agree with you though, the groupthink around "needing" lots of power just doesn't jive with the science behind how all this stuff works. 500mW is adequate for any headphone on Earth if the entire design is good, unless the goal is to give yourself hearing damage quickly.

I love my vintage receivers as headphone amps. They can drive the difficult -to-drive K240 sextett and K340 without breaking a sweat. Similarly on the OG LCD2.1 as well as the HE500, no probs. However the HE6 was woefully undertpowered on the same receivers and I only understood the hype on the HE6 once I got a speaker tap cable for it.
I used everything from a 55 watt Sansui to a 125 watt Kenwood to drive the HE6, and the difference between speaker taps and headphone jack was the obligatory night/day difference.

 

[bpiotrow13]

That could be fun, only tricky part - those tubes have pretty high plate resistance, lowest on the 20B at 3K3. That means you are looking at a output transformer with a primary impedance of 10K or more. Transformer parasitics and bandwidth can become issues with a high primary impedance. Transformers would probably have to be custom and still unlikely to get a nice flat response. But if you can sacrifice all that, it could be done! I actually have some LL9202 50mA transformers sitting around here somewhere, can be wired for 11K, could work but the HF would probably roll off around 15kHz based on what I've seen. They are good sounding transformers, even so. Gain also is an issue with an amplification factor of 20 using a high impedance OPT, for all that probably would get less than 250mW with a 11K load - probably would necessitate a step-up transformer to swing some voltage on the 20B grid. Could use some local negative feedback to improve the bandwidth and output impedance, but again, already limited on available gain. Then there's the filament supply, and with a SUT, another place to potentially couple magnetic noise. Things start to get complicated pretty fast lol it hurts the brain. Could definitely be done, but wouldn't be easy peasy.

I think Emission Labs's distributor in Europe published some schematics related to 20b amplifier, including headphone one.

http://www.jacmusic.com/kits/pre-amp/20B-V4-Low-Noise/20B-V4-index.htm

And, pardon my ignorance, if output transformer is a problem, could't be an OTL design? The more i listen ro my Verite the more I like this design.