I am going back to the beginning and building the Radiotron 2A3 single-ended amp. I plan to start with the original and work from there listening as I go. My first mod will be taking the bypass caps to 100uf and the coupling cap to 0.22uf. My question regards the resistor values which are given as what looks like .25, .5, and 1,5. I assume those values are in mega-ohms, correct?
Happy listening,
John
Happy listening,
John
The image posted is from RDH version 3... Chapter 8 page 53. RHD Version 4 page 601 is a slightly different version of the amplifier. The input coupling cap is changed to a 0.1uF cap and the input tube from a 6C6 and a 6J7.
I would also comment to the OP about simply changing cap values for coupling and bypass on the 2A3. I wouldn't expect much, if any, improvement. Increasing the lower frequency coupling and bypass (for the 2A3) could also result in earlier core saturation the OPT, depending on the source material and the specification of the OPT, which is not listed beyond the primary impedance. After extensive testing with over 60- 2A3 tubes over the decades, I would recommend a 3.5K load... better linearity, lower distortion and easily able to get 4-watts output.
Have fun.
I would also comment to the OP about simply changing cap values for coupling and bypass on the 2A3. I wouldn't expect much, if any, improvement. Increasing the lower frequency coupling and bypass (for the 2A3) could also result in earlier core saturation the OPT, depending on the source material and the specification of the OPT, which is not listed beyond the primary impedance. After extensive testing with over 60- 2A3 tubes over the decades, I would recommend a 3.5K load... better linearity, lower distortion and easily able to get 4-watts output.
Have fun.
I don't know if you know him, but he has some things to tell you about that amp.
https://jelabs.blogspot.com/search?q=ukraine
http://jelabsarch.blogspot.com/2012/08/radiotron-se2a3_5.html
https://jelabs.blogspot.com/search?q=ukraine
http://jelabsarch.blogspot.com/2012/08/radiotron-se2a3_5.html
Some interesting info, but not a bit surprise in the results.... one major improvement... OPT, a Tango U-808 at 3.5K primary, higher cathode resistor value of 880 ohms (more negative grid bias) and a higher plate supply voltage of 330V. Pretty much what I use for a simpler 2A3 output stage; 3.5K load, 866 ohm cathode bias resistor and similar plate supply voltage (350v). The 4.5 watt output spec is the same for my 2A3 design.
Other changes include the driver, increased idle current and a stronger drive ratio between the plate load resistor and the 2A3 grid resistor. Still, I don't see the need for such large bypass capacitors for both cathode bias resistors. Even the U-808 3.5K OPT has a frequency spec of 35Hz to 60KHz at -2 dB points.
However.... I used a completely different input/driver stage. I would also refer to Gordon Rankin's Bugle 45 document (attached) showing his design and the various formulae to calculate specific part values and such. One key area is the idle current in the driver stage. This is calculated against slew rate and voltage swing to ensure adequate bandwidth.
My earlier design (circa 2007) used for the 45 and 2A3 was based on Hashimoto iron. Using a drive current of 2.8ma and a plate load of 54.9K against a grid resistor of 249K (coupled with a 0.22uF cap) resulted in a frequency response of 25Hz - 50KHz within 1dB at 1-watt output. I've also opted to attach the schematics for these designs (45 and 2A3 SETs). Both use Hashimoto iron and cost a small fortune in parts, even back in 2007.
Needless to say, one can endlessly alter design values and some will make more of a difference than others... and some changes might even make it worse. In any case, have some fun with your build and whatever changes you make.
Other changes include the driver, increased idle current and a stronger drive ratio between the plate load resistor and the 2A3 grid resistor. Still, I don't see the need for such large bypass capacitors for both cathode bias resistors. Even the U-808 3.5K OPT has a frequency spec of 35Hz to 60KHz at -2 dB points.
However.... I used a completely different input/driver stage. I would also refer to Gordon Rankin's Bugle 45 document (attached) showing his design and the various formulae to calculate specific part values and such. One key area is the idle current in the driver stage. This is calculated against slew rate and voltage swing to ensure adequate bandwidth.
My earlier design (circa 2007) used for the 45 and 2A3 was based on Hashimoto iron. Using a drive current of 2.8ma and a plate load of 54.9K against a grid resistor of 249K (coupled with a 0.22uF cap) resulted in a frequency response of 25Hz - 50KHz within 1dB at 1-watt output. I've also opted to attach the schematics for these designs (45 and 2A3 SETs). Both use Hashimoto iron and cost a small fortune in parts, even back in 2007.
Needless to say, one can endlessly alter design values and some will make more of a difference than others... and some changes might even make it worse. In any case, have some fun with your build and whatever changes you make.
yeah, I don't understand it either, looking from the capacitor and the resistors as equivalent to thevening, it gives me something like 47uf (rounding) with a frequency of 20hz(1/2*PI*Fc*R)
I like his stuff! I had read the second one several times, but I had not seen the first. Thanks!
I will try a 3.5 k load as some I can find the appropriate iron. Thanks. You mentioned that you had done a lot of 2A3 testing. Favorites that are possible to obtain?
Hmmm... 2A3 tubes... in my collection I have quite a few of the old RCA Grey and Black plate tubes, which are the basic dual-triodes in parallel, many are in original JAN boxes. I also have quite a few of the Sylvania, Philco, Zenith, Tung-Sol and even a couple matched pairs of the original Cunningham/RCA single-plate versions. I have some Ken-Rad types as well.
From my own testing and listening, I prefer tubes made by Sylvania (which also made them for Zenith) and Philco first. Note however that the internal construction can vary quite a bit. In some of the preferred versions, the filament supports are done using 4- coil springs to tension them at the top mica... other good versions use, what I call, the "fishing rod" filament support for tension at the top mica. These versions typically have very robust internal supports and thick micas and also feature damping wires to the glass envelope.
I also test for mechanical issues... things like tapping the glass envelope to see comes out on the scope... and even playing trumpet next to the tube and observing the scope for any resonances. After all, the music being played by the amplifier itself will tend to be excited by the music itself and introduce distortion from the mechanical issues.
Oddly, I'm not much of a fan of the RCA versions. I know many people rave about them, but they just don't perform as well for me. The single-plate versions (aka the original) can be interesting, but use a split center-tapped filament. This basically requires using a D.C. filament, otherwise it's almost impossible to null the hum out. I would also tend to bias them at no more than 15-watts on the plate as they won't handle as much plate dissipation as the later dual triode versions.
One last note.... What I found many decades ago was that Zenith (my uncle was a dealer in NY for decades in their hey-day) sourced their tubes, typically from Sylvania, but they also required a very tightly spec'd tube. The real old ones were sealed in a rectangular packing with a metal bottom and top cap and an embedded string on the upper half to open them up. I've got NOS 2A3 and 45 triodes in this packaging and would easily state that these are about the best versions of these two DHTs that I have... Ken-Rad and Philco coming second.
Not sure this is overly helpful, but finding good quality NOS DHTs can be a challenge, but possibly worth it if you can score some good ones.
From my own testing and listening, I prefer tubes made by Sylvania (which also made them for Zenith) and Philco first. Note however that the internal construction can vary quite a bit. In some of the preferred versions, the filament supports are done using 4- coil springs to tension them at the top mica... other good versions use, what I call, the "fishing rod" filament support for tension at the top mica. These versions typically have very robust internal supports and thick micas and also feature damping wires to the glass envelope.
I also test for mechanical issues... things like tapping the glass envelope to see comes out on the scope... and even playing trumpet next to the tube and observing the scope for any resonances. After all, the music being played by the amplifier itself will tend to be excited by the music itself and introduce distortion from the mechanical issues.
Oddly, I'm not much of a fan of the RCA versions. I know many people rave about them, but they just don't perform as well for me. The single-plate versions (aka the original) can be interesting, but use a split center-tapped filament. This basically requires using a D.C. filament, otherwise it's almost impossible to null the hum out. I would also tend to bias them at no more than 15-watts on the plate as they won't handle as much plate dissipation as the later dual triode versions.
One last note.... What I found many decades ago was that Zenith (my uncle was a dealer in NY for decades in their hey-day) sourced their tubes, typically from Sylvania, but they also required a very tightly spec'd tube. The real old ones were sealed in a rectangular packing with a metal bottom and top cap and an embedded string on the upper half to open them up. I've got NOS 2A3 and 45 triodes in this packaging and would easily state that these are about the best versions of these two DHTs that I have... Ken-Rad and Philco coming second.
Not sure this is overly helpful, but finding good quality NOS DHTs can be a challenge, but possibly worth it if you can score some good ones.
Years ago, in Sound Practices, someone wrote:
"I never met a 2A3 amplifier that I did not like"
Want to build a 2A3 amplifier from a historical circuit . . . Good!
Later, try a more modern 2A3 circuit.
However, I suggest you always use modern output transformers, modern capacitors, etc.
Building a literal replica with parts created before 1945 is going may end up in poor performance.
Or you may have to take out a second mortgage.
Purcha$e an output transformer that has become OverPriced-ium $$$.
Usually, getting an original output transformer involves either Un-Obtain-ium, Obsolete-ium, or both.
It is Less Expensive to purchase a new, high performance modern transformer.
I remember an EL34 amplifier that used old Black Beauty Coupling Caps (There were many fans of those caps) . . . one was so leaky that one EL34 was at 60mA, and the other EL34 was at 40mA. The output transformer said OUCH! One EL34 said OUCH!
"I never met a 2A3 amplifier that I did not like"
Want to build a 2A3 amplifier from a historical circuit . . . Good!
Later, try a more modern 2A3 circuit.
However, I suggest you always use modern output transformers, modern capacitors, etc.
Building a literal replica with parts created before 1945 is going may end up in poor performance.
Or you may have to take out a second mortgage.
Purcha$e an output transformer that has become OverPriced-ium $$$.
Usually, getting an original output transformer involves either Un-Obtain-ium, Obsolete-ium, or both.
It is Less Expensive to purchase a new, high performance modern transformer.
I remember an EL34 amplifier that used old Black Beauty Coupling Caps (There were many fans of those caps) . . . one was so leaky that one EL34 was at 60mA, and the other EL34 was at 40mA. The output transformer said OUCH! One EL34 said OUCH!
That was Joe Roberts, SP's founder. I forgot completely about that. Will reread tonight.
Thanks,
John
Thanks,
John
Why? Modern are better? Which modern tube circuit do you do? Thanx for the info 6A3sUMMER
Joe was suggesting that you try the original first and then a more modern version for comparison. Exactly what I am doing.
John
John
Hello John,
FWIW, I'm the author of the two Radiotron SE2A3 blog posts referred to in this thread. The original article stored in my blog archive came from my defunct website from the early 2000s.
I revisited the Radiotron SE2A3 a couple of years ago because I still enjoy the sound of the same circuit I worked out with Sergey in the mid 2000s albeit now with G-Trans Finemet FM6WS OPTs (instead of the Tango/Hirata U808s) + Hammond 302AX power transformer as pictured above. This amp serves sort of a benchmark for all my other amp projects.
This is my main SE2A3 amp now. It uses nickel plate choke loaded 6A6s custom wound by dave slagle + Tamura F7002 permalloy output transformers which date back to the Sound Practices homebrewer article I wrote for Joe Roberts in the late 90s.
I can assure you that the schematic in the 3rd edition of the Radiotron works. Build it and listen to lots of music so you get to know it intimately before doing any mods. Being a trained classical musician, I've learned more about this hobby through listening!
Enjoy!
JE
