Greetings!

All new 2A3 Amplifer Design

5/2007
by Jeffery Larson
abraxasaudio

Pentode driven??
No one will believe it!


Index

The 2A3D3aP

This amp started out as an experiment. I wanted a two-stage 2A3 design. I also wanted to avoid the SRPP type input stage because it only works well when buffered correctly. Going SRPP straight to the 2A3 grid was not the ticket. I combined some of my earlier experimenting with pentode and tetrodes, and I liked what I had heard before. So for this design I grabbed the very high gm D3a tube from Siemens, determined to make something special happen.

The D3a. Considerable time and effort went into perfecting this arrangement. Experiments with many operating points and differing loads in the basic form of the 2A3 grid grounding resistors, as well as hard-driving the 2A3 grid resulted in the configuration here. Work on this stage took an entire week of steadfastness. Part of the picture at the time was that I wanted to bring this amp to the 2007 Lone Star Audio Fest and time was short. I dabbled and experimented a bit using the D3a as a triode, but just ran out of time and I was liking the pentode version very much anyway. The pentode design allows a larger than life, low distortion plate voltage swing more than adequate to drive the 2A3 to full power with only a little compromise into the grid current area as to be expected. Overall it's a good balance, and provides for the low distortion character of this amplifier.

The D3a stage is cathode biased with partial bypass. The unbypassed part is used as a feedback input point, and 2.5db global feedback is incorporated into this amplifier. This feedback is mainly used to lower the output impedance somewhat and for minor correction of the usual output transformer issues and NO output transformer is without issues. The feedback is slightly compensated, although if I did this again, I would put the compensation elsewhere.

The cathode bypass caps are Elna Cerafines and bypassed with some foils as well. Plate coupling to the 2A3 is done via some NOS CDE "greeny" caps. I have a handfulof these around and been wanting to try them. One of the few parts I've used I have to say has a break-in period. After that they are terrific.

The output section is grid biased, thus allowing the user to set the cathode current to any reasonable operating point. With the extremely low impedance power supply section feeding this stage, the B+ remains very stable regardles of current, thus allowing accurate bias settings and the usage of fragile NOS 2A3's if desired. Grid bias seems to give more authority to the 2A3 stage than convnetional cathode bias. The cathode circuit has a 10-ohm current sensing resistor for the bias metering circuit.

The output transformers are of a 3K design into 8 ohms with a single secondary winding interleaved over the entire primary for even impedance transfer. Thus they will operate the 2A3 successfully even into 16 ohm loads. Each 2A3 has its own heater transformer and heater hum adjustment control.

Power Supply needs its own section. From a single transformer I was able to create 3 totally separate power supplies. One of very high power and low impedance for the output section, a second heavily filtered moderate impedance section for the the input stage, and finally another low-Z section for the biasing voltage. This pretty much keeps any chance of audio on the B+ leads from feeding into other circuits, and is superior to trying to develop all the voltages from a single secondary winding. The power transformer also supplies the 6.3V heater for the D3a's. As a bonus, the amp could be wired for 230/240V operation if desired.

As mentioned, extensive filtering is used throughout the supply, with a heavy-duty old-stock choke for the output circuit, and a double-CLC configuration for the input stage providing for minimum noise. In the tradition of my previous pentode efforts, and as the result of much experimenting, this supply also provides the required (by me) regulated screen voltage for the D3a's. The bias supply is double RC filtered at very low current, so it is noise free. Solid state rectification for the 2A3 section is a pair of 1N4007's which are bypassed with anti-noise caps. The input stage gets a low-Z bridge type as does the bias section. While I like the "look and feel" of tube rectification, my bottom line is that SS rectification sounds better, with better definition especially in the lower registers. I'm not against using tube rectifiers, I just prefer SS.

Construction uses my newer 4-fold chassis bend. This gives a lower "lip" along the front and rear panels providing a better looking finish and stronger flex resistance. I use my standard clear-anodized .080 aluminum with oak interior side supports. I will only use oak for side supports from now on except for very small chassis. This particular chassis also has an oak support strip placed so that it could be used for mounting parts as well as to add structural strength preventing chassis flex when inserting the 2A3 near that area. The chassis fits neatly inside a standard 17" space.

Thanks for looking.
Schematic: ask me.
Your comments are welcome.
Critisms gracefully accepted.
Newbie questions courteously answered.
Email to: abraxasaudio at att dot net (this avoids the address scanners)
May I build one for YOU?
jef