Create the best low power tube amplifier?

Building the best sounding amplifier is always a life long mission for audiophiles in the world. There are some that chase sky high output power, lowest THD and such. And, there are some that seek purity and simplicity.

This project is all about purity as the ultimate goal. We are designing a single stage single ended vacuum single tube per channel amplifier using E55L (8233) pentode. It can be configured in pentode (ultra-linear) and triode mode. Although there are plenty of other tubes to choose, this looks to be one of the best choice we can have based on price and availability. You don't want us to build a Western Electric WE437 amplifier, do you?

There will be ZERO capacitor in the signal path. Capacitors are only employed in the power supply. In order to achieve the highest performance in terms of audio quality, and with this kind of simplicity and component count, the parts selected shall need to be of the highest quality.

For example: amorphous core silver input & output transformer, Z11 core power transformer, amorphous core power supply inductor, silver wires, teflon tube socket, and etc. All these will cost us a lot of money. This will not happen if we do not get your support!!!

So, come and join us to make this the best sounding low power tube amplifier ever in history! Aren't you excited about this? Do you want to witness the making and get to have a share on the project? We welcome all you inputs and support to make this a reality, and make it a success! If support is good, perhaps we could even make a kit out of it and all can enjoy the benefits of volume pricing and have all the hard work done by us. We win, you win, everybody wins!

P/S: Due to the simplicity and purist design, it will not have ample of power. Therefore, care must be taken on matching of the speakers to be used with this amplifier to get the best combined performance. High efficiency speakers are plenty nowadays fortunately! Best regards, Ken

Discussion (3 comments)

xeno 11 years ago

Check out this one.

Although with semiconductor rectification.

Have been listening to this amp for a while now and I am happy with the result.

e55l2.png (12kb)

1 Attachment(s) 0 Comment(s)

Ken 11 years ago

e55l-spud-023.png (23kb)

Robaroni 11 years ago

Hi Ken, It seems to me that if we are going to design a low power amp then using a a 5U4 is detrimental to that end. Do tube rectifiers really add to the sound? Are they worth the extra power they add?

Ken 12 years ago

Hello Sqrt.LC, 1. Good catch. I should have indicate that the voltage developed there will be used to lift the filament voltage of E55L to about 70V to reduce noise coupled from the filament to the amplifier. My bad. This has proven to be very effective in eliminating hum/buzz. 2. LED bias would be one of the bias methods I would like to experiment and measure (with the help from Elektor on the measurements gears). :-) When the project is realized, I would be testing a few bias methods and correlate the measurements with respect to the sound. Hopefully the data will show and we will use the best out of the ones listed below: a) fixed bias with voltage source at the grid b) constant current sink at cathode c) LED bias at cathode d) cathode bias with resistor / capacitor e) battery bias Thanks a lot for your feedback. Cheers Ken

Sqrt.LC 12 years ago

Please your attention for two questions: 1. Do you think that C3 (10uF) adds sufficiently to the quality of the sound to justify its presence in the schematic, or has it another function? 2. The LED impedance [in the cathode circuit] is highly non-linear and even adds to the non-linearity of the valve, are there any measures to counterbalance the resulting distortion?

1 Attachment(s) 3 Comment(s)

Ken 11 years ago

E55L Introduction

Before we dive into the circuit design, let's look into why we choose tube E55L as the main active component. Here's the detail of the tube:

Why we choose E55L?

Linear and wide bandwidth.
One tube does it all as an integrated amplifier: CD output -> Input -> E55L single stage of active component -> Output Transformer -> Speaker.
Gain (amplification factor = 30) is suffcient, with headroom of 6Vpp at input to obtain full power.
No couplig caps or interstage transformer needed.
Low plate resistance of 600 Ohms.
Decent power of 2W instead of micro power simple tube amplifiers that is <1W (1626, 5842, 5687).
Inexpensive tube, plenty of supply, plenty of excellent brands to choose from - Amperex, Mullard, & Philips; made in USA or UK.
SQ - special quality tube, tight tolerances of electrical ratings, shock and vibration resistant, gold-pin to give low contact resistance, corrosion and oxidation, stability of characteristics and long life - 10,000 hours. One can even boost it for higher power under reduced life conditions.
Produced in the 1960's, the older the better theory applies.
Alternative to ultra rare & expensive EC8020 or WE 437.

Most importantly, it sounds great!!!

Design Preview

As what Nelson Pass always preach, design over component quaity. We shall start with the design. Here's the teaser:

Amplification section:

Input transformer / TVC -> E55L -> output transformer
No capacitor in signal path
Input transformer / TVC can be removed if system gain is sufficient (2Vrms source output & high efficiency speakers)

Power supply

Tube rectified
Regulated or non-regulated (your say!)
Choke input
Semi-dual-mono

Your feedback is wanted on achieving the best design + sound!

Regulated vs non regulated PSU
Active or passive loading on cathode bias
Fixed or cathode bias
No / with input transformer
No / with TVC (transformer volume control)
Tube rectifier choice
PSU capacitor choice
Transformer material / brand / type

Till then!

Best rgds, Ken

e55l-amplifier.jpg (52kb)

1 Attachment(s) 0 Comment(s)

Embed Code

Mini-Wheelie Self-Balancing Robot I'm posting my code here. This is just about working although left and right doesn't do much. S
I'm posting my code here. This is just about worki...

9 hours ago

Mini-Wheelie Self-Balancing Robot Can the author of the code please help. I'm trying to understand the calls to the LMotorControl
Can the author of the code please help. I'm trying...

10 hours ago