Published on: Oct 11, 2011. Last updated: July 27, 2018.

Introduction

Ever since building the 2W single ended tube amplifier, I wanted to build something better, something that is considered a good amplifier among tube amplifier enthusiast.

It is in no way cheap to build a tube amplifier, so this project got a good kick start a day where I got a old broken 120W bass amplifier for free, the power transformer was burned out and it had been left in a barn for about 15 years.

Safety

WARNING!: Working with electricity is dangerous, all information found on my site is for educational purpose and I accept no responsibility for others actions using the information found on this site.

Read this document about safety! http://www.pupman.com/safety.htm

Considerations

When searching for a amplifier design to follow, I came across a good paper written by Claus Byrith where he took a well known Mullard design and gave it a proper discussion. Through this discussion  he came out with a improved design with valid arguments to his choices. You find the instructions and original schematics here: https://www.lundahltransformers.com/old-fasioned-30w-push-pull-amplifier/

His design was recommended by many on www.diyaudio.com and I chose to go ahead with his paper as it was well documented compared to other older designs that mostly consist of a schematic and what forum threads you can find throughout the internet.

Specifications

Schematic

Monostage

Power supply

EM800 VU meter

Construction

4th January 2009
Obtained and disassembled a old Sound City 120 guitar amplfier, parts scavenged 6x El34 tubes, 5x ECC83 tubes, sockets, 120W mono output transformer, various jacks and potentiometers.

27th March 2009
Traded the 120W output transformer for 2x 30W ultra liniar output transformers.

5th May 2009
Ordered a custom power transformer.

11th June 2009
Received the custom made power transformer along with covers, and 2 EM800 ( http://www.akh.se/tubes/htm/em800.htm ) magiceye tubes for free! These will look neat used as VU meters!

16th June 2009
Designed power supply PCB.

17th June 2009
Designed EM800 VU meter PCB.

22nd June 2009
Etched, assembled and tested EM800 VU meter.

24th June 2009
Etched and assembled power supply.

1st July 2009
Designed mono stage PCB.

7th July 2009
Tested power supply, bias voltage is -150VDC, tweaking is needed.
Etched and assembled first mono stage PCB.

11th October 2009
Plan for the baseplate is made

12th October 2009
Construction of the baseplate begins

30th October 2009
Construction of baseplate is done

4th January 2010
Polishing and varnishing of baseplate

5th January 2010
Assembly of baseplate with transformers and tube sockets.

14th January 2010
Etched second mono stage PCB

20th January 2010
assembled second mono stage PCB

7th March 2010
Assembly on baseplate begins

16th March 2010
Assembly progress

20th April 2010
Assembly progress

25th April 2010
Assembly progress. the amount of different colour wires and insuring a proper twisting made the process take much longer than expected. But in order to get a good result, one needs to invest the necessary work in making it so.

26th April 2010
First power on, test and adjusting. There are problems with the negative feedback.

12th August 2011
I found a wooden box for the amplifier to be installed in.

10th August 2011
It is a problem having two mono stages share a power supply, negative feedback voltage needs to be adjusted for each channel. Matching of some grid resistors necessary.

17th August 2011
The amplifier runs and can play audio, but there is noise problems from ground loops.

I got the ground loops sorted out in such a manner that you have to put your ear close to the speaker in order to hear the mains hum. All ground wires goes to a star point at the power supply and all ground wires that could, is twisted or braided together. When I install the amplifier in its final enclosure, I will take additional measures to wire the ground better.

24th August 2011
The amplifier is complete on the baseplate, everything and in the right sizes are soldered on and only the enclosure and mounting of jacks remain the last to do. Been listening to the amplifier all day on the Isophon BS35 speaker set. wonderful.

13th February 2012
The best piece of wood for roof constructions that I could find at the home improvement store was dragged home, I needed some good thick and tall wood in order to make a cut with a router for the amplifier base plate to slide into.

9th February 2013
Ever since I built the amplifier, there was an issue with a small amount of 100 Hz hum from the power supply. Realizing that I did not double the capacitance in the power supply, as described if it was to be used for a stereo amplifier, I wanted to correct that mistake. With only two 220 uF capacitance on the 450 VDC rail I would see 2,3% ripple. With the recommended four times 220 uF there would be 1,05% ripple. As I had spare 1000 uF capacitors I used those to bring it further down to a mere 0,4% ripple.

While I already had the amplifier on its back I also decided to add some other modifications that I felt was necessary to give it more years to live in. A soft start circuit for the filaments and a delay on the high voltage. These steps are taken to avoid that the cold filaments with their very low cold resistance would take damage from the magnetic forces when power is applied without soft start, this will eventually break the filament as it also glows up bright yellow at first. Delaying the high voltage will prevent the tube trying to conduct from a filament that is not properly heated, this could lead to the thoriated tungsten filament losing its ability of giving off electrons as the contamination is ruined. The soft start of the filaments is done through a 1 Ohm 100 Watt resistor and after one minute the filaments get full voltage and the high voltage is also applied. All done with just one relay and one timer. A NTC resistor was also added to the primary side of the power transformer in order to soft start the five times bigger capacitance in the power supply, this should remove a great deal of stress on the rectifier diodes.

The EM800 indicator tubes are also installed and their circuits connected to the amplifier. To obtain galvanic isolation between the EM800 circuit and the amplifier I used a 1:30 current transformer on the output lead to the speakers. With a proper ferrite ring core for high frequencies the indicator tubes now show the level of the treble output.

21st December 2013

Through some time I had noticed lower sound level in the right channel and found some time to turn the amplifier around and measure what was going on.

The left channel was almost in balance at 205mV and 195mV over the cathode 10R resistors. The right channel was however at 235mV and 145mV. The bias level was also on the low side from when I first adjusted it very conservatively and was for both channels set to 350mV with balance between push-pull tubes measured to 0,005V.

After adjusting the weak tube in the right channel began red plating and the plate took enough damage/discolouring for me to change it for another from the big box of tubes. It was most likely just burned out as they are all old used tubes.

Conclusion

Mono stage
Pros: Close to the speaker, easy adjustment of power supply, no critical matching with the other channel.
Cons: Need separate power supply which will increase total cost.

Stereo stage
Pros: Everything in one box, lower total cost, everything is equally coupled regarding heat and noise.
Cons: Critical matching of stages and power supply, ground loops appear easier.

It might seem that there are most cons that pros, but the fact is that the cons are much harder to deal with. For the future I would try to build mono stages to gain more experience in choosing between the two. Looking back at the process, matching the two stages to each other and the power supply took much of my time.

I have tried my Isophon BS35 speaker set on my regular amplifier for my stereo speaker set for my computer, they are miles ahead in sound quality over the stock speakers that came with this semi Hi-Fi set, Edifier S2000. The largest difference from the transistor amplifier to the EL34 tube amplifier is the bass. Its deep as the internal hell, clear like tuned piano and makes a world of difference to enjoying music. It have to be experienced as a description is far from enough compared to feeling it in your chest!

Cathode current settings are a bit on the low side, but for now they perform good enough for playing up my apartment. Setting AC balance with a distortion meter is also on the list of future improvements.

I still need a proper pre amplifier that will have the ability to deal with input impedance matching, ground potentials and noise. I will properly build this myself, but for now I just use my laptop as source and pre amplifier.

Demonstration