I finished building the power supply unit for my 4P1L parallel single-ended amplifier.

The wooden cabinet enclosure is exactly the same size like the amp one for a pleasant look.. Also the top plate is identical size ( 460x324mm) and same gold anodized 3mm aluminum from Proto Panel. However the PSU is much more heavier at 56 lbs compare to the amp. The inside available height is 120 mm allowing me to fit those huge Epcos polypropylene in oil motor run capacitors.

I like symmetry for visual balance, and I used this strategy to design this PSU. Along with symmetry, there several things that one should consider when building a power supply for a tube amp such as

• uniform weight distribution for easy handling and reduce stress on the enclosure
•  optimal component placement to minimize connection wire length, reduce EM influence
• correct wiring in order to minimize EM radiation, create a perfect path for ground signal and earth/chassis ground

The PSU uses a couple of TV damping tubes 6CJ3  ( one of the biggest challenge was to procure sockets for these tubes) to rectify the high voltage that are placed close to the main transformer and protected on the sides by the 4 Hammond 8H/150mA chokes. The main transformer is an Edcor XPWR019 (540V ct/450mA, 6.3V/4A and 15V/600mA) that was the best price per quality that I could find. The HV filtration calculated using PSU2 delivers about 290Vdc at  0.29mVpp ripple.

The HV designed has common rectifiers and first cap and separate LCLC for each channel. The low voltage raw supplies are separate per channel and stage. I am using Hammond 187F24 (sec 24V/4A) for final stage filament supply and Hammond 187D12 (sec 12.6 V/2.5A) for driver stage filament supply. All transformers and choke in this PSU are running slightly warm (showing the PSU has very good power reserve and output is little influenced by larger current swings of the amplifier). PSU is using 50 watts when not connected to the amp and about 270 watts when supplying to the amp. This is showing that the overall efficiency of the amp( delivering 2×3 Watts) is very low at 2.2%. However, the sound experience is overshadowing any concerns about efficiency. ( It is paradox to me that my daily job is in the energy efficiency sector, but I am totally messing it up with my amps ;))) ).

The raw supply for filaments delivers power to Rod Coleman’s regulators and are designed accordingly. (CRC common mode 10mF/1.5ohms)

I am using two 7 pins military grade connectors procured from Ebay/China origin. Each one connects Earth ground and three pairs (+/- HV, +/- Filament Final, +/- Filament Driver) to the corresponding channel.

Here are some pictures:

I am doing final tests and measurements on my amp. I will follow up soon..

Here is the schematics:

I chose Edcor main transformer XPWR019 as the closest option to what the design required. This is the reason I had to adjust the voltage by adding two power wirewound resistors before the rectifiers. The low voltage filament raw supplies shown in the schematics are for one channell. One must double that portion of the schematics. All transformers were selected with a margin of power reserve, therefore the PSU unit is running cold ( slightly warm) except for the rectifier tubes. Wiring the ground must be done with great care. Schematic shows a idea of running the ground earth and signal.

This PSU provides very low ripple. Amplifier is dead silent. I added a 1/4″ headphone jack and I tested it with Sennheiser HD595 headphones ( they have 104 dB sensitivity) and no hum can be heard.

Best,

Radu