Recently I finished the filament supply for the latest incarnation of my 4P1L pre-amplifier. Here is the next instalment of this project. The HT power supply was refined after builiding more than 7 stacked HT supplies for the 814 SE Amplifier.
The supply design is very simple. Perhaps the selection of components and the refinement of some aspects of it is what makes the difference to me:
The power transformer has multiple taps to adjust the secondary voltage by 50V steps. It also has a copper shielding between primary and secondary and an outer shielding as well. R1 and R2 provides the additional resistance to the LC filtering tank to avoid overshooting and a smoother response to the load demands. You can use PSUD2 to fine-tune their value. You also need to take into account the anode resistance and the secondary copper resistance as they also add up to the total R value in the LCR circuit. R5 and C5 are the secondary snubber. The rectifier is hybrid and it uses a pair of 6AU4GT damper diodes and a pair of SiC C3D02060F 600V/2A diodes. Clean switching and smooth voltage raise given by the damper tubes make this hybrid bridge ideal. I rather use this thank 4 damper valves, but you can if you want to. the heaters are referenced to +B by connecting one leg of the filament windings to +B.
The choke is a 20H/100mA LL1673 from Lundahl. You can get them at a great price in the UK from Big Bear Audio. I used a nice pair of ANSAR film 40μF/400V capacitors. The have a very low ESR measured to be around 7mΩ. You can use any other Oil Motor Run capacitor (e.g. ASC) or your prefered polypropylene capacitor. C6 and C7 are a pair of ceramic capacitors which they are intended to form the choke snubber (see Morgan Jones’ recommendations on his book). R3 is omitted in this case but I left it there in the diagram as you can add additional resistance to drop output voltage if needed and as a secondary benefit to improve the filtering stage. R4 is just a bleeder resistor, you need them!
I’m actually loading this supply with about 70mA. The shunt regulator is an SSHV2 which feeds the 4P1L pair. Each one of them are running at 30-32mA. Here is the first noise test of the raw supply output:
The total inband noise level is about 39mVrms. The 100Hz level is just 26mV, where the 50Hz is mainly the interference of my workbench. My workbench noise floor is 170μV.
Now, if we look at the output of the SSHV2 when tested with a CCS load set at 62mA we can see the additional benefit of the SSHV2 regulator:
The 100Hz noise is now down to 1.48mV which is an additional 24dB. SSHV2 rejection should be more than 24dB but it seems to be below my workbench noise and can’t measure the improvement 😉
Well, unless I add a 60dB differential preamplifier to my noise measurement jig! Let’s see what the improvement is:
Inband noise level is now down to 691μV. The SSHV has contributed a 35dB improvement. The rejection of 100Hz noise is about 39.2dB. Quite good, but cant improve this as my workbench noise floor with the differential preamplifier is improve down to 198nV (input short-circuit) but when measuring noise with HT supply powered off, noise is as high as 560μV. Just remember that the SSHV2 is working nearly at full tilt with 75mA across the CCS!
Just close to finish the preamp now!
Ale
Hi Ale,
Nice, detailed analysis. I wish I can get something similar with Weiss in the US.
Best,
Radu
Thanks for showing us what you are doing. I am anxiously following this project. Any chance of a schematic for what you are doing with the filament’s supply?
Take care,
Do you sell this stuff?
Hi Ale,
Did you tried 5U4 rectifier tube against 6AU4GT.