GoodSoundClub - Romy the Cat's Audio Site

In the Forum: Melquiades Amplifier
In the Thread: Single-stage Melquiades vs. DHT amps

Post Subject: Tu-be or not Tu-be, seems Tu-be the question.Posted by floobydust on: 2/8/2009

I've used both tube and solid-state rectification over the years and much is well documented for power supplies, i.e., old ARRL manuals, Tube manuals, etc.. In most cases small value filter caps were shown, as you really couldn't get much in larger values beyond 40-50uf with high voltage ratings and they were quite large... and electrolytic. As you start to examine older electrolytic capacitors they were somewhat odd and putting too much capacitance on the rectifier tube would result in some (internal) fireworks display. So recommended cap values were kept small.

Enter the age of solid-state rectifiers, high PIV and current ratings and many (me included) went in this direction. In my view, the real advantage was the introduction of better quality and high volumetric efficiency capacitors like Sprague DX and Mallory CGS computer-grade units. These made it possible to provide large reservoir low-impedance supplies. If you use good quality (electrolytic) caps, then you shouldn't need to parallel additional small-value caps to compensate for alleged poor performance. Multiple paralleled caps of disparate types typically result is some form of "tank circuit". If you hear audible differences by adding a small value cap in parallel with your large value filter caps, I would suspect you have the wrong caps in the first place or you're hearing the result "tank circuit" interactions.

In my recent Class-A SET designs, I've returned to a vacuum-tube rectifier with a classic pi-filter. Reasoning behind this is three-fold: 1st- with a small 2-4 watt Class-A SE mono amplifier, you do not have large changing current demands on the supply, 2nd- there are now some really excellent large value/high voltage film capacitors available and 3rd- A properly chosen rectifier tube will provide adequate delay to allow the filament in a DHT to heat up and stabilize before the B+ ramps up. As previously noted, designing such a supply can be a challenge, but if managed properly, you shouldn't hear any sonic change by swapping (good quality) rectifier tubes. If you do, I would suspect that you either have (or had) a bad rectifier and/or the power supply design does not isolate the voltage ripple properly and/or output impedance (relative to load) is too high making it sensitive to component changes.

My SET 45 amp uses power supply iron (power transformer and filter choke) that is rated 300% higher than what the circuit requires, this results in cool running, good regulation and a clean/quiet output. The rectifier tube is also rated for a typical operating current 300% more than what the circuit draws. Of course the proof is in the operation. Using a 5V4G rectifier, a 15 Henry choke and a pair of 40uF/630V Axon film caps and less than 40ma total current draw, scoping the B+ output at normal operating conditions reveals virtually no ripple and/or voltage fluctuation. One side note.... the initial prototype used a pair of 100uF/450V CGS caps. No issues with the 5V4G but I noticed random voltage spiking on the scope trace... switching to the Axon film caps eliminated this. I suspect the random voltage spiking to be the result of chemical processes within the electrolytic cap unless someone else has a better reason.

Regards, KM

Rerurn to Romy the Cat's Site