Why tube rectification?


This question is directed at the distinguished members of the forum who design and build tube amps or those who have knowledge of tube amp design. All the tube amps I own/have built us two diodes for rectification. Diodes are cheap, compact and last the life of the amplifier in most cases. Examples include the Dynaco ST-35, the Decware Zenkit1 (which is basically a Decware SE84) and the Elekit TU-8900. All reasonably well respected amps. Yet many of the more expensive amps go with tube rectification, which obviously involves the downside of another tube, more power, more space.

These two competing solutions both supply the basic power to the audio tubes and output transformers, so only indirectly interact with the sound signal. I have not read anything that explains what tube rectification brings to the party. But it must have some upside to offset it's obvious downsides. If I changed over one of the above amps to tube from diode rectification what would I be likely to  hear?

Ag insider logo xs@2xbruce19

I will add something that has not yet been mentioned explicitly.
 

What you are listening to from an amp/pre-amp is the power supply - not the signal being sent to it.

 

It is a bit of a misnomer that we call these things amplifiers. It implies that the signal itself is being amplified, and the other stuff is just there to help.

What is in fact happening is that the voltage from the power supply is being modulated by the input signal, and the input signal itself is swallowed by the tube/transistor. So what happens is the voltage from the power supply is either directed to ground or to the speaker, and the input signal simply regulates how much of that goes where.

 

This is why power supplies, especially in preamps and DACs, are so critical. You are literally listening to the sound of the rectifier after is has been smoothed out by some capacitors.

 

That is a good observation @joshua43214, yet, while I don’t doubt those who have said that changing rectifier tubes changes the sound, I do struggle a bit to understand why. After all if they are doing their job they should just be truning AC into inperfect DC, then the following caps, resistors and maybe inductors all smooth the DC. The real important thing about a power supply as I understand it is that it supplies clean and plentiful power. The audio signal doesn’t interact at all until the gainstage or preamp tubes and then again in the power tubes. So that is where desireable tube distortion or "tubiness" enters the picture. Now it is plausible that Rectifier tubes may introduce their own flavor of distortion into the power flow and that carries though to the final output. I can buy that and also accept that it might be desireable.

My own thinking on hifi these days is that music performance is to music reproduction as live viewing of a landscape is to an image of the landscape. Different expectations apply in each case and in the latter two intentional modification for effect is entirely permissible. That is, tone controls on an amp are just as permissible as impressionistic painting or jiggering the exposure of photos.

These two competing solutions both supply the basic power to the audio tubes and output transformers, so only indirectly interact with the sound signal. I have not read anything that explains what tube rectification brings to the party.

Voltage drop across the rectifier, especially when the amp is making power, is one very large reason rectifiers can affect the sound. The other reason is noise.

Solid state rectifiers are often blamed for making noise but its really how the power transformer interacts with the rectifiers. Between the two an electrical resonance can occur which is set into oscillation by the rectifiers turning on and off. The resonance might be at 2MHz but can cause the rectifiers and transformer to make noises much lower in frequency, which can leak into the signal chain and cause IMD. This is known as a 'swept resonance'. Its not something you often run into with a tube rectifier.

But it can be solved with proper application of snubber networks. The best approach is a small capacitance in series with a resistor across the input to the rectifier(s), directly across the output of the transformer.

Once the swept resonance is shut down there's no going back to tube rectifiers as the power supply sags less at higher power levels- which makes for less distortion at those power levels.

An obvious advantage of a solid state rectifier is the capacitances in the power supply can be quite a lot higher. There is often a fairly low limit to capacitor values when tubes are used. There are timing constants involved with these values. What you want is that none of the timing constants in the amplifier circuit be lower than those in the power supply- else the amp can much more easily modulate the power supply, resulting in higher distortion.