Question About Capacitor Upgrade in Tube Amp

It's more complicated than just hanging a film cap across B+. If you assume the designer knew what he was doing, had good ears and related equipment, you could make things a whole lot worse by making willy-nilly changes.

Too often audiophools make changes that make ZERO difference but think they have due to everyday occurrences like fatigue, humidity, temperature, atmospheric pressure, line voltage / noise, etc. etc. etc. 

I can definitely hear a difference in the clarity of vocals and the separation of instruments between the X7 vs the X9 and X10. So let’s see how this upgrade works out once I get the parts.

There’s a lot more going on than just the caps. Power, impedance, source, cables, loudspeaker, tube age, wire, sockets, etc. etc. Change one thing, change everything.

To those who don’t like the tone:

Are you fine with people being seen off by charlatans that give bogus reasons with ZERO supporting evidence other than their say so? They think their system sounds the same every day or that they can turn it off for some time, make a change and then determine a delta. Ha! Qualified engineers listen for weeks, making comparisons between a reference unit and the changed. Confirmation bias plays a huge part in an awful lot of opinions.

I’ve been at the HiFi game for more than 50 years, Grammy nominated engineer and if you saw a movie from about 1990 to 2010, quite possibly the soundtrack was recorded through something I designed.

Some posters blather on about changes without the slightest regard for rigor. To boot, they evaluate whilst intoxicated and at injurious levels. I endured too many sessions with doped up musicians to have any tolerance for audiophools.

PAX.

The elephant in the room that no one notices is Power Supply Impedance and Phase. It is perhaps the most important and neglected aspect of HiFi design.

Capacitors have two frequency dependent properties, ESR and ESL, often ignored by engineers and almost certainly by Joe Audiophool.

The amplifier circuit, if perfect, does nothing more than regulate the power supply. If the power supply has poor phase and impedance, it passed on to the next component. Sequential components alter the preceding and this is why a really great component may get so-so response in one reviewer's system and sterling accolades in another.

Sequential components & cables all add their sonic signature.

Hence, when Walter Wannabe audiophool swaps out ReallyGoodMatch for  BigFatSilver caps, he is altering his entire system and any perceived change, improvement or not is only valid for his system.

Add in the $$$$ and time incentive, and of course it's better.

As for capacitor used to shunt DC power supply, they are necessary to eliminate high-frequency noise present there, but never affect the sound quality as much as a different coupling (serial) capacitor used in the audio path.

@eworkflow Are you an engineer? Do you have a test lab? Verified the bypass capacitor values actually sink just the noise and haven't changed the DUT frequency / phase response?

'Noise' is not a constant and changing capacitor value / type may actually make things worse.

Changing type / value changes the power supply impedance and phase. Whether or not it is an improvement may be dependent on the DUT, the rest of the system AND the CBLF*.

Making willy-nilly recommendations is irresponsible.

*: the listener with all its inconsistencies.

there was an improvement in frequency response and lower distortion replacing the cheap factory caps with one of better and different material construction.

Did you measure the C or look up the ESR and ESL of the caps or just read the value printed on the schematics / old cap?

Did you measure the frequency response, distortion, phase shift, transient response, etc. or just listen?

@calieng 

Does anyone know how the power supply bypass capacitor effects high end frequency response?

Capacitors have properties Capacitance, Equivalent Series Resistance [ERS] and Equivalent Series Inductance [ESL], all of which are frequency dependent. Dielectrics, foils, leads and geometry affect the frequency response of these properties. At some frequency the combination of these properties will make the system self-resonant. Hence a band limited digital system may exhibit different response when fed by a disc system with much higher bandwidth.

Most think the capacitors just filter the DC from the rectifiers. However, current flows in @ 60/120Hz [50/100Hz] and out of the capacitor bank @ DC to many kHz.

Add in additional effects of construction, circuit design, wiring, printed or otherwise, program, level, related equipment, environmental conditions and universality approaches ZERO.

The expectation that the sonic effect in one device will translate to all others is pure fantasy.

Upgrading capacitors is a no-brainer. Ditto resistors, wire, tubes, sockets, transformers, connectors, layout, etc. 

Expecting 'warmer mids' or 'silky highs' when installing Caps-4-Bux or any other component change in myriad divers applications is hubris.

The lack of rigor in accommodating environmental and CBLF changes negate generalized comments.

I would also suggest that Dielectric Absorption (DA) is the most important measure for modding, because the other measures should have been considered in the design phase.

Nope. DA is not the most important by a long shot.

Of properties mentioned this far, left out are:

• Geometry, rolled or folded
• Lead material, length and attachment methods
• Foil material
• Foil inner / outer connection direction for DC blocking
• Microphonics
• Age

The properties you list are critical when using electrolytics, which are not (should not) be under consideration here.

OH, yes they should. Trying to clean up a marginal supply with a magic bullet is silly and rarely effective. A better supply can transform an amplifier.

Capacitor technology has improved immensely in recent years. More capacitance in smaller packages with reduction in ESR, ESL.

But my reading of the OP is that the question concerns coupling and bypass caps. Are you seriously suggesting bypassing with electrolytics?

BEFORE worrying about bypass or blocking caps, ALWAYS ensure the electrolytic PSU filter caps are up to snuff. Way too many tube amps have inadequate primary filter capacitance. Dropping the 120Hz by 10dB is far preferable to 100kHz by a couple.

Too often they are under voltage, too high ESR, too small, etc.

Prove it.

HF bypass caps are common in IC circuits as opamps can have multi-MHz bandwidths coupled with 100dB gain.

Tubes can also be multi MHz oscillators, but willy-nilly hanging a film cap across the B+ filter maybe just the ticket to convert an amplifier into an oscillator.

Tubes are capable of several MHz operation.

Oscillators are made of LCR and a device to drive them. The combination of the PSU, wiring - printed or P2P, transformer LCR and a tube to drive them can create a self-resonant circuit.

If one searches the DIY sites there are innumerable posts of capacitor changes turning tube amplifiers into oscillators.