Questions about Decoupling Values and Tone

[makinrose]

I have couple questions related to decoupling; specifically as shown in the TUT projects.  So here goes:

1.  Affect on tone?    Generally in those projects a 10K/22 uf combination is used per gain stage. It makes sense from the perceptive that most of the projects use 22 uf 450V caps and it makes it easy on someone building the projects.   However this seems to be a lot more filtering than what is seen in lots of vintage amps (which may be good thing in many cases) .  Does this tend to make the amp sound stiffer?  Is it advisable to lower the value of the cap if we are trying maintain a feel similar to certain vintage designs like tweed Fenders?  

2. Resistor and Cap Values? TUT 1 mentioned that the R and C values for decoupling can be selected pretty liberally but that with lower cap values the value of the resistor needs to increase to provide good isolation.  Is there rule of thumb to determine these values for good isolation?  How low of a cap value is advisable?  I can see where using 10 uf or 4.7uf caps would save lots of space especially with plastic caps.

Thanks!!!

[K O'Connor]

Hi Guys

Guitar and bass amps are almost universally under-filtered, making this an easy part of the circuit to upgrade. Where a certain noise level was acceptable in the infant days of electric guitars and related amplification AND caps were expensive then, things have changed in our modern era where caps are inexpensive AND we have options to use equipment that is very quiet; overall, our modern expectation of "quiet" is different.

Decoupling is done in various ways and has myriad benefits, as our "The Ultimate Tone" (TUT) series describes. The projects in TUT3 and TUT5 use a combination of high-node filtering and proportional filtering. It is proportional up to a master preamp node (Vp) from which power fans out to the individual preamp stages. This gives reasonable filtering while allowing each stage to have its own voltage adjustment.Because there are fewer stages of filtering overall, it is highly beneficial to use larger cap values.

Overall, it is easier to use a simple proportional supply with filter nodes for every stage. The only potential complication is where there are multiple signal paths.

Preamps draw constant current, or so nearly constant that sag is a non-issue. using 10uF filters is nonproblematic and lets one use physically small caps of 10mm diameter typically. As a proportional filter this works very well.

The isolation/dropping resistors are usually selected purely for the voltage drop desired and the RC constant falls where it may. For example, 1k + 10uF is a 10ms TC corresponding to a 3dB frequency response of 16Hz. This means that frequencies below 16Hz could interact much more greatly than higher frequencies do, with a possibility of motor-boating if the layout is poor. However, with a Galactic Ground layout there is no concern about this.

Note that in some specific gain stages or circuits, it is important to have a supply filter response that is much lower than the lowest signal frequency expected through the circuit.

In most OEM amplifiers, voltages tend to be rather high in the preamp stages and filtering is quite good. The charge stored is proportional to the voltage and the capacitance, so increasing one or the other will maintain a high energy storage well above what the stage needs to meet signal swing requirements. Fender often has a single 20uF cap supporting several stages where Marshall might have 50uF supporting a single stage or a couple of stages. If it were possible to isolate all other differences and hear what the filter cap influence is alone, which would require a half-ways correction of some layout problems, the higher cap values should provide better suppression of intermodulation artifacts and overall sound nicer.

You can "get away with" all kinds of decisions with respect to R and C values in the supply, along with layout choices, but if you tray too far from the ideal methods the circuit will be unstable or noisy at high-signal levels and if truly awful choices are made, at idle too.

[makinrose]

Kevin thanks for that informative reply. I hope I'm understanding it clearly. Here's few points I want make sure I'm interpreting correctly:

1. From what you wrote that its' difficult to tease apart the tonal effects of tying multiple pre-amp stages to single node partly because of how most amps are built and their poor grounding. Therefore direct comparison to the approach to filtering in the TUTs is difficult.

2. Better pre-amp filtering tends to sound better since it reduces inter-modulation. So London Power's approach to filtering (along with Galactic Grounding) every stage obviously solves all these inter-modulation woes and will sound better!

3. You get away with lower cap values if we use a proportional filtering approach ala the Standard in TUT5 but higher values are helpful if use the node approach like in the pre-am TUT3 Bassman and others?

Again thanks for your help. I can truly say Galactic Grounding and so much of the TUT info has been invaluable to me.

[K O'Connor]

Hi Guys

Yes, that's all correct.

What some hobbyists and builders find once they correct the grounding is that the amp "loses something". What it has lost is all that IM distortion that clutters the sound and contributes to some of the "fullness" for some kinds of playing - mostly for clean sounds, although one could imagine that hash fitting in better with distorted tones? It is subjective how you interpret the result.

For the most part, eliminating IM greatly improves note articulation and you can hear small details of the sound. This either reveals nuances of your playing that you could not appreciate before, or it shows up adaptations you have made to the previous cluttered tone, or both. It is the same with straight harmonic distortion where a small amount of even harmonics adds warmth but too much is muddy, a small amount of odd harmonics is crisp and too much is harsh. At least with the harmonic distortion the frequencies are musically related to the notes; IM is unrelated frequencies and therefore much nastier.

[makinrose]

Thanks for clarifying that for me. As you know it's difficult to get good engineering advice/info about tube amps. Some of trouble is that tone is so subjective but most of it is that many builders apply a lot of non-scientific thinking to them.

I figure that if I hear something is "lost" by getting rid of IM I ought to make changes elsewhere. It's better design in the tone into the amp than hope for it.

[K O'Connor]

Hi Guys

I choose components for reliability and circuits for tone. Galactic Grounding is a given, so IM will be very low and if I need to, I'll sub a ceramic cap some where that might otherwise have a plastic cap just to add some grit. The same applies when you swap out electrolytic filter caps and now have a cleaner PSU - you lose the "electrolytic distortion" and might have to compensate in some other way.

Overall, I stay true to tone NOT to circuits.

The tone-reference approach sometimes means I might have to add an entire tube to get the warmth I want, or use wildly different part values than the OEM tone reference. But.. that just makes it a more creative process.

How much noise and nuance you hear depends a lot on the speaker, which is a whole other discussion...