Capacitor voltage effects a tone circuit pot sweep?

[ssstonelover]

I came across the contention today (online) that if we compare two capacitors of equal value (say .022mfd) and one has a voltage rating of 400V, and the other has a voltage rating of 1.5V (as an example), they perform very differently when you roll the tone pot from 10 to 1. The contention is the small cap is too efficient and renders the pot hopeless at controlling the treble cut, whereas a large cap will be easier to control through the pot sweep.

I have had trouble precisely controlling treble cut (using standard Alpha A250K pots) so gave up on .022mfd caps a long time ago and now use 0.012mfd to eliminate the mud, yet cut treble in a more pleasing way. I use the small poly caps, never big caps....

I always through voltage ratings would not make no difference at the voltages a guitar operates at -- so just over-stuff a control cavity, for the point of looking cool and retro.

I'd be very interested to know if there is anything behind this theory in terms of cap choice.

[newey]

May 13, 2023 17:41:38 GMT -5 ssstonelover said:

I always through voltage ratings would not make no difference at the voltages a guitar operates at --

And you've always been correct in so thinking. The voltage rating of a capacitor is the maximum voltage that the manufacturer is willing to certify before the thing melts itself into a puddle. A guitar pickup puts out fractions of a volt. Whether the cap is rated for 400V or 1.5V won't make any difference.

Whenever I hear/read people saying that they have to have this or that type of cap, that they can hear a difference, my first thought is always "Double blind test or it didn't happen". Will a vintage paper-and-oil cap sound different? Sure, if it's aged to the point where it no longer has its rated capacitance.

Same here with this voltage question- if the person hasn't checked the cap with an LCR meter to be sure it is within spec with any others to which it is being compared, no valid results are possible.

[ssstonelover]

May 14, 2023 6:17:25 GMT -5 newey said:

if the person hasn't checked the cap with an LCR meter to be sure it is within spec with any others to which it is being compared, no valid results are possible.

Very likely they did not check the actual value, and yes it is some "bubblebee" repro so likely a PIO type....



If what they claimed was true, more people would have caught on after all, so I'll stay clear of their claim and control pot performance with better science instead

[kitwn]

I can forgive ignorance on any subject, part of my working life was spent doing my best to dispel it. What I find hard to tolerate is the number of people who are determined to propagate their ignorance as far and wide as possible. GN2 is a haven of reality in a morass of myth and repeated hearsay.

The type of capacitor only becomes critical if you're working at high voltage, high ripple current, high frequency, high temperature or require an unusually high accuracy and stability. The inside of your guitar ticks none of these boxes.

Kit

PS  Perfectly good, brand new guitar-suitable capacitors are available from many sources for a few cents each. Nearer a dollar if you MUST have one that looks like an orange blob. Anyone who tries to sell you 60 year old junk for $10 each should be drowned in their own snake oil.

[Yogi B]

May 13, 2023 17:41:38 GMT -5 ssstonelover said:

The contention is the small cap is too efficient

This makes me think we're (they're) talking about the capacitor's equivalent series resistance (ESR), since "efficiency" is a measure relating to energy transfer and increased series resistance would dissipate more energy (as heat, due to joule heating), thus be less efficient. This can be measured in terms of the dissipation factor (DF) as a ratio of the ESR to the magnitude of the reactance of the capacitance (at some given frequency) — if that definition seems familiar that may be because DF is the reciprocal of Q factor. Therefore, I think the logic behind the assertion is that lower efficiency caps have higher DF & ESR, lower Q, thus less treble cut (in terms of maximum attenuation, i.e. ignoring that the cutoff frequency is also affected).

The best example I've found so far relating to dissipation factor varying within capacitors of the same type with respect to their voltage rating, is that from Cornell Doubiler's 715P "Orange Drop" (polypropylene) range. I've seen these sold (for use in guitar tone controls) interchangeably with the 225P (again an Orange Drop, but polyester), but the latter does seem to be the preferred (or at least more abundant) choice. (Unfortunately the 225P datasheet does not include such a detailed breakdown). Below is reproduction of a table of dissipation factors for the 715P caps (tweaked to enable easier horizontal comparisons).

Maximum Dissipation Factor (D.F.) in %

Cap	100V-600V		800V/1000V		1200V		1600V/2000V
Range (μF)	20kHz	100kHz	20kHz	100kHz	20kHz	100kHz	20kHz	100kHz
.001 - .01	.029		.039		.035		.033
		.040		.088		.068		.059
.012 - .022	.031		.040		.043		.039
		.050		.098		.111		.093
.027 - .047	.034		.055		.047		.041
		.071		.175		.132		.102
.056 - .068	.037		.058
		.089		.193
.082 - .1	.042		.063
		.116		.220
.12 - .15	.049
		.158
.18 - .22	.059
		.217
.27 - .33	.074
		.309
.39 - 47	.093
		.427

Within this table, I think there are three things worth pointing out: first is the grouping, in that it either wasn't possible to or wasn't deemed necessary to separate caps in the 100V-600V range (and I don't think I've ever seen 800V+ Orange Drops being sold as guitar tone caps); second is that (in this case) a larger voltage rating corresponding to larger DF is only true up to a certain extent (800V/1000V is always larger than 100V-600V, but that's not true of the other columns); the final would be to look at a specific example of DF and calculate the corresponding ESR.

So using:

\begin{aligned}
\mathrm{DF} &= {\mathrm{ESR} \over \lvert X_C \rvert}
\\[3ex]
\mathrm{DF} \cdot \lvert X_C \rvert &= \mathrm{ESR}
\\[3ex]
\mathrm{DF} \cdot \left\lvert-{1 \over 2 \pi f C}\right\rvert &=
\\[3ex]
{\mathrm{DF} \over 2 \pi f C} &=
\end{aligned}
and looking at the table for a .012-.022μF (12-22nF) cap rated at 100V-600V measured at 20kHz, and generously assuming the DF value applies to the lower end of the capacitance range, the ESR is:

\begin{aligned}
\mathrm{ESR} &= {\mathrm{DF} \over 2 \pi f C}
\\
&= {.031\% \over 2 \pi \times 20 \!\cdot\! 10^3 \times 12 \!\cdot\! 10^{-9} }
\\[3ex]
&\approx 0.20558\,\mathrm{\Omega}
\end{aligned}
However taking the same cap at 100kHz, we get only:

\begin{aligned}
\mathrm{ESR} &= {.050\% \over 2 \pi \times 100 \!\cdot\! 10^3 \times 12 \!\cdot\! 10^{-9} }
\\[3ex]
&\approx 0.06631\,\mathrm{\Omega}
\end{aligned}
As the above demonstrates, unfortunately ESR isn't constant with frequency — a capacitor has more parasitics than just a series resistance — and with 20kHz being close to the upper end of audible effect that a tone control could have, I'd like a value at a lower frequency before rushing to a conclusion.

Thankfully, I found a document from Vishay (a previous manufacturer of Orange Drops), that gives typical DF measured at 1kHz, for 715P this is given as ".05%". It is only one value, which I'm taking to be valid for an arbitrary capacitance value — but looking back at the previous table, the range of DF as it varies with capacitance drops between 100kHz & 20kHz, so it's not unreasonable to assume this trend would continue down to 1kHz and give an even smaller spread of values. Plugging the numbers into the equation, again using a 12nF cap, (or coincidentally multiplying the previous result by 100) yields an ESR of about 6.63 ohms. Thinking about that in comparison to an A250k pot turned to 50% (25 kilohms) that's about one part in four thousand; much, much less than the tolerance of the pot's resistance.

There is also a final observation that can be had from that document: dissipation factor varies quite widely across caps of differing type. The 715P caps are specifically billed as low dissipation whereas 225P are not. Their typical DF being six times larger (0.3% vs. 0.05% at 1kHz), consequently makes their ESR larger by the same amount, so around 40 ohms (again for a 12nF cap at 1kHz). Recalling that I earlier mentioned the 225P was the preferred type, and larger ESR is apparently desirable, then maybe this could explain that preference. However, if that were true, then surely the often decried ceramic capacitors should be even better (as their DF is even larger, up to 5% at 1kHz for some compositions). Or better still electrolytics, which can clock in with a DF of up to around 20%.

---

May 15, 2023 22:52:52 GMT -5 kitwn said:

Anyone who tries to sell you 60 year old junk for $10 each should be drowned in their own snake oil.

Ha! ONLY $10!? You've missed a power (and a half, if we're talking AUD) of ten: (A pair of) NOS Bumble Bees.

[jandesign]

Hi everyone!

My experience started some years ago, after reading tons of threads on websites/forums. I decided to buy some 0.022mdf caps from different brands to see if I could hear and/or tell which is which. They all sounded the same, and one was a little muddier. Guess what: I measured them all and THAT one showed a different value. The tolerance plays in the game as well...

My favourite cap is now mounted on my personal guitar and it must be a "paper in oil" thing from an old 1961 italian radio. Some components from that radio weren't working anymore. My chosen cap is literally burnt from one side, but it does the job in its own way. The sound is very mellow and sweet. My explanation is that it must be leaking some frequencies while retaining others in a non-linear unexpected way.  The value is 0.011 mfd and, needless to say, nothing compared to other 0.011 mfd caps.

So what's the story? Should one buy an original NOS bumblebee or try to burn a modern cap hendrix-like?

I suggest to use what you already have in your bench, test it with proper instruments (including your ears ) and have a lot of fun doing it!

[gckelloch]

May 16, 2023 7:06:52 GMT -5 jandesign said:

Hi everyone!

My experience started some years ago, after reading tons of threads on websites/forums. I decided to buy some 0.022mdf caps from different brands to see if I could hear and/or tell which is which. They all sounded the same, and one was a little muddier. Guess what: I measured them all and THAT one showed a different value. The tolerance plays in the game as well...

My favourite cap is now mounted on my personal guitar and it must be a "paper in oil" thing from an old 1961 italian radio. Some components from that radio weren't working anymore. My chosen cap is literally burnt from one side, but it does the job in its own way. The sound is very mellow and sweet. My explanation is that it must be leaking some frequencies while retaining others in a non-linear unexpected way.  The value is 0.011 mfd and, needless to say, nothing compared to other 0.011 mfd caps.

So what's the story? Should one buy an original NOS bumblebee or try to burn a modern cap hendrix-like?

I suggest to use what you already have in your bench, test it with proper instruments (including your ears ) and have a lot of fun doing it!

All do respect, was the 22nF cap with the different value the last one you tried? Did it look different? Were your ears fatigued? Was there any level of tinnitus change or wax build up in your ears over the test? Did you play in exactly the same position on the strings, with the same pick angle and force, and listen in exactly the same place and angle from the speaker? Was the room temp and humidity the same for each comparison? Did you record the results and listen back to the same passage for each in quick succession on different days?...

My suggestion is to try a few very different values with your pickups and pick one that sounds closest to what you want. My experience is it isn't worth bothering with value differences of less than ~25%, but much smaller values can sound very different above 5 on the tone knob than larger ones.

[newey]

May 16, 2023 14:19:40 GMT -5 gckelloch said:

was the 22nF cap with the different value the last one you tried? Did it look different? Were your ears fatigued? Was there any level of tinnitus change or wax build up in your ears over the test? Did you play in exactly the same position on the strings, with the same pick angle and force, and listen in exactly the same place and angle from the speaker? Was the room temp and humidity the same for each comparison? Did you record the results and listen back to the same passage for each in quick succession on different days?...

All of that . . .And, there's also the blind part of any such testing.If the person doing the testing knows what cap is being tested, subjective results may be biased by preknowledge.

If we truly wanted to eliminate the potential personal biases, we'd need a group of about 10 sets of ears, none of whom know what is being tested.

And, Yogi B- I am not even going to pretend to follow your technical explanation, I'll have to defer to your expertise on that. But what I couldn't garner from your post was whether this theoretical difference between caps would translate into any real-world effects.

[kitwn]

May 16, 2023 0:19:22 GMT -5 Yogi B said:

looking at the table for a .012-.022μF (12-22nF) cap rated at 100V-600V measured at 20kHz, and generously assuming the DF value applies to the lower end of the capacitance range, ESR ≈ 0.20558Ω.
...
However taking the same cap at 100kHz, we get only ESR ≈ 0.06631Ω

Very interesting. I'm a retired radio frequency engineer and we usually think of component performance falling off as frequency goes up not down! A quick squint at Wikipedia shows that ESR is inversely proportional to frequency squared. Of course the reactive impedance of the capacitor goes up as well, but not as quickly. I still don't think you'd hear the difference at a live gig which is what really matters, and any subtle, frequency dependant effects you're after can be more precisely mimicked and experimented with using additional, inexpensive components of known behaviour anyway. Far better than paying hundreds of dollars for dried out crap from the 50s of completely unknown and inconsistent performance. There are good reasons why "they don't make them like they used to"

[Yogi B]

May 16, 2023 20:07:43 GMT -5 newey said:

And, Yogi B- I am not even going to pretend to follow your technical explanation, I'll have to defer to your expertise on that. But what I couldn't garner from your post was whether this theoretical difference between caps would translate into any real-world effects.

The kind that would be evident in a guitar tone control, nope! Mainly, I was just grasping at straws at how one could potentially (with some excessive extrapolation) arrive at to the original supposition.

The resistance of the tone pot swamps any comparatively miniscule difference in ESR between caps (of the values/types commonly used) — only with the pot all the way down would the ESR be significant, but it's still around one thousand times too small to be audible.

May 17, 2023 2:58:04 GMT -5 kitwn said:

I'm a retired radio frequency engineer and we usually think of component performance falling off as frequency goes up not down!

For the most part it's only the absolute ESR which gets worse at lower frequencies — measured relative to the magnitude of the impedance, performance is still better at lower frequencies (up to a point). Speaking of which...

A quick squint at Wikipedia shows that ESR is inversely proportional to frequency squared.

I can't find the article that states this, and I don't doubt that it's true, but only for (very) low frequencies. Remember that for a given expression that is a sum of powers of a variable, as that variable increases the terms with larger powers become dominant — i.e. an f^-1 term or a constant (f⁰ term) would become more significant than an f^-2 term.

We can somewhat see this in the data from the table, between 20kHz & 100kHz the average gradient of ESR for the largest caps is approximately -1dB/decade (i.e. changing very little, almost constant). Whereas for the smallest caps (which have a frequency response shifted upwards, and therefore are roughly similar to examining the larger caps at a lower frequency range), the gradient is approx -16dB/decade (getting closer, but still not surpassing inverse linear, -20dB/decade).

I've since found KEMET's K-SIM which, for the arbitrary selection of caps I've looked at, mostly shows ESR as being vaguely proportional to the total impedance — inversely proportional to frequency at low frequencies & proportional to frequency at high frequencies (once the cap's inductance becomes significant).

Only with the very smallest caps, such as the 500fF (0.5pF) shown in the plot below, does the behaviour of ESR being inversely proportional to the square of frequency begin to creep in at the lower end of the measured frequency range (100Hz—10GHz).

[kitwn]

Those of us with an engineering background generally have a feel for which aspects of the science involved will be significant, or the ability to work out what we don't know for sure, though I freely admit that my background means the the self-resonant frequency or specified max ripple current of a capacitor (neither of which are relevant in your guitar guys) have always been more relevant than ESR. Fact is your choice of pick will be 10^x times more significant in affecting your output tone than the ESR of your capacitor and I haven't seen anyone advertising genuine 1950's picks at $600 a pair!

For those people who's expertise is not in electronics the story is different and it is the exploitation of this ignorance by sellers who try to convince our young players that they MUST have this expensive capacitor, that expensive pickup with genuine 70 year old (probably hygroscopic) fibre coil formers etc. etc. if they want to sound any good that annoys me, probably more than it should.

Kit

[ssstonelover]

Thanks everyone,
The conclusion then that the breathless rave of the internet source is not credible and that a properly measured and accurate value in a large voltage capacitor would in reality behave similarly to one of equal value but low voltage cap as the tone knob was turned from 10 to 5 and that maybe possible very minor differences would be masked by the pickup, pot, and other components in the signal chain.
In a way this is a relief to hear, so tone control can go down paths which are effective, yeah.

[kitwn]

May 22, 2023 11:20:35 GMT -5 ssstonelover said:

Thanks everyone,
The conclusion then that the breathless rave of the internet source is not credible and that a properly measured and accurate value in a large voltage capacitor would in reality behave similarly to one of equal value but low voltage cap as the tone knob was turned from 10 to 5 and that maybe possible very minor differences would be masked by the pickup, pot, and other components in the signal chain.
In a way this is a relief to hear, so tone control can go down paths which are effective, yeah.

That's about it. Most people today would probably pick 10% tolerance components for non-critical uses such as a guitar tone circuit (50-70 years ago 20% was the norm) but If you're worried about accurate values then 5% tolerance capacitors are not exactly expensive. You're wiring will probably make a bigger difference anyway. The price below is $AUD by the way, so somewhat less if you're in the USA.

Kit

PS  I'd love to see Jim Lill do a video comparing the same guitar with one of these in and then one of those 1950s bumble bees.🤣