Les Paul Jr Capacitor Upgrade

[Jackie Treehorn]

Yep extremely tiring. I'm an electronics engineer and work with audio signal processing for a living, and I've posted all sorts of technical analysis in the past on this forum and others. Back in the days of the CD player bit wars, I was involved in studies of this sort of thing that were commissioned for no other reason than to convince non-technical (or semi-technical) purchasing people that we could (or couldn't) use a different component in a properly designed system with no audible effects that might have somehow escaped our circuit modelling or physical test systems.

The only problems we found were when a circuit was poorly designed, or other circuits (sample and hold, RF, power coupling, etc). In guitar circuits, the biggest measurable effect might by ESR or leakage. And those are easily compensated for with parallel or series resistance on a decent ceramic or poly cap - and almost always are more easily compensated by turning a knob a few degrees. That's ignoring the effect that temperature has on the rest of a typical guitar circuit. In the end, any effects that might be audible in a guitar tone cap are swamped by other variables in the system. Dynamic absorption and other non-linearities and so far below the thermal noise in a guitar tone circuit that it is laughable to even consider them.

The biggest issue is when people conflate knowledge from different domains. Ie. radio frequency, power, high impedance/low impedance, instrumentation, etc. This is rampant in marketing (and therefore consumer misunderstanding), specifically in other areas like audio interconnect/speaker/power supply/guitar cabling, power amplifiers, pre-amplifiers, DACs, speakers, etc.

What’s relevant and factual is that different types of capacitors measure differently. People say, if it’s audible, it’s measurable and the inverse of that is if it’s got a measurable difference it should be audible. Unlike you, I don’t have to speculate about whether the differences in capacitors are great enough to be audible. I’ve listened and they are, so your conclusions are wrong.

Further, no one has stated any particular causation, simply that different capacitors can sound different and it could be any combination of reasons.

So, you are free to use your imagination as you have been doing and run through every unsubstantiated straw man you can dream up or you can deal with reality, that everything sounds different to some degree and not everyone is going to hear them at the same time.

It is possible that capacitors sound different in guitars, but you cannot hear the difference in whatever context you listen. Have you thought of that as a possible explanation given your exhaustive research into human fallibility?

 

[pipedwho]

What’s relevant and factual is that different types of capacitors measure differently. People say, if it’s audible, it’s measurable and the inverse of that is if it’s got a measurable difference it should be audible. Unlike you, I don’t have to speculate about whether the differences in capacitors are great enough to be audible. I’ve listened and they are, so your conclusions are wrong.

Further, no one has stated any particular causation, simply that different capacitors can sound different and it could be any combination of reasons.

So, you are free to use your imagination as you have been doing and run through every unsubstantiated straw man you can dream up or you can deal with reality, that everything sounds different to some degree and not everyone is going to hear them at the same time.

It is possible that capacitors sound different in guitars, but you cannot hear the difference in whatever context you listen. Have you thought of that as a possible explanation given your exhaustive research into human fallibility?

You sound exactly like the 'audiophiles' that we used in the past for 'expert ears' testing. As soon as the properly conducted double blind testing started, their assertions no longer held. Guitar capacitor audibility is implausible unless you're using different values or some seriously deteriorated cap with huge ESR or leakage (both of which can be duplicated with series/parallel resistances).

As I said earlier, set up a double blind test party and join in the fun - I'll even come along with food and beer if it's local. If you can hear a difference in these guitar circuits across a statistically significant series of tests then further analysis may lead to a new discovery in the field of electronics.

 

[Jackie Treehorn]

You sound exactly like the 'audiophiles' that we used in the past for 'expert ears' testing. As soon as the properly conducted double blind testing started, their assertions no longer held. Guitar capacitor audibility is implausible unless you're using different values or some seriously deteriorated cap with huge ESR or leakage (both of which can be duplicated with series/parallel resistances).

As I said earlier, set up a double blind test party and join in the fun - I'll even come along with food and beer if it's local. If you can hear a difference in these guitar circuits across a statistically significant series of tests then further analysis may lead to a new discovery in the field of electronics.

Like I said earlier, I already did the blind test and picked out the bumblebee. We could blind test all you want, but the only new discovery for you will be you’re wrong.

 

[zombiwoof]

The Gibson Bumblebee caps are often just cheap poly wrapped in "Bumblebee" plastic. See this article - https://www.mylespaul.com/threads/historic-gibson-faux-bumblebee-caps.118027/

Gibson has already addressed that issue, they are now using real Luxe repro BB's, which these days are hand-made by the guy at Luxe. He used to use Russian PIO's inside the Luxe PIO's, but now builds them completely by hand, according to the Luxe website. Gibson no longer uses the fake Bees that had a film cap inside.
Al

 

[COYS]

Like I said earlier, I already did the blind test and picked out the bumblebee. We could blind test all you want, but the only new discovery for you will be you’re wrong.

A certain number of people would "pick it out" just by random guessing. That's why a more serious trial would be needed to actually prove anything.

 

[pipedwho]

Like I said earlier, I already did the blind test and picked out the bumblebee. We could blind test all you want, but the only new discovery for you will be you’re wrong.

If it was a properly conducted double blind experiment with sufficient tests to allow a statistically significant result, then I'm happy to be wrong. At which point the components could be investigated to establish what unknown physics are occurring and maybe make an advance in the field of electronics, or to establish a problem with the testing model.

 

[Eugene Wallace]

if you are talking about whether or not one cap of the same value inside the tone section of a guitar may sound better than another, then I'd say no.

It's like in an amp, if you just change one cap ( of the same value) , it will mean very little/nothing..even if it the first cap coming off the plate of v1a.
Once that 'slightly defined by the signal cap' signal reaches other caps, it's those other caps that are going to take over and re-shape whichever way that first cap shaped the tone. ( and tubes too)
There is a cumulative action thing going on.
If you changed them all the way down the line to the Phase inverter then, well that's another story. ( when, not if, a person tries to tell me changing out all the mustards in a marshall or blue fender caps in an old fender amp with orange drops makes no difference to the sound of that amp, I simply avoid considering anything they ever have to say to me about tone like they have the plague)

i may be wrong on this because I've never done any experiments with the tone cap of a guitar , but I have done a LOT of mucking around with signal caps before.
If I'm wrong, I'm happy to accept it.

this reminds me of the use of carbon comp resistors in pedals. The use thereof is waste of a good carbon comp resistor because the voltages are waaay too low for the carbon comp to do anything except add hiss. Ironic. you need over a hundred volts and some serious voltage swing for them to have an effect, to add nice harmonics.

And behind all these things, someone is making money, what a surprise. A LOT of fake signal caps are being sold on ebay , Astrons and Mustards and I daresay these waxy ones used in old teles etc.

 

[pipedwho]

A certain number of people would "pick it out" just by random guessing. That's why a more serious trial would be needed to actually prove anything.

The usual problem with these sort of non-rigorous tests is that the setup is prone to bias. For example, where a capacitor change in a guitar is used, the player's hand technique will alter the output signal and will be heard as a difference by the testers. Also, the capacitor value much be matched as precisely as possible to avoid differences due to capacitance value. The capacitor leakage and ESR should also be measured to make sure the cap under test is not faulty. If the goal is produce a match to a faulty 'golden sample' capacitor, then a parallel or series resistance should be added to the test set up as a third option to establish a baseline for the expected effects based on currently understood electronic principles. Even the order of testing is important with good (pseudo or real) randomisation to avoid temporally distributed correlations creeping in to bias the result.

With a series of a dozen tests where the player doesn't know if the guitar is using capacitor A, B, C, etc after a random change, the resultant test becomes statistically significant. The player and listener must not know which capacitor is switched in, or even if the switch has changed to another cap, otherwise the testing is invalidated due to bias.

And as you say, if the sample size is too small, the results become unreliable. If less people are being used for the test, then the number of tests must be increased per person to avoid pitfalls of finding a pattern in an otherwise random set of guesses.

 

[Eugene Wallace]

these blind tests -

not in relation to this matter as for me it's a no brainer as to what is true.

but generally, I think it has to be considered..' who is taking these tests?'
The better you play, the better you hear..whether the player learnt by repetition or just sort of plodding along/ studying whatever.
If the player learnt by repetition, then I think you'll find they (possibly ironically) have the best ear of all.( ask yourselves..how did they know what was ' good playing' and ' bad playing'?)
If they learnt by study, then they learnt to hear probably more slowly, simply because you can't think tone and hear it at the same time.
There are a lot of people who uh..want to play the guitar and like it very much and everything, but they don't have the hearing yet.
Are these the people taking the blind tests?
Be mindful that if results are being published on the merits of something solely based on tests conducted by people weren't in the top range of hearers ( believe me, it's just like that, some people have much, much better hearing than others in relation to tone).Then the results are critically compromised.

Richie Sambora and Ritchie Kotzen aren't gonna take the test, you know what I mean?

 

[pipedwho]

these blind tests -

not in relation to this matter as for me it's a no brainer as to what is true.

but generally, I think it has to be considered..' who is taking these tests?'
The better you play, the better you hear..whether the player learnt by repetition or just sort of plodding along/ studying whatever.
If the player learnt by repetition, then I think you'll find they (possibly ironically) have the best ear of all.( ask yourselves..how did they know what was ' good playing' and ' bad playing'?)
If they learnt by study, then they learnt to hear probably more slowly, simply because you can't think tone and hear it at the same time.
There are a lot of people who uh..want to play the guitar and like it very much and everything, but they don't have the hearing yet.
Are these the people taking the blind tests?
Be mindful that if results are being published on the merits of something solely based on tests conducted by people weren't in the top range of hearers ( believe me, it's just like that, some people have much, much better hearing than others in relation to tone).Then the results are critically compromised.

Richie Sambora and Ritchie Kotzen aren't gonna take the test, you know what I mean?

This is a major consideration when using human test subjects. Especially for critical listening scenarios. The approach we used back in the early days of audio codecs was to weight the results by the 'quality' level of the test subject. We spent a lot of time 'training' test subjects to hear various compression artefacts. Nearly anyone could be used to qualify low bit rates, but as we went higher, people needed to be able to 'hear' the psychoacoustic processing to be able to rate improvements and regressions. Most people couldn't hear any difference even at low standard rates.

The same process would be used for rating guitar tone. The trick here is to find a number people that otherwise already think that they can distinguish these differences - of these at least some might be near the top of their game. People like Sambora and Kotzen are no more likely to be any better than more accessible subjects, they are just famous examples.

Another technique we needed to use was to include a periodic test that was obvious enough that the test subject could barely but clearly hear a distinguishable change. Without this, people would get frustrated that they can't actually hear any differences, so they'd stop trying and just guess without properly listening.

We used regular paid test subjects that we'd bring in for this.

For testing guitar tone caps, the player and listener would likely be the same person. The best ones would be those that already have a specific tonal preference and practice regularly enough for it to be ingrained.

However, a guitar tone pot can be tested without any player required. Using an induction coil to inject a consistent signal into the pickup, various sweeps could be used to fully characterise the circuit. Harmonic distortion could be directly measured, correlations made between components under test, and numeric comparisons made. From this, a baseline could be established if a difference was found to exist. And further human testing could be included to determine the level of perceptibility. These differences could be exaggerated and used to train critical listeners.

 

[Eugene Wallace]

astounding!

 

[Eugene Wallace]

the reason I mentioned R. Kotzen is because of this type of activity.
The same hearing ability as more accessible musicians? I highly doubt that.
In order to get to his level, you have to hear real nice.

2:53 ..- and note the absence of a plectrum.
edit- the love actually starts at 2:33

https://www.youtube.com/watch?v=4SoTgZjokLU

 

[Jackie Treehorn]

This is a major consideration when using human test subjects. Especially for critical listening scenarios. The approach we used back in the early days of audio codecs was to weight the results by the 'quality' level of the test subject. We spent a lot of time 'training' test subjects to hear various compression artefacts. Nearly anyone could be used to qualify low bit rates, but as we went higher, people needed to be able to 'hear' the psychoacoustic processing to be able to rate improvements and regressions. Most people couldn't hear any difference even at low standard rates.

Fascinating. So, that’s exactly what I was saying previously, that not everyone is going to hear a difference at the same time. Imagine if what you term a quality listener and an average listener, or maybe someone who hasn’t listened at all but watched some YouTube videos, came to TGP to discuss codecs or whatever. How do you think it would go down? I bet you’d have the quality listener stating that the artifacts of the codec were audible and the rest saying no, the differences are not enough to be audible..thermal noise or something.

The same process would be used for rating guitar tone. The trick here is to find a number people that otherwise already think that they can distinguish these differences - of these at least some might be near the top of their game. People like Sambora and Kotzen are no more likely to be any better than more accessible subjects, they are just famous examples.

Another technique we needed to use was to include a periodic test that was obvious enough that the test subject could barely but clearly hear a distinguishable change. Without this, people would get frustrated that they can't actually hear any differences, so they'd stop trying and just guess without properly listening.

We used regular paid test subjects that we'd bring in for this.

For testing guitar tone caps, the player and listener would likely be the same person. The best ones would be those that already have a specific tonal preference and practice regularly enough for it to be ingrained.

However, a guitar tone pot can be tested without any player required. Using an induction coil to inject a consistent signal into the pickup, various sweeps could be used to fully characterise the circuit. Harmonic distortion could be directly measured, correlations made between components under test, and numeric comparisons made. From this, a baseline could be established if a difference was found to exist. And further human testing could be included to determine the level of perceptibility. These differences could be exaggerated and used to train critical listeners.

You’re right about ear training, but that’s not how it works on the internet. Strangely, not everyone is open to the possibility that there is more to hear. Imagine that sufficient testing and ear training has already been done by some, but it contradicts people’s present beliefs. How do you think that would play out on the internet?

I bet a good number of them would not look at it as an opportunity to learn something as you suggest, but become defensive of their present belief, say the differences are imagined, and the testing and ear training are not the right type or quantity.

I’ve actually seen that quite a bit on the internet and people like that can be quite vocal on Internet forums.

 

[Ron Kirn]

Cornell Dubilier is indeed an excellent product but 7 bux on eBay or a buck 82 from Mouser..

https://www.mouser.com/ProductDetail/Cornell-Dubilier-CDE/WMF1S22K-F?qs=TmH0BUXeWC2zuuqCouYO2g==

3 pages and 53 posts and your never gonna know until you do it.

and that is the dawggone truth... just about all this tone chasing is based on unfounded conjecture... and absolutely NO scientific research... and that is further exacerbated by, we all hear to differing levels... the older.. beyond our teenage years... the worse it gets, specially if "you" have been subjected to noisy environments like standing in front of a rack of Marshalls trying to crank out Metallica riffs at 110 db...

Just do it.. caps are cheep.. if you're head is in the choice... wanna 1000.00 24k gold foil PIO. .. God.. I hope not..

r

 

[biter]

Meaningless debate over capacitors and tonewood,,,,, It's been a great month at TGP!!!