So what in solder "sounds" bad, and what can be done about it?

[big_teee]

Mah, what chew talkin bout. Weldin is easy.

I thought so too.
I just Gob and Dob.
If you melt enough weldin Rods on top of it good enough.
Just Kiddin.
Soldering is pretty basic, 99% of the time either you Made a good Connection or you didn't.
Most of time when someone makes a cold joint, it is from movement before the solders cools.
It has to kept still while drying or it crystalizes in movement, and makes a cold joint, that will haunt you later.
B_T

[RedHouse]

I thought so too.
I just Gob and Dob.
If you melt enough weldin Rods on top of it good enough.
Just Kiddin.
Soldering is pretty basic, 99% of the time either you Made a good Connection or you didn't.
Most of time when someone makes a cold joint, it is from movement before the solders cools.
It has to kept still while drying or it crystalizes in movement, and makes a cold joint, that will haunt you later.
B_T

Yep, and that's why PCB mfr's have a step called "re-flow", keeps anything like that from going out the door.

It's a good thing (re-flow) but take care to watch your total time heat is applied (and iron temp) if it's semiconductors you are soldering, but wire and terminal joints (like pickups/pots/jacks, or ceramic tube sockets) are very fair game for a final re-flow after assembly.

[FunkyKikuchiyo]

Then two months later, I picked up my guitars and compare, I felt differently in the same room, using the same amp!!! Very different from what I remembered.

Bear in mind that some have had the experience of pickups changing over time. The exact reasons I'm not sure and I vacillate, but if you play a pickup immediately after first assembly and then wait a bit, they do sound a bit different. I'm actually yet to make a pickup that sounds the same from first build to settling in a week or two later. Sometimes it is dramatic, sometimes it is subtle.

Alan - I know of the method you speak, the "solder painting"... definitely bad! It also overheats components in my experience.

[J M Fahey]

I know of the method you speak, the "solder painting"... definitely bad! It also overheats components in my experience.

That´s why when I teach Musicians soldering (they *should* at least be able to solder a plug on their own) i emphasize:
"a soldering gun is not a brush, is not a brush, it´s a syringe, you (slowly) inject heat into the part to be soldered" "it takes from 1 (bare wire) to 5 seconds (pot case or metallic jack bolted to chassis)"
And: "do not heat the solder, heat the part until IT melts the solder on contact"
Also: "do not trust Rosin to clean on its own, scratch shiny what you want to solder" (unless it´s already shiny copper or gold plated or has recently been tinned)
Success goes from 10/20% to 90/100%.

[kleuck]

Fine strands don't help skin effect unless they're separately insulated. If the strands are bare, HF current hops from one to the other to stay at the surface of the bundle. Jim LeSurf shows evidence to prove that in the article I linked.

So ?
Just did the maths, considering a 5 meters and 0.5 cm diameter cable, found a 0,4 ohms resistance at 10Khz, wich would lead to a 0,02 db of attenuation in a 8 ohms speaker.
Bullshit imo, if you can hear a difference between speaker cable, it's certainly NOT because of the skin effect, or, the formulas are wrong.
Give it a try, here's how to calculate the resistance of a cable.

La résistance d’un conducteur est directement proportionnelle à sa résistivité () et à sa longueur (L) ; elle est inversement proportionnelle à la section (S) de ce conducteur.

R = x L/ S

Avec R la résistance électrique du conducteur en ohm
résistivité de la matière en mm²/ m
L longueur du conducteur en m
S section du conducteur en mm²

Pour rappel, la section d’un câble suivant son diamètre se calcule comme suit :
S= x d²/4 Ou suivant son rayon : S= x r²

Quelques exemples de valeurs en mm²/m à 20°C :
Cuivre 0.017
Argent 0.016
Alu 0.028
Fer 0.1

[Alan0354]

The matt "dry joint" appearance certainly sounds like lead free solder. I hate the stuff too!

Some high-end hi-fi speakers dip very low in impedance at certain frequencies. Low enough that if you had long speaker cables, you could maybe hear a lack of output at those frequencies with 18 gauge lamp cord, and four 12 gauge runs in parallel would fix it.

Fine strands don't help skin effect unless they're separately insulated. If the strands are bare, HF current hops from one to the other to stay at the surface of the bundle. Jim LeSurf shows evidence to prove that in the article I linked.

A rough surface always has more surface area than smooth surface. Even is 30% more is a big difference.

[Steve Conner]

So ?
Just did the maths, considering a 5 meters and 0.5 cm diameter cable, found a 0,4 ohms resistance at 10Khz, wich would lead to a 0,02 db of attenuation in a 8 ohms speaker.
Bullshit imo, if you can hear a difference between speaker cable, it's certainly NOT because of the skin effect, or, the formulas are wrong.

I've used that fine stranded speaker cable to carry something like 30A RMS at 200kHz in my power electronic experiments, and it didn't get hot or melt, so it seems to be a reasonable conductor of high frequencies. High in audio terms at least.

But bear in mind that cables are usually thinner than 5mm diameter, and speakers usually dip lower than 8 ohms. I've seen electrostatics that go down to 1.5 ohms at high frequencies. They're designed by SCAMP, the Society for Cruelty to Amplifiers.

Can you spot the audiophile grade wiring?

[kleuck]

Mistakes here.
Done again, i found for a 2.5 mm² & 5 meters cable a resistance of 0.032 ohms, and 0.036 at 10 Khz.
Can lead to an attenuation of 0.0087 db at 10 Khz.

[km6xz]

"minutiae was not lost to Leo.."
The range of variability of a pickup position, turns count, core style is what, 1-2-3 orders of magnitude greater than the variability in wire and gold-eyes anointed amplifier cone supports or twinkle fairy dust, you will have to admit. Or solder technique.

That is like equating a hot rodder's tinkering with gear ratios being the same as tinkering with the gauge of wire used in the heated side mirror. One has demonstrate-able impact and the other doesn't. The vast majority of fairy dust exotic parts hi-fi and gold-eye amp gurus are trying to sell, probably have less real world benefit than the $0.25 change in mirror heater wire.
The reason some techs or players get worked up about parts with questionable benefit is not having something of value to present. Not one person ever enjoyed more or less a song because was played with an amp that had its $100 transformer replaced with a $800 silver wire transformer. That is for an insecure player to tell others that something significant has been done when all the player really needed was to practice a lot more and get good enough to tell a story worth listening to with their music.

Just today a guy came to my place to have me look at his malfunctioning tube stop box. He was frantic because the part his needed would have to wait until Monday when the parts house was open. "I have a gig tonight...you have to get it done now" I asked what played with before the preamp was in-line. "Nothing, just into the amp so it sucked".
Are you saying audiences walked out and the songs were bad....or that you did not feel confident with only your fingers and skill? If the audience thinks you are bad or good has NOTHING to do with the specific nature of the distortion (actually this was just a preamp, with a little bit of a compressor, no distortion). I am sure during the performance tonight, about now, he either sucked or did fine regardless of the gear. Listening to a complex combination of playing, acoustics, material and skill with nothing else to compare with will sound just fine if the total performance was likable.
Players who have something interesting to say, and are skilled enough to have created a reproducible sound really do not stress about equipment as much as amateurs who think gear makes the player and song do. When players like Santana or B.B. King came to play in the studio, it was quite telling that there was so little concern over the gear. B.B. did not even care which amp be used, just brought his guitar, extra strings and a cord. Two notes however, and there was no mistake who on the playing end of that thing.
If someone is not going over well, it has nothing to do with gear, it is your stage personality and material, assuming there is a level of competency in playing that material.

[big_teee]

"The reason some techs or players get worked up about parts with questionable benefit is not having something of value to present."

AMEN Brother!
don't have a problem with the Cork Sniffer Customers. We will Always have those!
Do have a Problem With Builders and Sellers, Selling his Snake Oil, Con Job, and Using Smoke, Magic, and Mirrors to Persuade someone to buy something!
B_T

[salvarsan]

ObMarketing point: fetishistic customers are your legal prey. Don't gouge them too hard before they wise up.

RE: voltage coefficient of resistance:

There is scarce data on R(v) for old resistors. The mil spec JAN-R-11 established allowable limits as 0.02%/v and 0.035%/v at 10% and full rated voltages respectively. Thus, the remaining carbon comp resistor manufacturors have no incentive to report their specs differently. TT's IRC division, a modern manufacturor, uses exactly those figures, although claiming a NEGATIVE voltage coefficient. See:
http://www.irctt.com/file.aspx?produ...type=datasheet

Since the voltage coefficient changes with voltage, you could approximate the actual resistance as a 2nd order differential equation.
I won't because I'm lazy and not too good at it, either. Here's a quick and dirty linear approximation:

Ractual = R * (1 - (.0002 * V) - (.0035 * (V/Vmax))
where:

• R = resistor value
• Ractual = calculated resistance at voltage V
• V = applied voltage
• Vmax = maximum continuous voltage, usually 250VDC for 1/4 watt units

You MatLab dweebs can plug this in and get your 1st order approximation plot:

R = 10000
V = linspace(1,250,250) ; # DON'T FORGET THE SEMI-COLON!!
Rv = R .* (1 - (0.0002.*V) - (0.0035.*V/250)) ;
plot(V,Rv)

For a modern carbon comp 10k resistor, it looks sorta like this, decreasing from 10,000 ohms down to 9465 at the limit, about 5% full range.
CC resistors made prior to the JAN-R-11 recommendation will show a larger change, so far unknown.

If you are using it in the standard Fender preamp stage, the 160-240VDC range is a good place to do your simulations.



Feel free to bitch and take cheap shots in the usual Internet traditions of intellectual laziness.
...or refine the model and report back.

[km6xz]

ObMarketing point: fetishistic customers are your legal prey. Don't gouge them too hard before they wise up.

RE: voltage coefficient of resistance:

Feel free to bitch and take cheap shots in the usual Internet traditions of intellectual laziness.
...or refine the model and report back.

No reason to bitch, you are making a good wrap point that hobbyists are usually forgetting(or think some golden ears type "discovered" voltage coefficient" but is a consideration and basic element in design for engineers.

The problem with "fetishistic customers" is that their unrealistic assumptions can be swayed by the next wild claim read in a magazine and suddenly all your good work for them is not up to par and even deconstructive to their sound, in their new opinion. You loose a customer and gain a bad rep because those types are very vocal. For longer term positive relationships with customers I have found that it is better to educated them about some aspect they are deluded about instead of profiting by exploiting their willful ignorance.

It was interesting to read here that some golden ears types were cited as actually doing measurements and found that resistors vary with voltage. Some apparently were impressed by the originality of that discovery but I was more astonished that someone felt it was a new concept, one that had not been presented in basic electricity class or that it is something routinely considered in almost all designs.
So far I have seen no hi-end hi-fi or guitar amp guru discover anything that advances the art, only misleading marketing hype, used for the express purpose to mislead customers. I have never seen one paper submitted to journals. Much of the claims and hucksterism is not even original, much was borrow directly from the alternative health industry which owns the crown for hucksterism. Only there, the results of mass ignorance is serious deterioration of health and possibly death. With sound systems, the only destruction is to the wallet.

Did you ever notice how all these "breakthroughs" touted in esoteric audio publications are really really expensive? Generally, in other fields real breakthroughs tend to lower costs and increase reliability by the time it gets to market.

[Alan0354]

ObMarketing point: fetishistic customers are your legal prey. Don't gouge them too hard before they wise up.

RE: voltage coefficient of resistance:

There is scarce data on R(v) for old resistors. The mil spec JAN-R-11 established allowable limits as 0.02%/v and 0.035%/v at 10% and full rated voltages respectively. Thus, the remaining carbon comp resistor manufacturors have no incentive to report their specs differently. TT's IRC division, a modern manufacturor, uses exactly those figures, although claiming a NEGATIVE voltage coefficient. See:
http://www.irctt.com/file.aspx?produ...type=datasheet

Since the voltage coefficient changes with voltage, you could approximate the actual resistance as a 2nd order differential equation.
I won't because I'm lazy and not too good at it, either. Here's a quick and dirty linear approximation:

Ractual = R * (1 - (.0002 * V) - (.0035 * (V/Vmax))
where:

• R = resistor value
• Ractual = calculated resistance at voltage V
• V = applied voltage
• Vmax = maximum continuous voltage, usually 250VDC for 1/4 watt units

You MatLab dweebs can plug this in and get your 1st order approximation plot:

R = 10000
V = linspace(1,250,250) ; # DON'T FORGET THE SEMI-COLON!!
Rv = R .* (1 - (0.0002.*V) - (0.0035.*V/250)) ;
plot(V,Rv)

For a modern carbon comp 10k resistor, it looks sorta like this, decreasing from 10,000 ohms down to 9465 at the limit, about 5% full range.
CC resistors made prior to the JAN-R-11 recommendation will show a larger change, so far unknown.

If you are using it in the standard Fender preamp stage, the 160-240VDC range is a good place to do your simulations.

[ATTACH=CONFIG]15963[/ATTACH]

Feel free to bitch and take cheap shots in the usual Internet traditions of intellectual laziness.
...or refine the model and report back.

Can you give the 2nd order differential equation formula? I fail to see the reasoning of the approximation of Ractual formula you gave.

One thing I want to say, so far everyone is talking about the dR/dV of the resistor that make it sound like a more even function. But in reality, amp has multiple stage, if you stack two stages together, one goes up when the other goes down, this will simply make the waveform symmetrical and become a compression at both end instead of a 2nd order harmonics heavy waveform.

[Steve Conner]

The problem with "fetishistic customers" is that their unrealistic assumptions can be swayed by the next wild claim read in a magazine and suddenly all your good work for them is not up to par and even deconstructive to their sound, in their new opinion. You loose a customer and gain a bad rep...

Did you ever notice how all these "breakthroughs" touted in esoteric audio publications are really really expensive? Generally, in other fields real breakthroughs tend to lower costs and increase reliability by the time it gets to market.

I just finished restoring a high-end Class-A hi-fi amp for someone. I took it on as a kind of speculative project and didn't charge a great deal of money for it, and I warned him that all I could guarantee was that it met the original power output and THD spec, there was no guarantee that whatever mojo it had would survive.

He got it back and reported to me that it lacked weight and mastery. Now what on earth part of the circuit controls that?

Well, translated into engineer speak, it probably meant that it was lacking bass and/or simply wouldn't go loud enough without distortion. On further investigation, it turned out that the amp had an abnormally low input impedance that would be hard for a preamp to drive. (Another esoteric breakthrough? ) His DAC/preamp could drive it before, but while the amp was away, he had been busy modding the line output stage, changing it from an op-amp to a discrete circuit.

I suggested he plug it into the DAC's headphone socket, as a headphone amp would almost certainly have a lot of drive capability. Lo and behold he said that it was now muscular, robust and almost scarily loud.

Why am I telling this story? In my heart of hearts I believe that the subjective and objective camps are talking about the same phenomena and at the end of the day they must eventually agree. And somehow we waddle forward with one foot in each.

[km6xz]

Technically the ceramics were less than perfect....technically speaking, but guitar amps are not precision amps, they are tone creators that are not meant to be linear, low distortion or uniform in any parameter. For instrumentation amplifiers or even hi-fi reproduction amplifiers I would think they are not a good choice. Guitar amps, however, are production devices, not re-production devices. That is the big difference and reason why some thrown together amp with poor engineering ends up sounding very pleasant to listen to as it creates sounds that are not intended to be analogs of what went into them. A guitar pickup and strings alone create a signal if reproduced faithfully would sound very unappealing. Mixing precision electronics with sound production is like oil and water, they have different intents and functions.
Once the total sound of the instrument, amplifier, speaker, acoustic space and player influence has created a sound, faithfulness of RE-reproduction becomes very important and desired. If the total sonic experience was pleasing to you, using ceramic caps or two nickels taped together with a piece of rolling paper between, should not be the criteria....it the sound when all those production elements come together and interact with the listener's perception systems is all that matters. EVERY single point of signal path after the sound has been created however becomes evaluated by faithfulness. Often, as in the case of recording, hi-fi is not what is needed to create the impression of faithfulness, as long as the listener believes that it is a faithful reproduction, that is all that matters in that chain. Hi-fi golden ears have forgotten or not known this so assign objective terms to subjective perception of a fake or modified analog intended to give the impression that the recording is a faithful representative of the actual acoustic event. They will argue for hours about which amp or wire presents the best "sound stage" and placement of sound sources....yet fail to mention that all position perception is fake, intended to give the impression that in fact worked.
So the worlds of "sounds appealing" clashes with "sounds accurate" in every step of music production and reproduction. Assigning too much importance to engineering is folly. Engineering is great in its own right, and leads to better devices, more reliable, more predictable and often even cheaper but it is not so important in sound creation.
For example I measure parameters for every amp I look at, not because its Thd, noise floor, mains idle current or spectral plots are important in a working production device but to give clues as to the state of the unit in case there are not so obvious technical problems. If the data collected deviates much from expectations for that model, it is a good starting point to return the unit to normal operation. In a re-production device, those parameters do matter because there is an inferred goal set of parameters.