resistor replacement?

All depends on the circuit, not all caps are critical, it all depends on how one uses it. I only have one film cap in each channel, at the input to cut off low frequency below 10Hz or so. It is critical to have film caps, other than that, I use ceramic NPO caps for compensation and all. I got incredible THD result in measurement. And sound wise, in my opinion ( just my opinion, might even be partial), my amp beats the Nakamichi PA-7 Stasis amp designed by Nelson Pass and is almost identical to his famous Threshold S300 and even slightly better.

Like I said before, I did not know thick film resistor has 1/f noise, I use as many thick film SMD resistor as I could, I have not noise problem what so ever. In fact my amp is so much quieter than the Nakamichi and Acurus I have. I can't hear anything even when I pressed my ears to the speakers.
 
Is it safe to assume:

Dark brown tubular resistor = CC
Tan figure-8 shape resistor = CF
Blue figure-8 shape resistor = MF

Gray ceramic = Fusible (film?) resistor
White ceramic = wire wound resistor?

Green candy roll = Inductor

Sometimes it's hard to guess what's what...
 
All depends on the circuit, not all caps are critical, it all depends on how one uses it
very true. The ceramic caps I have changed and gotten improvement from have all been audio coupling caps, often in the tone control circuits. Used as bypass caps it doesn't really matter.
 
Is it safe to assume:
Dark brown tubular resistor = CC
Tan figure-8 shape resistor = CF
Blue figure-8 shape resistor = MF
...
Sometimes it's hard to guess what's what...

The blue may be precision wirewound. I remember those from the 1980s. The tan may be either CC or CF, depending. The coating is a not a formaldehyde-phenolic, but that's all one can say. It was common to coat both this way.

If you don't know, replace them.
 
very true. The ceramic caps I have changed and gotten improvement from have all been audio coupling caps, often in the tone control circuits. Used as bypass caps it doesn't really matter.

It's a bit more complicated than that.

Ceramic capacitors are today used for bypass purposes because of a combination of low cost, high density, small size, better temperature resistance than film, higher ESR, and overall smaller form-factor with greater ease of handling, which is particularly important for dense motherboards assembled by pick-and-place machines and then wave soldered. The non-linear properties, reduction in capacitance from DC bias, and other drawbacks are not that important for supply decoupling.

Film has very low ESR, as do certain other monolithics. The capacitor thus has high Q which can permit oscillation on high-speed mothersboards because the Q is so high (i.e. losses are so low) that high-frequency (again, low MHz to GHz) oscillations ripple back and forth through the capacitor without resistive losses. One technique combines low ESR ceramics with high ESR ones to solve this, but that can itself cause problems, including oscillation.

Mixing capacitors in a high-frequency (low MHz to GHz) circuit is tricky because mutual tank circuits form using the inductance from the printed-circuit board traces or another capacitor. (The trivial amount of inductance forms a very high-frequency tank.) The solution is, again, a combination of higher ESR devices or adding external resistors to form attenuators and thus lower the Q to the point that oscillations damp out. A related problem is different potentials arising on ground planes.

These problems do not, however, generally affect audio circuits. The exception is the poor quality of ceramics can reduce filter accuracy in FM tuners and that translates into sloppy decoding of MPX which degrades the sound. Changing ceramics to film can improve the quality of an FM tuner even when the number of gangs remains constant.
 
Those are another component I have no special love for. Shame, electrically they are rarely faulty but they sound bad. I rarely claim to be able to hear a component, but every time I've pulled ceramic caps for film it sounds better.

This is easily explained. Ceramic capacitors undergo a reduction in capacitance with applied DC bias. Such DC bias is common in audio circuits where ceramics are used as coupling capacitors, particularly plate to grid, blocking DC but passing AC. So the DC bias issue reduces the capacitance. This can be worse when the DC bias is shifting as the capacitance is similarly shifting.

Ceramic capacitors were only used because of lower cost. Modern manufacturing has lowered the cost of small-value film capacitors to a comparable point, so unless one is building equipment on a large scale the minor price differential should be largely irrelevant.

The accuracy of filters is better with film than ceramic.
 
yeah, that and the piezoelectric thing if your amp is near the speakers. I don't know if modern ones are any better than the kind used in the 50s and 60s, but thats the stuff I work on most of the time.

I just don't care for the sound, regardless of the physics behind it. The fact that there is actually a reasonable explanation is a comfort though, proves its not just my general insanity. For the most part I don't buy much into component sonic signatures but there are exceptions. Well, that and the ones caused by poor quality / defective parts but thats not really a "sound" so much as a "fault".
 
very true. The ceramic caps I have changed and gotten improvement from have all been audio coupling caps, often in the tone control circuits. Used as bypass caps it doesn't really matter.
Oh yeh, those matters. The ones that determine the frequency response are the critical ones, one wants to have the best cap. It's all about dC/dV, the more the voltage change across the cap during operation, the more critical it is to use a cap that has low voltage coefficient. of cause ideally, use caps that are dC/dV= 0, that is no change of capacitance when voltage across the cap changes. Tone circuit is one of those that the voltage across the cap change a lot during operation. Also, I think it is a lot more critical for tube amp than SS amp. You can get voltage swing of over 100V across the cap easily in tube circuit just by the nature of the tube circuits. But for SS amp, swing is a lot lower.

By that, when one design the circuit, identify the ones that see a lot of voltage change in normal operation and use the good caps. Try to design the circuit in such that you have minimal change in voltage across the cap as much as possible so you can use either electrolytic or ceramic caps. My experience is it's possible. Also, if accuracy is not important, use NPO ceramic caps if possible.

Resistor is the same thing, it's all about change of value with change of voltage across it.

All to save money.
 
yeah, that and the piezoelectric thing if your amp is near the speakers. I don't know if modern ones are any better than the kind used in the 50s and 60s, but thats the stuff I work on most of the time.

Barium titanate is inherently piezeoelectric. This cannot be fixed.

I just don't care for the sound, regardless of the physics behind it. The fact that there is actually a reasonable explanation is a comfort though, proves its not just my general insanity. For the most part I don't buy much into component sonic signatures but there are exceptions. Well, that and the ones caused by poor quality / defective parts but thats not really a "sound" so much as a "fault".

Many components do have a sound. Engineers call this "distortion" and try to eliminate it. Audiophiles describe it in terms used for wine, chocolate, and food, and seek it out.

Components should not have a sound, because that means the signal is being altered. All components should be neutral. I believe it was a Harman-Kardon ad which first proposed that an amplifier should be a straight wire with gain.
 
.........................
Components should not have a sound, because that means the signal is being altered. All components should be neutral. I believe it was a Harman-Kardon ad which first proposed that an amplifier should be a straight wire with gain.

Ha ha, That's how it supposed to be, but I found a lot of people that are into tube amps really don't want sound to be accurate, they want it to sound "nice", they want the tube sound, not the true sound.

I am building a tube amp, I want to hear it with my own ears what is all the fuzz about.
 
Just for completeness, dB relative to what?

The dB is a measurement of the noise voltage to the applied voltage which allows it to be compared to other any other dB measurement.

Here are some examples I worked out for you and anyone else who cares. (I'm never sure anyone is really interested in the math so I probably should just have stopped at the one sentence summary. But understanding the math often helps to understand how components work in actual circuits.)

Given:
NV = Noise Voltage (μVolts)
AV = Applied DC Voltage (Volts)


Then:

dB = 20 x log10(NV / AV)

Examples

The noisy end of the carbon-composite resistor at 6 dB has a ratio of:
6 dB = 20 x log10(NV / AV)
log10(NV / AV) = 6 / 20
NV / AV = antilog10(0.3)
NV / AV = 2 μVolt / Volt

Given 0 dB of resistor noise for carbon composite the ratio is:
0 dB = 20 x log10(NV / AV)
log10(NV / AV) = 0
NV / AV = antilog10(0)
NV / AV = 1 μVolt / Volt

Given –12 dB of resistor noise for carbon composite the ratio is:
-12 dB = 20 x log10(NV / AV)
log10(NV / AV) = -12 / 20
NV / AV = antilog10(–0.6)
NV / AV = 0.25 μVolt / Volt

Given –18 dB of resistor noise for carbon film the ratio is:
-18 dB = 20 x log10(NV / AV)
log10(NV / AV) = -18 / 20
NV / AV = antilog10(–0.9)
NV / AV = 0.13 μVolt / Volt

Given –32 dB of resistor noise for metal film the ratio is:
-32 dB = 20 x log10(NV / AV)
log10(NV / AV) = -32 / 20
NV / AV = antilog10(–1.6)
NV / AV = 0.025 μVolt / Volt
 
The dB is a measurement of the noise voltage to the applied voltage

Noise is normally measured against an absolute reference, be it a voltage, or a power. Because noise is usually thermal, and therefore not related to the applied voltage. Decibels are always a ratio; either to a reference (as in dBV, dBm, dBA, etc), or as a ratio of input to output, when it's just dB (as in an amp or attenuator).

I think you need to look at your numbers again. -12dB is not 0.25uV/V; you've got a micro in there that is not right. -12dB would be 0.25V/V. 0.25uV/V would be -132dBV, or -12dBuV.

The fact that you have positive dB figures for noise is what alerted me; a positive dB is a gain...

I'm not trying to score points or stupid stuff like that; just trying to be clear on terminology.

And I wholeheartedly agree that CC and even CF resistors should be replaced with MF devices.
 
Noise is normally measured against an absolute reference, be it a voltage, or a power. Because noise is usually thermal, and therefore not related to the applied voltage. Decibels are always a ratio; either to a reference (as in dBV, dBm, dBA, etc), or as a ratio of input to output, when it's just dB (as in an amp or attenuator).

Shrug. That's how noise is specified. log10 of μVolt / Volt.

I think you need to look at your numbers again. -12dB is not 0.25uV/V; you've got a micro in there that is not right. -12dB would be 0.25V/V. 0.25uV/V would be -132dBV, or -12dBuV.

Here's the calculation:
Given –12 dB of resistor noise for carbon composite the ratio is:
-12 dB = 20 x log10(NV / AV)
log10(NV / AV) = -12 / 20
NV / AV = antilog10(–0.6)
NV / AV = 0.25 μVolt / Volt

The ratio is given as μVolt / Volt, so the scaling is already present.

Where is the mistake in the math? -12/20 = -0.6, 10^-0.6 is 0.25. (Check: 10^-1 = 0.1, 1-^-0.5 is about 0.3, so that's right.)

The fact that you have positive dB figures for noise is what alerted me; a positive dB is a gain...

Arrgh! You are correct! That should be –6. Keypunching error replicated it on the cut and paste.

In the future, instead of WASTING MY TIME NEEDLESSLY EXPLAINING SOMETHING TO YOU SO YOU CAN SHOW OFF, why not say, I think you missed a negative sign on the first one, and I would have corrected it.
 
Then you need to specify what it's referred to, which is 1uV, so it's -12dBuV.

I copied it out of a table from my notes, which did not reproduce, and I ended up keypunching it and then just replicated the numbers in the calculations not realizing that the sign was wrong. At that point it was just crunching math.

As I said, I'm not trying to show off; not my thing.

Yeah, sure. You knew the sign was wrong and didn't correct it which caused more work.

Next time just correct the error instead of asking questions which look like you didn't take EE 101.
 
You knew the sign was wrong and didn't correct it which caused more work.

No, I didn't. I didn't know what you meant with the figures you quoted.

All I asked was "Just for completeness, dB relative to what?"

Your answer needed only to have been "one microvolt", and I'd have considered that a perfect clarification. I didn't need you to explain to me what a decibel is, or how to calculate it.

I'll leave you to it.
 
I just have a pile of NOS resistors and would like to classify what is what. I'm sure I have a lot of old NOS CC resistors, maybe should keep them separate unless I find a situation where the value isn't as critical.

Not to mention I have a lot of CC resistors with 10% tolerance (silver band)... Yow.
 
Further to my prior posts in this thread, I've been reflecting on my Pioneer receivers; I had refurbished an SX-434 (low-mileage, very clean, stored in dry conditions for many years) and was VERY impressed with the little devil after new transistors and good quality caps throughout. I didn't change the resistors, I left the carbon films alone. I liked the SX-434 so much that I had to finish refurbishing my SX-1010 to see how it compared. The SX-1010 is a bugger to work on, I did a full recap and really overhauled the power supply. New transistors, LED lamps, as much as I could renew, I did. But not the carbon film resistors.

After all the work, the SX-1010 sounded good - but no better than it's baby brother! I guess I'm going to have to do the resistors and see what that does. And the myriad ceramic caps in the tuner section have to go.
 
I'm sure I have a lot of old NOS CC resistors, maybe should keep them separate unless I find a situation where the value isn't as critical. Not to mention I have a lot of CC resistors with 10% tolerance (silver band)... Yow.

NOS? Sell them to guitar players at huge markup, purchase a pile of metal film instead.
 
Back
Top Bottom