Hi,
That miniaturization process has been going on for about a decode now.
But did you check you were operating those resistors within the specced voltage range?
Cheers,
P.S. @Wavebourn: a good place to find out what a resistor sounds like is in the GNFB-loop indeed.
Hard to fault those VSH bulkfoils.
astouffer said:
That miniaturization process has been going on for about a decode now.
But did you check you were operating those resistors within the specced voltage range?
Cheers,
P.S. @Wavebourn: a good place to find out what a resistor sounds like is in the GNFB-loop indeed.
Hard to fault those VSH bulkfoils.
Hi,
Well, a large group of resistors in a stepped attenuator would or should have an audible effect provided you'd build a series ladder.
One of the worst of its kind BTW.
There most certainly is a BIG audible difference between cheap plastic track pots and ALPS pots and even within ALPS range of pots there is a marked difference between the various models offered.
Oh, it's not a matter of L/R balance either although the more expensive ones are usually better calibrated though.
Just trust your ears, don't expect anything and.....listen.
Cheers,
Conrad Hoffman said:Frank, along those lines I think if there's anyplace that a large group of resistors makes a difference, it's stepped attenuators. Still, I don't trust my ears because of that expectation bias. It's too hard for me to accept that a cheap Radio Shack pot sounds the same as an Alps pot or a high quality resistive divider chain soldered to a silver contact switch. The sound difference might just be the accuracy of balance, but given my uncertainty and inability to measure, I go with the stepped attenuator.
Well, a large group of resistors in a stepped attenuator would or should have an audible effect provided you'd build a series ladder.
One of the worst of its kind BTW.
There most certainly is a BIG audible difference between cheap plastic track pots and ALPS pots and even within ALPS range of pots there is a marked difference between the various models offered.
Oh, it's not a matter of L/R balance either although the more expensive ones are usually better calibrated though.
Just trust your ears, don't expect anything and.....listen.
Cheers,
Hi,
That big string's probably even worse than a single one of the same value.
Nothing daft about shunt attenuators really.
You'll need to figure out a few parameters to see if it fits the circuit impedance-wise, use a make before break switch, but other than that?
My hearing isn't what it used to be either but then it doesn't have to be to detect differences unless they happen at the very high end of the audio spectrum and then some...
Cheers,
Conrad Hoffman said:So, do you think a string of resistors soldered together is the sonic equal of a single resistor of the same total value? I've always thought the logic behind shunt attenuators was a bit daft. But who am I to judge; my hearing was better 30 years ago.
That big string's probably even worse than a single one of the same value.
Nothing daft about shunt attenuators really.
You'll need to figure out a few parameters to see if it fits the circuit impedance-wise, use a make before break switch, but other than that?
My hearing isn't what it used to be either but then it doesn't have to be to detect differences unless they happen at the very high end of the audio spectrum and then some...
Cheers,
fdegrove said:
When I was a kid I damaged lots of them!!!
They were like small fabric rugs made of manganine, or constantane, or nichrome wire, in lacquer and silk insulation. Such rugs were wrapped around Bakelite spools with 4 legs each.
...I want a time machine! Badly! I damaged high precision hermetically sealed in copper boxes silver mica caps, in nanofarade to tens of nanofarade ranges! Also, I damaged many line transformers with permalloy and litz in silk wires, some of them were hermetically sealed in copper boxes.
Shame on me!
How is the quality of surface mount resistor
Seem there are many kinds, and difficult to choose
Seems in some places they would be practical to use, even with a mostly hole through design
In picture R5/R9 ordinary resistors are good to do the needed "jump"
But C14 could be a SMD
Well, here its a cap, but Im sure it could just as well have been a resistor in other places
Seem there are many kinds, and difficult to choose
Seems in some places they would be practical to use, even with a mostly hole through design
In picture R5/R9 ordinary resistors are good to do the needed "jump"
But C14 could be a SMD
Well, here its a cap, but Im sure it could just as well have been a resistor in other places
Attachments
All this talk about resistors is tempting me to get on my soap box. It is not correct to assume that since there exists a linear relation called ohm's law, that every conductor in every situation will conform strictly to that linear relation. In fact, I consider it somewhat miraculous that enough materials conduct in a sufficiently linear fashion for such a linear relationship to be deduced in the first place. For a prominent counterexample, look at electrical conduction in gases. Ohm's law is a useful first order approximation that doesn't work for a lot of materials and situations (this can be useful or annoying, depending on what you're trying to accomplish).
There are several ways to approach making something resistive ( hopefully, in a linear fashion).
1) Take something not too resistive to begin with, and make it long and thin (wirewound resistors, and to some extent, metal film resistors).
2) Take something that is not too terribly resistive, grind it to a fine powder, mix it in various proportions with a powdered non-conductor, and crunch it together so that the conduction path is through random grains of conductive material touching throughout the body of the composite (carbon comp resistors). This can be a nest of snakes, as the material may be microphonic or even arc randomly between grains, injecting a lot of excess noise. If there's arcing, then the conductivity can be affected by voltage gradient, especially if the conductor is a little bit non-linear to begin with (carbon? what sort of carbon? ( there's lots of ways to make carbon, and the properties can depend on the method of manufacture), how many/what impurities, etc)
3) Start with something not too terribly conductive to begin with, and vary thickness and length by various techniques (deposition thickness, spiral grooving). Be careful! The material you choose may be a semiconductor to a certain extent, and exhibit variations in conductivity as a function of voltage gradient, temperature, etc. Carbon film and metal oxide resistors fall into this category.
Even if the material you start with is linear, the method of deposition can affect the properties and quality of a resistive film (sputtering? thermal vapor deposition?, chemical deposition? What sort of buffer gas is used? Even deposition rate can vary the properties of the deposited material. Is the substrate heated? All this falls under trade secrets, and can differentiate one manufacturer's product from another, even if they start with similar materials.
Add picky points such as to how one interfaces the resistive material to the wire leads, lugs, etc. (end caps, nail-head turrets, or simply poking the lead wires into the resistive material as was done with the cheaper carbon comp resistors), other quality factors such as moisture absorption due to improper encapsulation, and you have just a whole host of things that can affect the performance of a given resistor. One thing seems fairly clear - the better sounding resistors use a linear bulk conductor rather than a semiconductor or composite. The practice of using a larger resistor than necessarily called for by the power dissipation in a given application can also reduce non-linear effects due to both voltage gradient and thermal modulation. I won't even get into magnetic vs. non-magnetic leads and end caps.
The practice of downsizing components and giving them inflated power ratings really gets up my nose. I run into that at work all the time, and try to buy resistors from vendors that size them appropriately for a given power rating. No, a 2W resistor cannot be made the size of a 1/2W resistor without it getting really hot - the copper leads (if they are copper) degrade, the solder joints can crystallize and become unreliable, ,and the hot resistor can damage the circuit board or adjacent components. The smaller resistor will have a higher voltage gradient across it, magnifying any effects due to voltage coefficient of resistance, a real consideration for plate load resistors.
In short, I don't find it too surprising that resistors can alter the sound of a given circuit.
There are several ways to approach making something resistive ( hopefully, in a linear fashion).
1) Take something not too resistive to begin with, and make it long and thin (wirewound resistors, and to some extent, metal film resistors).
2) Take something that is not too terribly resistive, grind it to a fine powder, mix it in various proportions with a powdered non-conductor, and crunch it together so that the conduction path is through random grains of conductive material touching throughout the body of the composite (carbon comp resistors). This can be a nest of snakes, as the material may be microphonic or even arc randomly between grains, injecting a lot of excess noise. If there's arcing, then the conductivity can be affected by voltage gradient, especially if the conductor is a little bit non-linear to begin with (carbon? what sort of carbon? ( there's lots of ways to make carbon, and the properties can depend on the method of manufacture), how many/what impurities, etc)
3) Start with something not too terribly conductive to begin with, and vary thickness and length by various techniques (deposition thickness, spiral grooving). Be careful! The material you choose may be a semiconductor to a certain extent, and exhibit variations in conductivity as a function of voltage gradient, temperature, etc. Carbon film and metal oxide resistors fall into this category.
Even if the material you start with is linear, the method of deposition can affect the properties and quality of a resistive film (sputtering? thermal vapor deposition?, chemical deposition? What sort of buffer gas is used? Even deposition rate can vary the properties of the deposited material. Is the substrate heated? All this falls under trade secrets, and can differentiate one manufacturer's product from another, even if they start with similar materials.
Add picky points such as to how one interfaces the resistive material to the wire leads, lugs, etc. (end caps, nail-head turrets, or simply poking the lead wires into the resistive material as was done with the cheaper carbon comp resistors), other quality factors such as moisture absorption due to improper encapsulation, and you have just a whole host of things that can affect the performance of a given resistor. One thing seems fairly clear - the better sounding resistors use a linear bulk conductor rather than a semiconductor or composite. The practice of using a larger resistor than necessarily called for by the power dissipation in a given application can also reduce non-linear effects due to both voltage gradient and thermal modulation. I won't even get into magnetic vs. non-magnetic leads and end caps.
The practice of downsizing components and giving them inflated power ratings really gets up my nose. I run into that at work all the time, and try to buy resistors from vendors that size them appropriately for a given power rating. No, a 2W resistor cannot be made the size of a 1/2W resistor without it getting really hot - the copper leads (if they are copper) degrade, the solder joints can crystallize and become unreliable, ,and the hot resistor can damage the circuit board or adjacent components. The smaller resistor will have a higher voltage gradient across it, magnifying any effects due to voltage coefficient of resistance, a real consideration for plate load resistors.
In short, I don't find it too surprising that resistors can alter the sound of a given circuit.
Wrenchone,
OK, you have given an admirable description of the effect of materials, way of manufacture, etc. In that sense one can only agree: One does not find it surprising that resistors are not perfect (as in fact nothing is!).
But in practice we have ready-made resistors of defined ratings, standards and characteristics (and how grateful I am that I do not need to manufacture them from the crusher stage upward!). To go from qualitative to quantative (as has been posted here before), one needs to compare figures with figures where these are available!
Scientists (engineers) do not know everything, but happily by now we do know certain things. Enough is known about hearing characteristics to determine what is inaudible to human hearing (and add say 15% safety factor or whatever to make the paranoid happy) and given likely operating conditions, one can work out what effect imperfections will have.
If now that is more than comfortably 'out of reach' of hearing after every likely circumstance has been accounted for, I believe it is safe to accept that certain things will not affect the performance of, in this case, an amplifier to any audible degree.
That is what this thread is all about. I (and others) have stated before the dangers to keep talking qualitatively about possible deleterious effects, where we have the knowledge to quantify the same. Decades of research by dedicated brains have given us the tools; one should at least gratefully honour that by using them.
OK, you have given an admirable description of the effect of materials, way of manufacture, etc. In that sense one can only agree: One does not find it surprising that resistors are not perfect (as in fact nothing is!).
But in practice we have ready-made resistors of defined ratings, standards and characteristics (and how grateful I am that I do not need to manufacture them from the crusher stage upward!). To go from qualitative to quantative (as has been posted here before), one needs to compare figures with figures where these are available!
Scientists (engineers) do not know everything, but happily by now we do know certain things. Enough is known about hearing characteristics to determine what is inaudible to human hearing (and add say 15% safety factor or whatever to make the paranoid happy) and given likely operating conditions, one can work out what effect imperfections will have.
If now that is more than comfortably 'out of reach' of hearing after every likely circumstance has been accounted for, I believe it is safe to accept that certain things will not affect the performance of, in this case, an amplifier to any audible degree.
That is what this thread is all about. I (and others) have stated before the dangers to keep talking qualitatively about possible deleterious effects, where we have the knowledge to quantify the same. Decades of research by dedicated brains have given us the tools; one should at least gratefully honour that by using them.
Yes, I've accidentally purchased those darn miniaturized power resistors, only to find that if I ran them near their ratings, the solder would melt and the resistor would fall out of the circuit board. Fan cooling or some other attachment method was required. And for goodness sakes, space them well off the circuit board. Also, if you want tempco to be a minor factor, and the reliability to be high, don't let the temperature change much!
I asked the resistor string question because I've had many instances where a string of resistors was technically better in every regard. More surface area to dissipate power with less temperature rise in case tempco is an issue. Lower stray capacitance across the resistors by getting the terminals further apart. Lower voltage drop across each resistor in case voltage coefficient was an issue. Higher voltage handling capability. The ability to use a more desirable resistor type, regardless of power and voltage. The only downside might be higher inductance, but that seems to be the least important stray term associated with resistors, being so small to begin with.
I'm not a fan of smt parts- too small to see! I've built very high gain amplifiers where the noise of thick film resistors was obvious. Metal film was fine.
I keep coming back to ordinary 1/4W 1% metal film resistors as having no objectionable parameters for audio use, provided they're operated at low power levels. I buy 'em from Digikey 200 at a time and often match them if needed. For convenience I often wind my own low value resistors from Manganin or Constantan. The sound fine as well, but don't kid yourself into thinking "non-inductive" wire wounds are really non-inductive. Perfect cancellation of inductance is near impossible to achieve, and comes at the cost of high capacitance. Still, the stray terms are negligible compared to the low values of resistance where you'd use a wire wound anyway. The only place I'd use caution is a feedback network because the amplifier needs the ratio to remain correct far beyond the audio region for stability. That's the place for those bulk metal resistors.
I asked the resistor string question because I've had many instances where a string of resistors was technically better in every regard. More surface area to dissipate power with less temperature rise in case tempco is an issue. Lower stray capacitance across the resistors by getting the terminals further apart. Lower voltage drop across each resistor in case voltage coefficient was an issue. Higher voltage handling capability. The ability to use a more desirable resistor type, regardless of power and voltage. The only downside might be higher inductance, but that seems to be the least important stray term associated with resistors, being so small to begin with.
I'm not a fan of smt parts- too small to see! I've built very high gain amplifiers where the noise of thick film resistors was obvious. Metal film was fine.
I keep coming back to ordinary 1/4W 1% metal film resistors as having no objectionable parameters for audio use, provided they're operated at low power levels. I buy 'em from Digikey 200 at a time and often match them if needed. For convenience I often wind my own low value resistors from Manganin or Constantan. The sound fine as well, but don't kid yourself into thinking "non-inductive" wire wounds are really non-inductive. Perfect cancellation of inductance is near impossible to achieve, and comes at the cost of high capacitance. Still, the stray terms are negligible compared to the low values of resistance where you'd use a wire wound anyway. The only place I'd use caution is a feedback network because the amplifier needs the ratio to remain correct far beyond the audio region for stability. That's the place for those bulk metal resistors.
.... and space them far away from electrolytic capacitors! How often have I seen a poor electrolytic fried between two heat-radiating resistors.
Yes, good post Conrad. I would just add that, yet again one keeps hearing about 'inductance - beware!' Sorry to sound like a broken record - but what is the Q of the final result?? As you say, feedback circuits ... But that is super-audio, and the result is under one's control; add a freaking capacitor wherever if needs be for stability; one probably has to do that anyway.
Yes, good post Conrad. I would just add that, yet again one keeps hearing about 'inductance - beware!' Sorry to sound like a broken record - but what is the Q of the final result?? As you say, feedback circuits ... But that is super-audio, and the result is under one's control; add a freaking capacitor wherever if needs be for stability; one probably has to do that anyway.
Morgan Jones page 204. Resistors, Valve amps 3rd ed.
When using more recent M/O or M/F types; I always use a 3W rated device for a 1W application or far better CCS it.
My favourite place for not using a power resistor is the cathode of a Williamson diff driver. I do like this technique because it's a good responder.
A 15dB reduction of even order triode harmonics is possible when reasonable matched tubes and a dynamic balancing pot is used part of anode load. The 3rd harmonic remains the same. When preceeded with a split-load phasesplitter and near matched tubes; the signal at the cathodes can be near equal amplitude 2F; the AC balancing often coincides with a reduction of the 2nd &4th harmonics. This near perfect cancellation is my claim is the CCSink in the cathode actually improves the sound quality of the Williamson diff driver as it begins to behave similiaries to the true LTP coupling.
When replacing the resistive anode loads with a CCSource, the near infinite AC resistance can make high gm tubes oscilllate due to non resistive damping. So a resistor shouldn't be totally decoyed.
Some users might say why bother using a CCsink in the Williamson diff driver, that the p-p o/p stage cancels the even order harmonics. True, but I am trying to quantify that replacing a resistor in the vital regions can actually change the sound quality.
For those using a scope with an FFT; the analysis is just beginning.
richy
When using more recent M/O or M/F types; I always use a 3W rated device for a 1W application or far better CCS it.
My favourite place for not using a power resistor is the cathode of a Williamson diff driver. I do like this technique because it's a good responder.
A 15dB reduction of even order triode harmonics is possible when reasonable matched tubes and a dynamic balancing pot is used part of anode load. The 3rd harmonic remains the same. When preceeded with a split-load phasesplitter and near matched tubes; the signal at the cathodes can be near equal amplitude 2F; the AC balancing often coincides with a reduction of the 2nd &4th harmonics. This near perfect cancellation is my claim is the CCSink in the cathode actually improves the sound quality of the Williamson diff driver as it begins to behave similiaries to the true LTP coupling.
When replacing the resistive anode loads with a CCSource, the near infinite AC resistance can make high gm tubes oscilllate due to non resistive damping. So a resistor shouldn't be totally decoyed.
Some users might say why bother using a CCsink in the Williamson diff driver, that the p-p o/p stage cancels the even order harmonics. True, but I am trying to quantify that replacing a resistor in the vital regions can actually change the sound quality.
For those using a scope with an FFT; the analysis is just beginning.
richy
astouffer said:
I believe I got them from Mouser. Since these were to be plate load resistors, I didn't believe they'd stand up to the voltage without that flash-over possibility. I sent 'em back and got ones specifically rated to withstand that voltage. Not gonna trust these tiny resistors to stand up to nearly 200Vdc even if the power rating was twice the proposed operating dissipation.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.