Hi gino,
The following are the resistors that are worth looking into, in no particular order:
1) I/V
2) Feedback
3) Cathode
4) Signal input/output
IMO, resistors matter, even those in preamps. Just ask anyone who uses TX2575 resistors in their DCB1.
The audio chain goes from CD/LP to sound.
Which means a transformance from a physical state to an electrical signal, then an amplification stage, after which the signal has to be shape-shifted into sound.
The transformance stages at the extremities of the process are the most difficult.
Hence the extremely large differences in cost between a bottom rank CD playback module and e.g. a CDM12-pro/CD-Pro2m, or between various phono cartridges, mine does about $3k.
Same story at the other end of the chain gang, from a $100 loudspeaker to a $1M Kharma Enigma set, or a >$1M Elliptica home cinema.
(made in Breda, where Jan Didden first sucked an udder, me 5 miles down the road)
Between plastic and sound, there's just an electrical signal, which has to be amplified.
The further the signal from the source goes through the amplification process, the stronger it becomes.
Stronger means less sensitive, the source output signal is therefore the most delicate.
Implies that for the amplifying, the order of importance is also DAC/phono-pre - then preamplifier - power amp last.
What follows is that transformer choice for a preamp is much more vital than for a power amp, as goes for the filtering level, same same with various other bits and pieces.
On top of that, the preamp has to deal with input selection, volume attenuation. (more so with output selection, mute, mono/stereo switching, or even equalisisation)
Means a power amp has a relative easy task set out.
Which also corresponds with retail price levels of commercial gear, not uncommon is that the preamp costs more than the matching power amp.
When it's the other way round, it's due to a massive hardware display in the power amps.
(my g/f had to get her blood checked and returned with $7k from 2 hours casino, so I was in bad need of a frustration write-off after 45m of JappadiJap. mea culpa)
The line stage, if present, has by far the easiest job to do so it should be the easiest to make completely transparent. Signal levels are high enough that noise is not too difficult an issue. Signal levels are low enough that distortion is not too difficult an issue. Where needed (which is the minority of systems) all it has to do is provide a little voltage gain (with the emphasis on little) and lowish output impedance (so it can drive a long cable into a non-linear power amp input impedance). If no extra gain or driving ability is needed then no line stage is needed.
Having said that, some people seem to like to include a line stage which modifies the signal by adding some low order distortion or shaping the frequency response or adding a little noise. Nothing wrong with that, except that they often claim that this modification is actually restoring 'musicality' - replacing something which the evil digital engineers have removed. Complete nonsense, but widely believed.
Having said that, some people seem to like to include a line stage which modifies the signal by adding some low order distortion or shaping the frequency response or adding a little noise. Nothing wrong with that, except that they often claim that this modification is actually restoring 'musicality' - replacing something which the evil digital engineers have removed. Complete nonsense, but widely believed.
Mr Wayne,
I have a favorite string color, by no means implies that I'm not off my rocker, just reaches the top of the light house faster.
On the other hand, I scored A+ levels at a Navy's hearing test in straightjacket surroundings.
Even decades later, I can still hear the feet of a bug thump at 3ft distance, and suffer from insomnia while auditing a mouse do the runabout.
Oddly enough, my favorite flavors are usually back-evidenced by numbers, and by pro candy store operators.
Still, if I felt like adressing an audience on the matter at hand, I'd feel obliged to fold a newspaper hat first.
I have a favorite string color, by no means implies that I'm not off my rocker, just reaches the top of the light house faster.
On the other hand, I scored A+ levels at a Navy's hearing test in straightjacket surroundings.
Even decades later, I can still hear the feet of a bug thump at 3ft distance, and suffer from insomnia while auditing a mouse do the runabout.
Oddly enough, my favorite flavors are usually back-evidenced by numbers, and by pro candy store operators.
Still, if I felt like adressing an audience on the matter at hand, I'd feel obliged to fold a newspaper hat first.
Here's an example from 1957:
http://www.tubecad.com/2007/12/23/The Sad Tale of a Half-Watt Resistor.pdf
Oh, I thought everyone knew that antique black body carbon composition resistors, specifically those made before 1930, were the best sounding in the world...
not me I am sorry
I like these ..
and similar .. zero noise things
Hi gino,
The following are the resistors that are worth looking into, in no particular order:
1) I/V
2) Feedback
3) Cathode
4) Signal input/output
IMO, resistors matter, even those in preamps.
Just ask anyone who uses TX2575 resistors in their DCB1.
Hello ! thank you for the kind and valuable advice
As I said I tend to trust my friend
The cost is quite limited all considered
By the way i think that some cheap industrial parts are better than other
It is difficult to spot them
Hello !
Thanks for the advice
Actually the signal that arrives at the pot of the preamp is very weak
I should have declared that I prefer solid state equipment
and very often I see people using tube line stages with solid state power amps
So I tend to think that to build a nice ss line stage is not the easiest of the tasks
Regards,
bg
Instead of arguing whether resistors sound different, why not identify the factors in their composition that could lead to different sound?
[Directed sort of at everyone, including the respected Fahey]
For instance - I have actually measured nonlinear resistance (voltage and polarity dependent) of a few percent over reasonable current/voltage rages in especially cheap, or, especially old "NOS" composition resistors. Its real. And not linear (though mostly monotonic toward lower resistance at higher current flow, bidirectionally).
And, in the old days, there wasn't necessarily a standard that called for winding-reversal in wirewound resistors (giving them considerable inductance). For that matter, modern film resistors often have a vexing amount of inductance precisely for these same reasons: the spiral of film, usually laser cut and/or trimmed, all goes one way. No magnetic flux cancellation. Inductance.
Both the wirewound and spiral-cut resistors also seem to have capacitance beyond what one would expect in a 2 terminal "resistor" device. Again, like with the windings of a transformer or inductor, the turn-to-turn spacing, the dielectric path and the voltage-potential across the turns ... yields distributed capacitance. One might hope that the inductance (which would cause an increase in 'resistance' (impedance) with frequency), would be counteracted by the distributed capacitance (decreasing impedance with increasing frequency) in some nice, linear way. Well ... that seems to be something that none of the manufacturers hint at.
Back to the voltage and polarity dependent non-linear resistance. It appears in all cases to be non-reactive ... i.e. neither increases nor decreases the HF body impedance of the devices that exhibit the behavior. Instead, and especially with higher value carbon composition resistors, grains that aren't nominally touching each other (thus insulated), may well be close enough that when the potential voltage between them exceeds "X", they can then enter into conduction by way of quantum tunneling. Its real, and it can be measured.
Irritatingly, the most important "sound" that one "hears" from long (and somewhat ridiculous, but of course entertaining) stints of replacing all the resistors in a quality amplifier with "better ones", is that all the operating points of all the active devices are all slightly changed. Then, by gum, the thing sounds different! So one can "prove" that all them consarned resistors were the culprit. And like the pride that comes from digging a long ditch and filling it with concrete, once the dozens of painstaking hours are over, that amplifier sounds better, doesn't it?
More precise, better bass response, tighter midrange, more presence in the sound stage, clearer definition of passages that were 'hopelessly muddled' before. Fast treble response, agile dynamic range, and of course more restrained treatment of wild passages with huge dynamic range.
Ummm... OK. That ditch is still just a ditch, and the pride-of-work thing taints all.
Don't forget though - the nonlinear and frequency-domain characteristics of resistors are very real, and easily measured with 4.5 digit precision measurements.
GoatGuy
[Directed sort of at everyone, including the respected Fahey]
For instance - I have actually measured nonlinear resistance (voltage and polarity dependent) of a few percent over reasonable current/voltage rages in especially cheap, or, especially old "NOS" composition resistors. Its real. And not linear (though mostly monotonic toward lower resistance at higher current flow, bidirectionally).
And, in the old days, there wasn't necessarily a standard that called for winding-reversal in wirewound resistors (giving them considerable inductance). For that matter, modern film resistors often have a vexing amount of inductance precisely for these same reasons: the spiral of film, usually laser cut and/or trimmed, all goes one way. No magnetic flux cancellation. Inductance.
Both the wirewound and spiral-cut resistors also seem to have capacitance beyond what one would expect in a 2 terminal "resistor" device. Again, like with the windings of a transformer or inductor, the turn-to-turn spacing, the dielectric path and the voltage-potential across the turns ... yields distributed capacitance. One might hope that the inductance (which would cause an increase in 'resistance' (impedance) with frequency), would be counteracted by the distributed capacitance (decreasing impedance with increasing frequency) in some nice, linear way. Well ... that seems to be something that none of the manufacturers hint at.
Back to the voltage and polarity dependent non-linear resistance. It appears in all cases to be non-reactive ... i.e. neither increases nor decreases the HF body impedance of the devices that exhibit the behavior. Instead, and especially with higher value carbon composition resistors, grains that aren't nominally touching each other (thus insulated), may well be close enough that when the potential voltage between them exceeds "X", they can then enter into conduction by way of quantum tunneling. Its real, and it can be measured.
Irritatingly, the most important "sound" that one "hears" from long (and somewhat ridiculous, but of course entertaining) stints of replacing all the resistors in a quality amplifier with "better ones", is that all the operating points of all the active devices are all slightly changed. Then, by gum, the thing sounds different! So one can "prove" that all them consarned resistors were the culprit. And like the pride that comes from digging a long ditch and filling it with concrete, once the dozens of painstaking hours are over, that amplifier sounds better, doesn't it?
More precise, better bass response, tighter midrange, more presence in the sound stage, clearer definition of passages that were 'hopelessly muddled' before. Fast treble response, agile dynamic range, and of course more restrained treatment of wild passages with huge dynamic range.
Ummm... OK. That ditch is still just a ditch, and the pride-of-work thing taints all.
Don't forget though - the nonlinear and frequency-domain characteristics of resistors are very real, and easily measured with 4.5 digit precision measurements.
GoatGuy
Here are the distortion measurements for an Ohmite carbon composition resistor and a Resista brand one.
Both are measured at 1 Khz and the test tone nulled as much as possible. 0 dB reference is 15.8 volts the resistors under test have 1/2 of that across them or in other words the power is 1/16 watt.
For full details see Linear Audio Vol. 1.
Attachments
almost ,,, extremely low I mean
Better look up "Johnson Noise." Fundamental property of the universe.
It is the excess noise that resistors make when a voltage is applied that is why some resistors are quieter than others. If you look at the Ohmite curve you will notice a widening at the base of the 1 kHz spike, also the base noise level is higher and rises as the frequency goes down. That is caused by noise. The better resistors don't do that.
The poster still needs to understand that "low noise resistor" has to be a qualified term- understanding the basics of Johnson vs excess noise is critical.
Carbon comp is a lousy choice for most signal circuits where voltage is dropped, that's pretty well known. See the article I linked to from 1957. They're fine for stoppers in power amps, but no-one on the sane side of engineering uses them for, say, feedback resistors.
Decent quality (cheap name brand) metal film have very low excess noise. Wirewound is even lower, but we're splitting very fine hairs at that point.
Carbon comp is a lousy choice for most signal circuits where voltage is dropped, that's pretty well known. See the article I linked to from 1957. They're fine for stoppers in power amps, but no-one on the sane side of engineering uses them for, say, feedback resistors.
Decent quality (cheap name brand) metal film have very low excess noise. Wirewound is even lower, but we're splitting very fine hairs at that point.
- Status
- Not open for further replies.