I recently bought a venerable Rotel RA-820BX4 for my PC-audio rig.
I needed a system which will 'work' at low volumes, that is at small scale. If I were to play music at 'realisitc' levels I'd get evicted from my flat.
So - a pair of JPW mini-monitor-golds (on DIY solid timber pillar stands) and the Rotel, which is 'class A/B' - it uses an array of output transistors (4 per channel) and delivers the few few watts in class-A. This is confirmed by the BIG heatsinks and the heat that comes off them even when the amp is idling.
Many (most, in fact) integrated amps don't really sing til the wick is turned up - I found this with an Arcam Alpha 3 I owned a few years back, powering an identical pair of JPW's (given to friend who's still using them).
HERE'S THE THING.
I didn't have any 'speaker-cable' to hand on setting up, so I cannablized an good 'OFC' interconnect. Very thin multi-strand.
Well, it works as I hoped it would - at low volume levels I still get 'scale', detail and ambience from recordings. Result.
But today, when I knew my neighbours were out, I cranked it up. it got loud, for sure, the Rotel is rated at 30+30 RMS into 8ohms, and double that into 4 ohms ....
BUT it got, to put it simply, 'congested' and not a simple scaling up of the delicious soundstage and detail I had been listening to at the diminutive levels I've got accustomed to in the last few weeks.
This was listening to Supertramp's 'Crime Of THe Century' and Stereolabs's 'Margerine Eclipse'.
I suspect that the cable I'm using, the chopped up interconnect, has a bearing on this.
If I were to use 'proper' 15-amp-of-mains-current speaker cables, I think I'd improve the sound for party-level volume levels like I tried this afternoon, but I then likely lose the pleasing small-scale sound-stage and detail I hear (and need) in day-to-day listening.
So - speaker cables - worth some thought when setting up your rig, and you're not intending on declaring sonic-war on your neighbours.
Rainer.
Edit - So what's my point? That using thin, high resistance, (low capacitance?) wire (of good quality - hence interconnect cable) for driving speakers, might serve many peoples' purposes where getting 'good' sound at low in-room levels is an issue.
Using interconnect as speaker cable is very dangerous. Cheap interconnects often develop short circuits, and good ones can do if they are DYIed. And a short circuit can mean the death of the amplifier.
The two parameters that significantly affect the sound of a speaker cable are resistance and capacity. Both will tend to reduce treble.
A big cable is as capable of faithfully transmitting a very weak signal as a small one. If interconnect sounds softer to your ears, maybe you speakers have a treble response that is agressive. That can be the case if they are set on your side rather than in front of you.
Edit : being coaxial, an interconnect will usually have a higher capacity than a speaker cable.
Using interconnect as speaker cable is very dangerous. Cheap interconnects often develop short circuits, and good ones can do if they are DYIed. And a short circuit can mean the death of the amplifier.
The two parameters that significantly affect the sound of a speaker cable are resistance and capacity. Both will tend to reduce treble.
A big cable is as capable of faithfully transmitting a very weak signal as a small one. If interconnect sounds softer to your ears, maybe you speakers have a treble response that is agressive. That can be the case if they are set on your side rather than in front of you.
Edit : being coaxial, an interconnect will usually have a higher capacity than a speaker cable.
[a href="index.php?act=findpost&pid=307622"][{POST_SNAPBACK}][/a]
I agree, care is needed, but i mentioned using
good quality wire, and of course care will be required in the terminations..
But you're incorrect regarding capacitance, it blocks DC, and hence is used in crossover networks as a basic high-pass for tweeters. The simplest 2-way cross-over consists of nothing but a capacitor in series with the tweeter.
I'm suggesting (to repeat) that amp/speaker systems intended for low-volume in-room listening might benefit from the use of thin, high resistence, low capacitance wire , which good interconnects are, as speaker cables.
At high volumes, considerable current flows as the amp drives the speakers and, as I said, such wire will probably not work well. This seems actually to be my experience so far,
Rainer.
edit - my prose is apalling.
edit 2 - I mean
appalling. sigh.
Of course, the AB I need to try is the wire I'm using against a big multi-strand cable such as mains cable, QED whatever.
This is easier said then done, as usually I listen to music when I want to listen to music, and for no other reason. I like to think I have a grasp of the basics/theory and can set the kit up right from the get-go.
I really don't enjoy the manic swapping of components in and out and then tryng to compose myself for 'the test'.
R.
Heany gauge zip cord will do the trick. If you think that expensive "sound pipe" and similar wire is any different, go for it. Then try to ABX the difference between that and heavy gauge zip cord. I have never seen any demonstration shown or quoted in 40+ years that can point to "better sound" from more expensive wires.
If you want to be really scared, look inside the amps. Skinny wire there.
Heany gauge zip cord will do the trick. If you think that expensive "sound pipe" and similar wire is any different, go for it. Then try to ABX the difference between that and heavy gauge zip cord. I have never seen any demonstration shown or quoted in 40+ years that can point to "better sound" from more expensive wires.
If you want to be really scared, look inside the amps. Skinny wire there.
[a href="index.php?act=findpost&pid=307650"][{POST_SNAPBACK}][/a]
what?
R.
The two parameters that significantly affect the sound of a speaker cable are resistance and capacity. Both will tend to reduce treble.
...
Edit : being coaxial, an interconnect will usually have a higher capacity than a speaker cable.
[a href="index.php?act=findpost&pid=307622"][{POST_SNAPBACK}][/a]
But you're incorrect regarding capacitance, it blocks DC, and hence is used in crossover networks as a basic high-pass for tweeters. The simplest 2-way cross-over consists of nothing but a capacitor in series with the tweeter.
[a href="index.php?act=findpost&pid=307634"][{POST_SNAPBACK}][/a]
Someone has to say it.... Aughh! Ever study AC circuits?
Yes, a capacitor blocks DC and passes high frequencies. The capacitance in a speaker cable is between the 2 lines (red/black, +/-, whatever), not between the input and output. So this *parallel* capacitance acts as a low pass filter by allowing some high frequency signal to go from + to - without going through the speaker.. How much effect it has depends on other things like the output impedance of the amp, resistance of the wire, and input impedance of the speaker.
And to confuse you further, do not confuse resistance or capacitance with "characteristic impedance", in ohms, which is a parameter of transmission lines that results in a lowpass filter effect in the megahertz range that is generally not considered relevant to audibility.
.... and do not confuse "characteristic impedance" with "impedance", in ohms, a complex value formed by the resistance (real axis) and capacitance/inductance (imaginary axis), which can be used to simplify the calculation of circuit/filter solutions and is applicable at all frequencies.
The two parameters that significantly affect the sound of a speaker cable are resistance and capacity. Both will tend to reduce treble.
...
Edit : being coaxial, an interconnect will usually have a higher capacity than a speaker cable.
[a href="index.php?act=findpost&pid=307622"][{POST_SNAPBACK}][/a]
But you're incorrect regarding capacitance, it blocks DC, and hence is used in crossover networks as a basic high-pass for tweeters. The simplest 2-way cross-over consists of nothing but a capacitor in series with the tweeter.
[a href="index.php?act=findpost&pid=307634"][{POST_SNAPBACK}][/a]
Someone has to say it.... Aughh! Ever study AC circuits?
Yes, a capacitor blocks DC and passes high frequencies. The capacitance in a speaker cable is between the 2 lines (red/black, +/-, whatever), not between the input and output. So this *parallel* capacitance acts as a low pass filter by allowing some high frequency signal to go from + to - without going through the speaker.. How much effect it has depends on other things like the output impedance of the amp, resistance of the wire, and input impedance of the speaker.
[a href="index.php?act=findpost&pid=307660"][{POST_SNAPBACK}][/a]
So, what are your recommendations for speaker cable for a given application?
Because it seems to me that, unlike any other part of the signal path, speaker cables are expected to carry audio signals at a huge range of currents, and I don't think a single wire can 'do it all'.
In setting a system up it seems obvious to me that you have to decide whether you want those conductors to work at fractions of an amp (to deliver a believable soundscape in your little flat at sound pressure levels that won't get you evicted) , or many amps for palpitating bass at party levels
Which is to be? I want a system that will work at fractions of an amp - that is a couple of watts at the most. I think I should use an approriate wire, and that isn't going to be a cable that looks like it's designed to power an arc-welder.
It's a kinda intuitive thing, but if you think a cable as thick as your wrist is 'all things to all men', go for it.
R.
The argument that larger cables don't work as well for small currents is sort of specious. Maybe if you thought of the cable as a tank of electrons, and the amp has to work really hard to add/remove electrons for the tank, would that make sense.
Let me go all fundie electromagnetism on you for a second. The amp is short for ampere, which is a unit of electric current defined as coulombs per second. What a current is, in some sense, is electrons moving along a wire so that a certain amount of charge passes through a cross section of wire at a given speed. charge divided by time equals current.
Pay close attention to the word "cross-section". It does not factor into the current calculation. That implies a certain surface area, which is related to the thickness of the conductor. If you double the thickness of the conductor, the surface area of the cross section doubles, but the current does not change. The amplifier only cares about the voltage and current it needs to drive; both of those quantities do not change as the cross section changes. As long as the cable doesn't overheat from being too thin to handle the current, it does not matter how thick it is. It could be as thick as your wrist and your amp could drive it!
Using a big cable for low signals is a bit like using a truck to carry one tomato. But the tomato is delivered in time and at the right location all the same.
In fact, the cable size is determined according to safety. For a given current, the cable must not overheat, which could start a fire. For high currents, a minimal size is given, but there is no maximum size.
However, for speaker cables, there is another consideration : sound. If the cable is too thin, or too capacitive, or too long, it will act as a filter. Resistance will interact with the speaker impedance and change the frequency response. Capacity will act as a lowpass filter because it is in parallel with the speaker (it would be a high pass filter if it was in serial). This does not depend on the amount of current. It occurs at high levels as well as low levels.
Thus it leads to a maximum resistance and maximum capacity for the whole cable, for the sound to be unaltered. That's why thin cables are not a good choice for speaker cable. This is not a problem for interconnect cables, because the impedance of the termination is much higher (47000 Ohm versus 8 Ohm for speakers), which turns the cable impedance (resistance and capacity) neglectible in comparison.
In setting a system up it seems obvious to me that you have to decide whether you want those conductors to work at fractions of an amp (to deliver a believable soundscape in your little flat at sound pressure levels that won't get you evicted) , or many amps for palpitating bass at party levels
Which is to be? I want a system that will work at fractions of an amp - that is a couple of watts at the most. I think I should use an approriate wire, and that isn't going to be a cable that looks like it's designed to power an arc-welder.
[a href="index.php?act=findpost&pid=307691"][{POST_SNAPBACK}][/a]
Hmmm... it was also obvious to you that a cable with lots of capacitance would act as a high pass filter when in fact it is more the opposite.. Don't always go for the obvious.
14 or 16 guage zip cord actually does a pretty good job of getting a signal to the speakers for reasonable (home applicaton) lengths and power.
Reasonable changes (that is, with good cause) from that baseline could be things like:
1) flatter cable to run under carpet
3) smaller for space constraints
4) some adaptation to make it pull easier through conduit
5) more flexible.
6) esthetic considerations like color, or you just think round cables look better.
In setting a system up it seems obvious to me that you have to decide whether you want those conductors to work at fractions of an amp (to deliver a believable soundscape in your little flat at sound pressure levels that won't get you evicted) , or many amps for palpitating bass at party levels
Which is to be? I want a system that will work at fractions of an amp - that is a couple of watts at the most. I think I should use an approriate wire, and that isn't going to be a cable that looks like it's designed to power an arc-welder.
[a href="index.php?act=findpost&pid=307691"][{POST_SNAPBACK}][/a]
Hmmm... it was also obvious to you that a cable with lots of capacitance would act as a high pass filter when in fact it is more the opposite.. Don't always go for the obvious.
14 or 16 guage zip cord actually does a pretty good job of getting a signal to the speakers for reasonable (home applicaton) lengths and power.
Reasonable changes (that is, with good cause) from that baseline could be things like:
1) flatter cable to run under carpet
3) smaller for space constraints
4) some adaptation to make it pull easier through conduit
5) more flexible.
6) esthetic considerations like color, or you just think round cables look better.
[a href="index.php?act=findpost&pid=307800"][{POST_SNAPBACK}][/a]
I must admit I had the nature of capacitance, how it's generated in a audio conductor wrong, and I thank you for the correction. But the fact is that it tends to increase in audio/AC conductors with the guage and number of strands. (although Pio might be right and the close proximity of the hot and earth conductors in coaxial wire might in itself actually create some).
For example, DNM Reson cables are single core with the conductors kept apart by the 'ribbon' cross-section of the insulator, and they are known to be about as low-capacitance as you can get.
Whatever, to repeat, I've found that my ad-hoc use of a reasonable quality signal wire as speaker wire
does seem to work at low volumes (plenty of low-level detail that I wouldn't normally expect to hear) (
edit - that includes 'timbre' and 'ambience'), but conversely (and not unexpectedly) the sound with the volume cranked up (but certainly not anywhere near the Rotels rated/continous) isn't good at all - rather flat sounding.
I owned exactly the same speakers (gave them ot friend who still uses them) with an Arcam Alpha 3 back in the early 90's, so I'm pretty familiar with their sound, and I also remember that combination didn't really 'work' unless the 40+40w Arcam was turned up - so I otherwise used it's tone controls to help compensate.
As I said, I bought the Rotel (which has no tone or 'loudness' controls) because I had a hunch it's design would lend itself better to low volume listening than some amps (including the otherwise pretty good Arcam) and happily it does indeed.
I know I have a couple of lengths of half-decent multistrand that a friend got with Bose system and didn't need, so the answer is to swap the cable and confirm (or not) what I think I hear, although I'm quite content with the sound as is.
Rainer.
I'm afraid multiple strands of cable have no effect at all over capacity, because all strands from each conductor are together and touching, and because capacity happens between positive and negative conductors, not inside each conductor. Coaxial or twisted pair cable does have higher capacitance. Multi-stranded has not. Also, capacitance, or it's effects, don't vary with the amount of output power, because cables are linear devices.
I'm afraid multiple strands of cable have no effect at all over capacity, because all strands from each conductor are together and touching, and because capacity happens between positive and negative conductors, not inside each conductor. Coaxial or twisted pair cable does have higher capacitance. Multi-stranded has not. Also, capacitance, or it's effects, don't vary with the amount of output power, because cables are linear devices.
[a href="index.php?act=findpost&pid=307944"][{POST_SNAPBACK}][/a]
The strands might well be in contact, but the surface of the individual strands forms a slight 'dialectric' or insulator between them. Whether this has any bearing on the measured capactitance of a given cable, I don't know, but generally speaking, single core conductors are measurably less capacitive.
In choosing a wire for a tonearm, and the external run from it's base, it's well known that if high-capacitance stuff is used it can profoundly affect the tonal balance of a cartridge, producing as it is a signal measured in millivolts.
R.
The strands might well be in contact, but the surface of the individual strands forms a slight 'dialectric' or insulator between them.
First time I have notion of such weird phenomena. If there's no actual dielectric between them (they are conductive cables with conductive surfaces and really touching), then there's no dielectric and no possible capacitive effect, even more when, again, capacitive effect happens exclusively between + and - conductors, not inside each conductor. See, even if there was an actual dielectric insulating each strand, there would be no capacitive effect, given that the + and - conductors were far enough from each other. Please check your electric theory.
Whether this has any bearing on the measured capactitance of a given cable
, I din't know, but generally speaking, single core conductors are measurably less capacitive.
[a href="index.php?act=findpost&pid=307961"][{POST_SNAPBACK}][/a]
Any actual proof of that (in form of measurements)?
Edit: more explanations and examples.
The strands might well be in contact, but the surface of the individual strands forms a slight 'dialectric' or insulator between them.
First time I have notion of such weird phenomena. If there's no actual dielectric between them (they are conductive cables with conductive surfaces and really touching), then there's no dielectric and no possible capacitive effect, even more when, again, capacitive effect happens exclusively between + and - conductors, not inside each conductor. See, even if there was an actual dielectric insulating each strand, there would be no capacitive effect, given that the + and - conductors were far enough from each other. Please check your electric theory.
Edit: more explanations and examples.
[a href="index.php?act=findpost&pid=307964"][{POST_SNAPBACK}][/a]
It's not weird at all, each strand has a slightly work-hardened and oxidised surface created as they are drawn thru dies in manufacture, and which is
not perfectly conductive (no material is, apart from 'room-temperature' superconductors which are the holy grail of researchers) , before being brought together as a cable.
It's the reason it's a good idea to 'tin' the ends of multi-strand wire (such as bare speaker cables) with solder to make sure all of them are equally coupled to the terminations at both ends of the run.
R.
It's not weird at all, each strand has a slightly work-hardened and oxidised surface created as they are drawn thru dies in manufacture, and which is not perfectly conductive (no material is, apart from 'room-temperature' superconductors which are the holy grail of researchers) , before being brought together as a cable.
[a href="index.php?act=findpost&pid=307974"][{POST_SNAPBACK}][/a]
That has nothing to do with the existance of a dielectric between strands. Despite they not being perfectly conductive (currently no known material is at regular temperatures), every strand is at same potential, so there's no possible capacitive effect.
The biggest immediate effect is that stranded wire has less area of copper for the same guage due to the gaps between close-packed circular cross sections... so more resistance per foot. So for the same resistance the stranded would be larger and create more capacitance.. but the percentage difference is pretty small. Other effects are that it is more flexable and gets less work hardening than solid for the same amount of flexing. Work hardening will lead to cracks which sort of defeats the purpose.
If you want to check for capacitive effects try Litz wire which actually has insulation between the strands.
If there were a problem with bad contact between strands on the same wire you will have a lot more audible effect with diodes being formed than capacitors.
For example, DNM Reson cables are single core with the conductors kept apart by the 'ribbon' cross-section of the insulator, and they are known to be about as low-capacitance as you can get.
Can I direct you to a couple of books about this stuff? I mean it would really help the conversation if you knew what capacitance is and what affects it. It irks me a bit because for over 10 years my job was mostly about analyzing small scale parasitic (unintentional) capacitance. Even with that it all starts with the basics, not ad copy or a review in some high end audio magazine.
The main reasons those cables are low capacitance are a) the conductors are relatively far apart, and b) the dielectric is mostly air.
Plus copper (I) oxide (CuO) is a semiconducter and thus not a very good dielectric material. So even if there was a potential in different strands there would not be any signifigant capactience.
Terminal ends of the wires are tinned mostly for strength, secondly for oxidation.
In choosing a wire for a tonearm, and the external run from it's base, it's well known that if high-capacitance stuff is used it can profoundly affect the tonal balance of a cartridge, producing as it is a signal measured in millivolts.[a href="index.php?act=findpost&pid=307961"][{POST_SNAPBACK}][/a]
That's right. Because of its impedance, the cartridge is sensitive to the capacity of its load.
However, if it outputted several volts, but with the same output impedance, it will still be as sensitive. It depends on the output equivalent impedance, and on the load impedance, but not on the voltage output.
In choosing a wire for a tonearm, and the external run from it's base, it's well known that if high-capacitance stuff is used it can profoundly affect the tonal balance of a cartridge, producing as it is a signal measured in millivolts.
R.
[a href="index.php?act=findpost&pid=307961"][{POST_SNAPBACK}][/a]
Yes that is so much a different kettle of fish that it's not funny. The tone arm circuit is relatively high impedance and very much load dependant, the load circuit of the tone arm is actually required to tune the circuit!. The amplifier speaker interconnect is not in the least bit similar to this. Yours is a classic case of having just enough snippets of knowlegde in this area to be able to misue it.
Cable capacitance is determined by a combination of geometric factors and the dielectric constant of the insulator material used. Gerenerally both of these factors contribute to a higher capacitance in coaxial cable as compaired to ribbon or parallel pair type cable.
The formula for capacatance (per meter) of the two different types is:
Coax : C = 2 Pi * 8.85* k / ln(d2/d1)) : pF/mwhere ln(.) is natural logarithm, d1 and d2 are the inner and outer conductor diameters respectively and k is the relative dielectric constant.
Parallel Pair : C = Pi * 8.85 * k / ln(d/r) : pF/mwhere d is the pair seperation and r is the radius of each conductor. Strictly this eqaution is accurate for the case for relatively widely spaced conducters in free air where the rel-dielectric constant is unity. For a typical insulated cable the effective dielectric constant will be somewhere in between unity and that of the actual insulator used. This is becasue while some of the electric fux is confined to the insulator dialectric a good deal of it is not and so passes outside the insulator and through free air. With coax on the otherhand all of the electric flux is confined to the dielectric. For this reason the parallel pair generally has a lower effective dielectric constant and hence a lower capacitance. In addition the geometric factor ln(d/r) versus ln(d2/d1) general has a higher ratio for a parallel pair again contributing to lower capacitance.
BTW. The primary reason why coax cables generally have lower characteristic impedance (not to be confused with DC resistance) than parallel pairs is due to the higher capacitance of the coax!
So now I've explained that now let me make it perfectly clear that it is totally irrelevent for a speaker interconnect anyway!!!!!!!!!!!!!!!!!!!!!!!!!!!The typical capacitance per meter for coax is in the range of 100pF to 250pF per meter while typical values for paraller pairs is about 30 to 60 pF/meter.
So no matter what you use it's likely to contribute less than 1nF (and likely much less than this if you use a parallel pair). Now the output resistance of you amplifier is probably something like 0.1 or 0.2 ohms so the pole frequence of the 1nF would be something like 1/(2*Pi*0.1*1e-9) which is about a Giga-Hertz. So are you sure that your hearing and the rest of you system are up to handleing this frequency range Rockfan? No, just what I thought.
Of course the cable capacitance will be very small compared to the self capacticance of the speaker coil in any case, making all of the above even more futile. Pull a speaker voice coil apart some time Rockfan and take a good look at the wire from which it's wound. Meters and meters of wire, all thiner than your precious interconnects and all very closley spaced with only very thin enamel coating to seperate adjacent turns leading to a fair amount of self capacitance.
Thanks for the enlightenment.
I recall reading somewhere cable impedance is insignificant so long as the cable length does not become significant relative to the wave length of the signal, which means you need a mighty long speaker cable to notice. Is this correct?
Thanks for the enlightenment.
I recall reading somewhere cable impedance is insignificant so long as the cable length does not become significant relative to the wave length of the signal, which means you need a mighty long speaker cable to notice. Is this correct?
[a href="index.php?act=findpost&pid=308186"][{POST_SNAPBACK}][/a]
Cable impedence only starts to matter when you find yourself with substantially different voltage at the ends of the cable, and those different voltages are due to time propagation.
For audio cables, this is quite some distance, only the telcos really get into this at audio frequencies under most conditions.
Cable impedence only starts to matter when you find yourself with substantially different voltage at the ends of the cable, and those different voltages are due to time propagation.
For audio cables, this is quite some distance, only the telcos really get into this at audio frequencies under most conditions.
Is this any different for line level audio cables vs speaker cables?
I know from experience that video will have degradation if you try to string a composite or s-video cable too far. But video has a lot more frequency band required. Video also degrades less gracefully. I've occasionally put line audio through twenty or thirty feet of cable and not heard much difference, but I don't think my ears are good enough.
Is this any different for line level audio cables vs speaker cables?
[a href="index.php?act=findpost&pid=308254"][{POST_SNAPBACK}][/a]
Well, yes and no. Low-level cables are more sensitive to noise injection, but this has nothing at all to do with cable impedence, rather it has more to do with shielding, balance of twisted pairs, etc.
But in terms of simple propagation, consider what the wavelength of 20kHz is at, say, .5 C, which is a conservative speed of propagation for most any cable. Consider what 1/10 of a wavelength is, and that's roughly where you have to start worrying, for propagation reasons. It's most possible that other issues, say drive current, driving amp stability, etc, will bite you first.
It's most possible that other issues, say drive current, driving amp stability, etc, will bite you first.
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=308261")
I was about to bring that up. The stability of the amp driving the speaker cables and the output impedence of the amplifier has a large effect on sound. Highly capacitive cables (like Litz wire) can reduce the phase margin of the driving amp and cause it to oscillate or ring on some signals. I personally haven't seen this happen at audio frequencies, but I have seen it happen with 40kHz sonar transponders. Amplifier oscillation is audible - although I can see how people could percieve low amplitude oscillation to be pleasant sounding.
An interesting article on the effects of cable impedence can be found here:
[a href="http://sound.westhost.com/cable-z.htm]http://sound.westhost.com/cable-z.htm[/url]
I recall reading somewhere cable impedance is insignificant so long as the cable length does not become significant relative to the wave length of the signal, which means you need a mighty long speaker cable to notice. Is this correct?
[a href="index.php?act=findpost&pid=308186"][{POST_SNAPBACK}][/a]
No. It's the characteristic impedance, that matters only for long distances, not the impedance. The impedance matters even for short distances.
But even if you had a very long speaker cable (more than one kilometer), the characteristic impedance would be difficult to deal with, because it varies greatly across the audio frequencies.
Cable impedence only starts to matter when you find yourself with substantially different voltage at the ends of the cable,
...the difference in voltage being caused by the cable impedance.
and those different voltages are due to time propagation.
No, you can have different voltages with permanent DC, because of the resistance of the cable.
I think that the "cable impedance" Woodinville refers to, is actually cable characteristic impedance, that is the one that begins to have importance when there are "substantially different voltage at the ends of the cable, and those different voltages are due to time propagation", that is, at very high frequencies.
Of course, regular impedance can have importance at low frequencies and even at DC.
Going back to the start of this thread, I would be most interested to hear what happens if you connect a high quality pair of headphones to the Rotel and listen at very low and then at safe(your ears) high levels.
If the headphone circuit is any good, the amplifier should drive moderatly heavily at the higher levels and you can tell if the characterisitic 'sound' of the Rotel changes. If it doesn't, I would be interested to see the results of different speakers.
You want the designed output from your Rotel to reach the speakers unchanged, the only way to achieve this is with a cable large enough that all of it's characteristics can be safely ignored!
Steve.
Cable impedence only starts to matter when you find yourself with substantially different voltage at the ends of the cable,
...the difference in voltage being caused by the cable impedance.[a href="index.php?act=findpost&pid=308316"][{POST_SNAPBACK}][/a]
Well, thank you, but we are in fact talking about characteristic impedence, not lumped-paramater impedence.
On the other hand, you would be just as correct to say that the dC/dx and dL/dx of the cable are responsible for propagation speed. Was that what you meant?
Going back to the start of this thread, I would be most interested to hear what happens if you connect a high quality pair of headphones to the Rotel and listen at very low and then at safe(your ears) high levels.
If the headphone circuit is any good, the amplifier should drive moderatly heavily at the higher levels and you can tell if the characterisitic 'sound' of the Rotel changes. If it doesn't, I would be interested to see the results of different speakers.
You want the designed output from your Rotel to reach the speakers unchanged, the only way to achieve this is with a cable large enough that all of it's characteristics can be safely ignored!
Steve.
[a href="index.php?act=findpost&pid=308367"][{POST_SNAPBACK}][/a]
Well, as it happens, I've been listening to the Rotel via my Senn HD580's quite a lot since I got it, and it's actually acceptable enough that i just sold my Creek OBH-11 headphone amp, (on promising myself I'll get an OBH-21SE as soon as I can afford it).
I've had the cover off the Rotel, and the phone socket is 'standard-issue' as far as I can see - pair of series resistors on the power-amp stage..
The problem with this arrangement is that it creates a 'high-impedence' output.
This works up to a point, but ironically, despite the abundance of current-capacity, the amp effectively can't deliver current via these resistors that 'follows' the voltage it's trying to swing, and the sound get's muddier as the volume increases. It's quantified as lack of 'damping', I believe (or is it too much damping?).
The actual current flowing is also very low and a MOSET, or whatever, designed to deliver 30 or more watts isn't going to work properly delivering a fraction of a watt - it won't be 'linear' at these current levels.
I remember often being frustrated when listening via phones plugged into an integrated amp and finding that the 'dynamics' seem to disappear when listening at head-banging levels, and I'm sure many others will have noticed this.
Rega used to make a 'shunt' device with a heatsink for connecting phones to a speaker output to get round this, and I remember a gizmo called "the can-opener" that did something similar.
R.
PS - I think the reason the Rotel's headphone output sounds OK ( for a budget integrated), is that it works in class-A up to a couple of watts, as i mentioned, and this is also the reason it sounds so good at low volume 'in-room'.
Well, as it happens, I've been listening to the Rotel via my Senn HD580's quite a lot since I got it, and it's actually acceptable enough that i just sold my Creek OBH-11 headphone amp, (on promising myself I'll get an OBH-21SE as soon as I can afford it).
I've had the cover off the Rotel, and the phone socket is 'standard-issue' as far as I can see - pair of series resistors on the power-amp stage..
The problem with this arrangement is that it creates a 'high-impedence' output.
This works up to a point, but ironically, despite the abundance of current-capacity, the amp effectively can't deliver current via these resistors that 'follows' the voltage it's trying to swing, and the sound get's muddier as the volume increases. It's quantified as lack of 'damping', I believe (or is it too much damping?).
The actual current flowing is also very low and a MOSET, or whatever, designed to deliver 30 or more watts isn't going to work properly delivering a fraction of a watt - it won't be 'linear' at these current levels.
I remember often being frustrated when listening via phones plugged into an integrated amp and finding that the 'dynamics' seem to disappear when listening at head-banging levels, and I'm sure many others will have noticed this.
Rega used to make a 'shunt' device with a heatsink for connecting phones to a speaker output to get round this, and I remember a gizmo called "the can-opener" that did something similar.
R.
PS - I think the reason the Rotel's headphone output sounds OK ( for a budget integrated), is that it works in class-A up to a couple of watts, as i mentioned, and this is also the reason it sounds so good at low volume 'in-room'.
[a href="index.php?act=findpost&pid=308459"][{POST_SNAPBACK}][/a]
Rockfan, I think you've nailed it! The audible differences you hear at different levels on your Rotel are caused by the amp, not the cables. I don't really have time for this now (I need to be getting to work) but you are wrong about the output devices not being linear at only low level. Try a different pair of speakers on your amp with decent cable and report back. I for one am fascinated by this thread!
Steve
Going back to the start of this thread, I would be most interested to hear what happens if you connect a high quality pair of headphones to the Rotel and listen at very low and then at safe(your ears) high levels.
If the headphone circuit is any good, the amplifier should drive moderatly heavily at the higher levels and you can tell if the characterisitic 'sound' of the Rotel changes. If it doesn't, I would be interested to see the results of different speakers.
You want the designed output from your Rotel to reach the speakers unchanged, the only way to achieve this is with a cable large enough that all of it's characteristics can be safely ignored!
Steve.
[a href="index.php?act=findpost&pid=308367"][{POST_SNAPBACK}][/a]
Well, as it happens, I've been listening to the Rotel via my Senn HD580's quite a lot since I got it, and it's actually acceptable enough that i just sold my Creek OBH-11 headphone amp, (on promising myself I'll get an OBH-21SE as soon as I can afford it).
I've had the cover off the Rotel, and the phone socket is 'standard-issue' as far as I can see - pair of series resistors on the power-amp stage..
The problem with this arrangement is that it creates a 'high-impedence' output.
This works up to a point, but ironically, despite the abundance of current-capacity, the amp effectively can't deliver current via these resistors that 'follows' the voltage it's trying to swing, and the sound get's muddier as the volume increases. It's quantified as lack of 'damping', I believe (or is it too much damping?).
The actual current flowing is also very low and a MOSET, or whatever, designed to deliver 30 or more watts isn't going to work properly delivering a fraction of a watt - it won't be 'linear' at these current levels.
I remember often being frustrated when listening via phones plugged into an integrated amp and finding that the 'dynamics' seem to disappear when listening at head-banging levels, and I'm sure many others will have noticed this.
Rega used to make a 'shunt' device with a heatsink for connecting phones to a speaker output to get round this, and I remember a gizmo called "the can-opener" that did something similar.
R.
PS - I think the reason the Rotel's headphone output sounds OK ( for a budget integrated), is that it works in class-A up to a couple of watts, as i mentioned, and this is also the reason it sounds so good at low volume 'in-room'.
[a href="index.php?act=findpost&pid=308459"][{POST_SNAPBACK}][/a]
The whole point of that "pair of resistors" is to limit the amp's current, otheriwse either your headphones or your ears will blow! It doesn't neccessarily create a high impedance either, just high resistance. Consider that any headphone or loudspeaker is essentially current-operated anyway - it actually creates the ideal current-driver stage by using those resistors (ok, not quite true, but it's nearer to the ideal if anything).
Wrong again about the MOSFETs. The amp uses something called "bias" to keep a certain amount of current flowing, enough to keep them off the "knee" of the characteristic and hence keep them in the linear range. You statement is like saying a 160mph sports car can only be driven at that speed - wrong!
As for the original bit - interconnect is designed for just that, interconnects. In the crazy world of "hi-fi" where people pay riduculous amounts of money for things that provide no provable improvements, do you not think that someone would market interconnect marked as speaker cable for low power if it was even vaguely sensible? Sensible cable rated appropriately is more than adequate for low power as well as high. It simply isn't true that a high-power cable can't handle low powers as well - analogy as for the MOSFETs.
The whole point of that "pair of resistors" is to limit the amp's current, otheriwse either your headphones or your ears will blow! It doesn't neccessarily create a high impedance either, just high resistance. Consider that any headphone or loudspeaker is essentially current-operated anyway - it actually creates the ideal current-driver stage by using those resistors (ok, not quite true, but it's nearer to the ideal if anything).
Wrong again about the MOSFETs. The amp uses something called "bias" to keep a certain amount of current flowing, enough to keep them off the "knee" of the characteristic and hence keep them in the linear range. You statement is like saying a 160mph sports car can only be driven at that speed - wrong!
[a href="index.php?act=findpost&pid=308538"][{POST_SNAPBACK}][/a]
I have to disagree, impedence across the audio band is a figure which includes resistivity, and those resistors add substantially to it on a power-amp stage.
This isn't something I just made up, it's received wisdom, and is the whole reason for using a dedicated headpone amp, or one of those shunt gizmos on a speaker output.
The effect, as I said, is quite commonly heard in the inabilty of such headaphone outputs (ironically) to go loud convicingly, ie. deliver the current necessary (even if it results in no more than a few 100 millwatts), while keeping dynamics intact, to 'follow' the output device's voltage faithfully.
I also don't understand what you mean about maintaining 'linearity' in a high-current output device when flowing tiny fractions of that current to headphone tranducers. I just can see how, for an extreme example, a 2000 watt Krell can be 'optimised' in any sense of the word to drive tiny headphone transducers.
And 'supercars' like Ferrari's have long been known for being absolute pigs to drive around town! Until the advent of electronic fuel and ingnition management, you had to put up with chores like taking the plugs out every other day to get the soot off them! And they still don't exactly 'work' for commuting and shopping.
To me, the 'horses for course' axiom seems applicable again.
R.
PS - one thing to realize is that those headphone resistors primary function is to make the volume control usable - without them you's barely be able to get it 'off the stop', and given the lack of accuracy of most pots, levels would probably be heavily skewed to left or right. Of course there's the added benefit of preventing people instantaneously blowing their headphones and/or eardrums out, which is prbably a good idea!
This isn't something I just made up, it's received wisdom,
Well, OK then.
The effect, as I said, is quite commonly heard in the inabilty of such headaphone outputs (ironically) to go loud convicingly, ie. deliver the current necessary (even if it results in no more than a few 100 millwatts), while keeping dynamics intact, to 'follow' the output device's voltage faithfully.
Just hipshotting here, but I think the point is that (as long as neither amp nor phones are starting to compress) the resistive network may be detrimental, but should be equally so at low levels.
The whole point of that "pair of resistors" is to limit the amp's current, otheriwse either your headphones or your ears will blow! It doesn't neccessarily create a high impedance either, just high resistance. Consider that any headphone or loudspeaker is essentially current-operated anyway - it actually creates the ideal current-driver stage by using those resistors (ok, not quite true, but it's nearer to the ideal if anything).
Wrong again about the MOSFETs. The amp uses something called "bias" to keep a certain amount of current flowing, enough to keep them off the "knee" of the characteristic and hence keep them in the linear range. You statement is like saying a 160mph sports car can only be driven at that speed - wrong!
[a href="index.php?act=findpost&pid=308538"][{POST_SNAPBACK}][/a]
I have to disagree, impedence across the audio band is a figure which includes resistivity, and those resistors add substantially to it on a power-amp stage.
This isn't something I just made up, it's received wisdom, and is the whole reason for using a dedicated headpone amp, or one of those shunt gizmos on a speaker output.
The effect, as I said, is quite commonly heard in the inabilty of such headaphone outputs (ironically) to go loud convicingly, ie. deliver the current necessary (even if it results in no more than a few 100 millwatts), while keeping dynamics intact, to 'follow' the output device's voltage faithfully.
I also don't understand what you mean about maintaining 'linearity' in a high-current output device when flowing tiny fractions of that current to headphone tranducers. I just can see how, for an extreme example, a 2000 watt Krell can be 'optimised' in any sense of the word to drive tiny headphone transducers.
And 'supercars' like Ferrari's have long been known for being absolute pigs to drive around town! Until the advent of electronic fuel and ingnition management, you had to put up with chores like taking the plugs out every other day to get the soot off them! And they still don't exactly 'work' for commuting and shopping.
To me, the 'horses for course' axiom seems applicable again.
R.
PS - one thing to realize is that those headphone resistors primary function is to make the volume control usable - without them you's barely be able to get it 'off the stop', and given the lack of accuracy of most pots, levels would probably be heavily skewed to left or right. Of course there's the added benefit of preventing people instantaneously blowing their headphones and/or eardrums out, which is prbably a good idea!
[a href="index.php?act=findpost&pid=308564"][{POST_SNAPBACK}][/a]
The point of the resistors is to limit the amp's output current to a safe value - headphones never require anywhere near the current a speaker does. If the output doesn't go "convincingly loud", try reducing them in value. The primary function of the resistors is certainly NOT to make the volume control usable, though you're quite right that they have that rather beneficial side-effect. Omit them and you'll get not convincing volume, but convincing smoke from your headphones!
The analogy with a "supercar" was maybe not the best, but the fact is that it's possible to drive them slowly.
If you fail to understand the bit about bias, go do some reading. Bias means running the output devices at some minimum current to keep them in their linear region. That current appears as waste heat, essentially. A 2kw amp needs no "optimisation" to run headphones, but it sure as hell needs the resistive chain in the output to limit current!
As I said before, headphones and most speakers are current-driven devices. The ideal driver (amp) for them is one whose current is independent of the load impedance. Adding the resistors to the amp's output takes it nearer to the ideal, not further away.
The impedance curve of headphones is not flat. If they are driven through resistors, their frequency response will differ from the one that they have if they are voltage-driven.
Does this mean that the headphone jack of an amp is basicaly a source of current, rather than one of voltage? Which in turn means your headphones output more power at frequencies with high impedance, and less power at frequencies with lower impedance, right?
How big is this effect, and how are headphones supposed to be driven?
Does this mean that the headphone jack of an amp is basicaly a source of current, rather than one of voltage? Which in turn means your headphones output more power at frequencies with high impedance, and less power at frequencies with lower impedance, right?
How big is this effect, and how are headphones supposed to be driven?
[a href="index.php?act=findpost&pid=309845"][{POST_SNAPBACK}][/a]
Yes up to a point - they are a source of current, but it's limited. The end result depends on the amp's output voltage capabilities, the headphones' impedance and efficiency, the value of the resistors and so on, so it's not a simple picture. I'm simply making the point that conventional-design (i.e. moving-coil) speakers and headphones are current-driven. The current passing through the coil at any given instant determines the force on the diaphragm. In reality of course that's a simplification since the suspension, inertial mass of the diphragm and a whole heap of other factors affect the diaphragm too. Ideally, the amp would supply a current that was independent of impedance and frequency. Arranging that isn't quite so easy if you want 100% accuracy, but considering the distortion inherent in the actual speaker/headphone design I can't see it being a major factor.
It doesn't follow that your headphones output more power at frequencies where they have higher impedance. You're equating higher impedance with higher voltage for a given current and, therefore, higher power. That's only true for a resistive load, which headphones aren't. The acoustic efficiency might be lower as well (I've no idea if that's true in real designs though). Remember, the whole process is horribly inefficient and non-linear anyway. I'd be quite interested to know, if an amp was driven at low level directly into headphones, how it would compare to the output-limited setup at the same listening level. Noise would probably be worse, but I wonder how the quality would be if you ignore that.
I've often wondered why no-one designs an amplifier to operate on the current-source idea, maybe I'm missing something here?
The impedance curve of headphones is not flat. If they are driven through resistors, their frequency response will differ from the one that they have if they are voltage-driven.
[a href="index.php?act=findpost&pid=309766"][{POST_SNAPBACK}][/a]
Can't argue with that :-) Which response is "correct" though? Neither, at a guess!
I've often wondered why no-one designs an amplifier to operate on the current-source idea, maybe I'm missing something here?
[a href="index.php?act=findpost&pid=309878"][{POST_SNAPBACK}][/a]
Most amps, at least solid state ones, are designed to be as much like a voltage source as possible, and most speakers are designed to be driven by a voltage source.
Using resistors to drop the output for headphones is the problem. Since the resistors are outside the amplifier's feedback loop, they add to the output impedance, making the output to the phones less like a voltage source and more like a current source (but still not close to being either). If the phones have an impedance that varies with frequency, then the percentage of the amplifier's output voltage that reaches the phones will vary with frequency. If the impedance across the board is >> the resistors, then it won't make much difference. If it is comparable or less, then sound will suffer.
Would it be a good idea to connect two resistors, say two and six ohms, in series to the amplifier's speaker output, and then connect the headphones parallel to the two ohms resistor? It'd be terribly wasteful of power, but it would turn a big amp into more or less a voltage source at headphone levels, right? And has the benefit of loading amp more so it may behave more linear...
Something like a 20 Ohm and a 2 Ohm resistors in series with amp speaker terminals, with the headphones in parallel with the 2 Ohm resistor, will drive any headphone so that the frequency response deviation is negligible (voltage source-like), providing a 20 dB attenuation. The 20 Ohm resistor must be rated several watts.
Something like a 20 Ohm and a 2 Ohm resistors in series with amp speaker terminals, with the headphones in parallel with the 2 Ohm resistor, will drive any headphone so that the frequency response deviation is negligible (voltage source-like), providing a 20 dB attenuation. The 20 Ohm resistor must be rated several watts.
[a href="index.php?act=findpost&pid=310062"][{POST_SNAPBACK}][/a]
Only thing is it raises the damping factor considerably. Series resistors would do that too, to an even greater extent though. The 20 Ohm resisors would need to be quite a high-wattage rating, roughly 1/5 of the amp's output rating (into 8 Ohms) if you're going to wind the amp up. The 2 Ohm ones need to be about a tenth of the 20 Ohm ones, power-wise. It'll work though.
I'm curious how you calculated the 20dB attenuation - that'd be a voltage attenuation, ignoring the headphone impedance?
I have a couple of questions. These may be stupid, as the discussion went beyond my electrical knowledge about the time people started talking about voltage source vs current source.
First, if a resistor in series with the phones is a negative presence, what about volume control? That's just a variable resistor. Wouldn't it have the same effect?
Second, are the headphone jacks made this way because they have to deal with two totally different designs of headphone? (ie, cheap ones and audiophile ones, which tend to have totally different impedance and response characteristics)
It seems to me like the best thing to have, if one tended to listen to both speakers and headphones equally and only wanted a single device, would be a receiver that had two amp circuits in it. The standard high power speaker driving amp, and a little headphone level amp for when the phones are plugged in. Much more efficient too. I guess until someone makes such a device (hah! fat chance) we will have to just use headphone amps on the line in.
First, if a resistor in series with the phones is a negative presence, what about volume control? That's just a variable resistor. Wouldn't it have the same effect?[a href="index.php?act=findpost&pid=310643"][{POST_SNAPBACK}][/a]
No, because the volume control is located before the power amplifier. It changes the voltage of the input, and the output impedance remains the same.
No, because the volume control is located before the power amplifier. It changes the voltage of the input, and the output impedance remains the same.
A durrrr. I'm an idiot. I should have knew that, there aren't many variable resistors that can handle tens, much less hundreds, of watts.
Would it be a good idea to connect two resistors, say two and six ohms, in series to the amplifier's speaker output, and then connect the headphones parallel to the two ohms resistor? It'd be terribly wasteful of power, but it would turn a big amp into more or less a voltage source at headphone levels, right? And has the benefit of loading amp more so it may behave more linear...
[{POST_SNAPBACK}][/a] (http://index.php?act=findpost&pid=310041")
I assume that's what the 'can-opener' and Rega gizmos did - basically a simple voltage-divider/shunt attenuator.
Both had substantial heatsinks to dispose of (presumably) several watts of power in 'pretending' to be loudspeakers.
I think you could probably go for a total of at least 16 ohms, maybe even 30 or so, which might be useful if the amp was a powerful one.
I did think seriously about DIYing someting along these lines after the Rega and can-opener both disappeared, but bought a Creek headphone amp in the end.
R.
PS - found this;
[a href="http://cgim.audiogon.com/i/vs/i/f/1082826750.jpg]http://cgim.audiogon.com/i/vs/i/f/1082826750.jpg[/url]
It seems to me like the best thing to have, if one tended to listen to both speakers and headphones equally and only wanted a single device, would be a receiver that had two amp circuits in it. The standard high power speaker driving amp, and a little headphone level amp for when the phones are plugged in.
[a href="index.php?act=findpost&pid=310643"][{POST_SNAPBACK}][/a]
THat's one of the things you get as you move up the price scale.
Something like a 20 Ohm and a 2 Ohm resistors in series with amp speaker terminals, with the headphones in parallel with the 2 Ohm resistor, will drive any headphone so that the frequency response deviation is negligible (voltage source-like), providing a 20 dB attenuation. The 20 Ohm resistor must be rated several watts.
[a href="index.php?act=findpost&pid=310062"][{POST_SNAPBACK}][/a]
Only thing is it raises the damping factor considerably. Series resistors would do that too, to an even greater extent though. The 20 Ohm resisors would need to be quite a high-wattage rating, roughly 1/5 of the amp's output rating (into 8 Ohms) if you're going to wind the amp up.
I'm curious how you calculated the 20dB attenuation - that'd be a voltage attenuation, ignoring the headphone impedance?
[a href="index.php?act=findpost&pid=310505"][{POST_SNAPBACK}][/a]
Damping factor won't be an issue. A headphone has a minimum impedance that is several times greater than the 2 Ohm load it is in parallel with, so the amp sees a near 22 Ohm load. The headphones see 2 Ohm in parallel with a 20 Ohm, so it's nearly 2 Ohm output impedance that they see. Their impedance is quite greater, so electrical damping won't be very bad. But also, headphones are already very damped mechanically, and need very little electrical damping.
As to the 20 dB calculation, yes, that's a voltage (level), attenuation. In practice it will be a little bit more than 20 dB, depending on how low is headphone impedance.
Something like a 20 Ohm and a 2 Ohm resistors in series with amp speaker terminals, with the headphones in parallel with the 2 Ohm resistor, will drive any headphone so that the frequency response deviation is negligible (voltage source-like), providing a 20 dB attenuation. The 20 Ohm resistor must be rated several watts.
[a href="index.php?act=findpost&pid=310062"][{POST_SNAPBACK}][/a]
Only thing is it raises the damping factor considerably. Series resistors would do that too, to an even greater extent though. The 20 Ohm resisors would need to be quite a high-wattage rating, roughly 1/5 of the amp's output rating (into 8 Ohms) if you're going to wind the amp up.
I'm curious how you calculated the 20dB attenuation - that'd be a voltage attenuation, ignoring the headphone impedance?
[a href="index.php?act=findpost&pid=310505"][{POST_SNAPBACK}][/a]
Damping factor won't be an issue. A headphone has a minimum impedance that is several times greater than the 2 Ohm load it is in parallel with, so the amp sees a near 22 Ohm load. The headphones see 2 Ohm in parallel with a 20 Ohm, so it's nearly 2 Ohm output impedance that they see. Their impedance is quite greater, so electrical damping won't be very bad. But also, headphones are already very damped mechanically, and need very little electrical damping.
As to the 20 dB calculation, yes, that's a voltage (level), attenuation. In practice it will be a little bit more than 20 dB, depending on how low is headphone impedance.
[a href="index.php?act=findpost&pid=311406"][{POST_SNAPBACK}][/a]
So, for a headphone of 16 Ohms, the damping factor is 8, which isn't brilliant, though I'm not sure how much it really matters. For the same impedance I calculate the power attenuation to be about 17 dB. I'm not sure if 16 Ohms is a typical figure, I think they are often higher, but mine are certainly 16 Ohm (2 pairs). Be interesting to know how this set-up compares in reality against series resistors - someone care to try it?
In your case, damping factor wouldn't still probably be an issue, due to headphones being enough mechanically damped. But if your headphones have varying impedance across frequency, a small deviation from flat frequency response at headphone terminals (due to the 2 Ohm output impedance) could be audible. See basic Ohm's law calculation:
http://www.kikeg.arrakis.es/various/zout_zl.xls (http://www.kikeg.arrakis.es/various/zout_zl.xls)
If it is, better use a 1 Ohm / 10 Ohm combination instead.
Edit: with this same spreadsheet, compare the effect on using a simple 120 Ohm resistor in series with the headphones.