Watts Up...?
- Thread starter wink
- Start date
Jawed
1000+ Head-Fier
- Joined
- Dec 7, 2014
- Posts
- 1,121
- Likes
- 580
HD 800 S, which does not need lots of power (being rated at 300 ohm), has tighter bass when driven by DAVE than when driven by TT. TT "softens" at higher levels, adding "phatness" to the bass sound, e.g. around -20dB, and this problem disappears at lower levels. So, no matter the theories on "enough power" for such a supposedly easy load, in practice there is something about it that is not "easy".
Now playing: The Microphones - Microphone pt. 2
Now playing: The Microphones - Microphone pt. 2
JaZZ
Headphoneus Supremus
I must confess that I haven't noticed something like that with my Hugo.
jarnopp
Headphoneus Supremus
- Joined
- Nov 26, 2014
- Posts
- 2,323
- Likes
- 1,657
I can support this third paragraph and statement about power (from Mojo's perspective). The more I listen and try new things, the more appreciation for Mojo's versatility and power. I can say that driving the HiFiMan HE-6, Mojo can do it to satisfying level and sound at -8 clicks from line is compared to either a speaker amp (Odyssey Stratos glass ceiling edition) or HP amp, Cavalli Liquid Carbon. I must say, I prefer these from the LC due to the warmer signature on most of my music (alternative, rock), but I also hear the Mojo as more transparent clearly, and not lacking in bass or power to drive the HE-6.
Rob Watts
Member of the Trade: Chord Electronics
- Joined
- Apr 1, 2014
- Posts
- 3,141
- Likes
- 12,368
When current for the load is drawn from the power supplies, then you get distorted currents in the negative and positive rails; these distorted currents create an error voltage due to the PSU output impedance; this error voltage is fed back into the amp via the amp's power supply rejection (this will depend upon closed loop gain and the innate PSRR) and becomes an output error and contributes to overall distortion. Now when the amp switches from Class A mode to Class B this problem becomes more serious as higher order harmonics (up to infinite frequencies) dominate; and higher order harmonics add a hardness and aggression to the sound so must be eliminated.
There is another error mechanism in PSU's too - the DC voltage voltage modulates with the power envelope of the music; this creates low frequency error voltages through the same mechanism; this LF error adds smear and bloom to the bass (a bit more phat) which some people like. I don't - I want my bass to be accurate with good pitch perception (the LF error degrades the ability to follow the pitch of bass). There are other mechanisms too, but they are second order problems and more complicated to explain.
Now when you halve the impedance of the power supply, then you reduce both these problems by two - and this has led to the wrong idea that you need huge power supplies to give completely unneeded power and current delivery - its nothing to do with the current delivery but the minimization of PSU induced errors at the output.
Now when I design my amplifiers I have always taken this problem into account, and calculate the levels of PSU impedance required against the amp's PSRR. With Dave for example, I eliminated this problem by using very high levels of PSRR and milli-ohm levels of PSU impedance driving the OP amp. That is why when you see plots of Dave's distortion against load, with no changes whatsoever in distortion (apart from an insignificant rise in 2nd harmonic) with low loads you can see how difficult that was to achieve as I had to eliminate power supply induced distortion (together with very many other things too).
Rob
PS - external amps will not eliminate this problem as it is a designer knowledge issue. Indeed, amps that employ big PSU's are probably heading in the wrong direction, as it's clear that the designer is not cognisant of the PSU problems as there are much more elegant and effective ways of removing the problem; but these ways are not easily visible to an audiophile; moreover, large PSU's can create more RF noise - but that is another problem.
rkt31
Headphoneus Supremus
- Joined
- Oct 14, 2014
- Posts
- 2,108
- Likes
- 776
I have few questions for rob . will the upcoming chord digital amp have analog inputs ? will it have both power and pre functionality ie will it take dave's full scale digital output or volume controlled digital output ? if it allows analog inputs , what will be the limitations for input sensitivity ?
Rob Watts
Member of the Trade: Chord Electronics
- Joined
- Apr 1, 2014
- Posts
- 3,141
- Likes
- 12,368
No analog inputs; and the first unit can be used with Dave, or independently with its own volume.
rkt31
Headphoneus Supremus
- Joined
- Oct 14, 2014
- Posts
- 2,108
- Likes
- 776
thanks for the clarification rob ! may we see all in one box solution in future like m scaler , Dave and digital amp , ie a chord super dac with 1m taps and variable output as high as say 25w or even 50w ?
JaZZ
Headphoneus Supremus
Rob, I have another question. I really hope I'm not bothering you while being engaged in some of your groundbreaking developments (of course I am!
), but my issue is really earthshatteringly important, and you are the only person able to solve it.
I remember you stating that every electronics component has an impact on the sound, even resistors, cables, solder joints and plugs. My own experience tells me the same. My favorite headphone to date, the HiFiMan HE1000, sounds dramatically better with a replacement cable (with silver conductors in my case). Another dynamic headphone, the Sennheiser HD 800, reacts almost equally sensitive to cables (although its original cable isn't really bad), different cables alter the sound in different ways.
You seem to be right: The human hearing is much more sensitive than commonly believed. Nevertheless, it's amazing that such tiny signal deviations «survive» on their way through a sound transducer with its infinitely greater signal corruption. But that's how it is, and I don't expect you to have an answer to that.
However, since you're in possession of extremely sensitive measuring equipment: Have you ever measured the effect of cables on the signal? I'm explicitly not speaking of electrical values such as capacitance or inductance, but measurable signal alterations with some likelihood to have an audible impact. In what way do cable geometries and conductor materials alter the signal shape? I can somehow reproduce how abrupt impedance steps and maybe even sharp corners (e.g. from cable plugs) may produce «transmissionline resonances», but even with those it's unclear how they print themselves on the signal. With conductor materials it's hard to find any explanation in the first place. Silver is the better conductor than copper. So what? Increase the copper conductor's cross-section accordingly and you get the same resistance! Moreover, differences in the tenths of ohms absolutely don't matter with (say) headphones with load impedances of e.g. 35 ohms. And resistance per se is frequency-neutral.
However, I can't help but confirm the notion that silver conductors tend to sound treble-friendlier than copper conductors. What's the cause for this? And how would it translate into measuring results? As far as I know, silver and copper conductors measure absolutely indentical in the audio band – with sine waves. Do silver conductors take more care of the signal, or do they add treble? (Unthinkable!) Are copper conductors unable to transfer an electrical audio signal adequately, without audible losses? Is there a conductor material at all that guarantees absolute signal integrity?
I'm somewhat familiar with the skin effect and have «heard» it in my own cable experiments. However, according to the classic electroacoustics doctrine it's reserved to the ultrasonic range. Now even if it causes a tiny treble roll-off in the audio band (let's assume it!), how can it have a decisive impact on the perceived sound quality when at the same time the signal on its way has to run through any arbitrary sound transducer which may in fact benefit from it due to synergetic effects? It seems the skin effect is always perceived as negative (well, according to my experience after all). How can this be? It seems that it's more than just a treble roll-off, it may in fact have a separate quality. How does it measure, if it's real?
Regards
Marcel
I remember you stating that every electronics component has an impact on the sound, even resistors, cables, solder joints and plugs. My own experience tells me the same. My favorite headphone to date, the HiFiMan HE1000, sounds dramatically better with a replacement cable (with silver conductors in my case). Another dynamic headphone, the Sennheiser HD 800, reacts almost equally sensitive to cables (although its original cable isn't really bad), different cables alter the sound in different ways.
You seem to be right: The human hearing is much more sensitive than commonly believed. Nevertheless, it's amazing that such tiny signal deviations «survive» on their way through a sound transducer with its infinitely greater signal corruption. But that's how it is, and I don't expect you to have an answer to that.
However, since you're in possession of extremely sensitive measuring equipment: Have you ever measured the effect of cables on the signal? I'm explicitly not speaking of electrical values such as capacitance or inductance, but measurable signal alterations with some likelihood to have an audible impact. In what way do cable geometries and conductor materials alter the signal shape? I can somehow reproduce how abrupt impedance steps and maybe even sharp corners (e.g. from cable plugs) may produce «transmissionline resonances», but even with those it's unclear how they print themselves on the signal. With conductor materials it's hard to find any explanation in the first place. Silver is the better conductor than copper. So what? Increase the copper conductor's cross-section accordingly and you get the same resistance! Moreover, differences in the tenths of ohms absolutely don't matter with (say) headphones with load impedances of e.g. 35 ohms. And resistance per se is frequency-neutral.
However, I can't help but confirm the notion that silver conductors tend to sound treble-friendlier than copper conductors. What's the cause for this? And how would it translate into measuring results? As far as I know, silver and copper conductors measure absolutely indentical in the audio band – with sine waves. Do silver conductors take more care of the signal, or do they add treble? (Unthinkable!) Are copper conductors unable to transfer an electrical audio signal adequately, without audible losses? Is there a conductor material at all that guarantees absolute signal integrity?
I'm somewhat familiar with the skin effect and have «heard» it in my own cable experiments. However, according to the classic electroacoustics doctrine it's reserved to the ultrasonic range. Now even if it causes a tiny treble roll-off in the audio band (let's assume it!), how can it have a decisive impact on the perceived sound quality when at the same time the signal on its way has to run through any arbitrary sound transducer which may in fact benefit from it due to synergetic effects? It seems the skin effect is always perceived as negative (well, according to my experience after all). How can this be? It seems that it's more than just a treble roll-off, it may in fact have a separate quality. How does it measure, if it's real?
Regards
Marcel
Rob Watts
Member of the Trade: Chord Electronics
- Joined
- Apr 1, 2014
- Posts
- 3,141
- Likes
- 12,368
Good question. A long time ago I spent a lot of time designing cables, and found out that small signal non-linearity was crucial for depth and detail resolution.
So clean oxide free contacts have better depth; higher purity copper is better too; silver bearing solders had better detail resolution - etc., etc. The common strand is impurities and oxides; and oxides create small signal non-linearities - indeed you can make diodes from copper oxide for example.
Also I experimented with solid core cables, and found they sounded better than stranded - also I found that silver plated stranded wire sounded more like solid core. The problem actually turned out to be skin effect, but the oxide layer was isolating each individual strands for small signals, so the skin effect would vary depending upon signal level due to the oxides isolating for small signals but not isolating at large signals. This is why skin effect is important - its nothing to do with the impedance variation itself, but the non-linearity caused by surface oxides on the strands. You also have a similar problem with crystal boundaries too. The long and short of it is small individually isolated solid core strands sound so much better than single thick solid core or stranded wire and it is all explainable in terms of tiny small signal non-linearity.
Now I could easily hear very large changes with this; so since it was easy to hear, we should be able to measure it. So I used 100M of poor grade stranded copper wire and measured it - and could measure no change at all. Later on in my career I was measuring one of my designs - an extremely low distortion digital amp. And I noticed distortion was higher than usual, and it turned out to be the load switch - it had corroded, and the oxides were creating distortion. I also measured improvements when crimped connections were replaced with soldered ones on the load. So loudspeaker cables these effects are measurable - albeit with the worst case. But for interconnects I have failed to measure any kind of distortion, but nonetheless it is very audible.
Now with the Dave design, I started to notice some very odd things - noise shapers were changing the sound quality. Now noise shapers have a low level resolution problem just like cables; small signals are attenuated compared to big signals. Fortunately, we can measure this with simulation. And I could correlate noise shaper performance with depth perception - 220 dB noise shapers have better depth than 200 dB ones - and this effect continues onwards, so one can hear 350 dB noise shapers having better depth than 330 dB. I have repeated these tests many many times, and I always get the same answer - better noise shaper performance equals better depth perception.
What this tells us is that the small signal error, no matter how small, is audible. I do not know why this is, and its a puzzle as to why the ear/brain is so sensitive. But it does provide a magnificent explanation why depth and detail resolution for interconnects is not measurable but audible - we will never be able to measure -330 dB non-linearity.
Rob
So clean oxide free contacts have better depth; higher purity copper is better too; silver bearing solders had better detail resolution - etc., etc. The common strand is impurities and oxides; and oxides create small signal non-linearities - indeed you can make diodes from copper oxide for example.
Also I experimented with solid core cables, and found they sounded better than stranded - also I found that silver plated stranded wire sounded more like solid core. The problem actually turned out to be skin effect, but the oxide layer was isolating each individual strands for small signals, so the skin effect would vary depending upon signal level due to the oxides isolating for small signals but not isolating at large signals. This is why skin effect is important - its nothing to do with the impedance variation itself, but the non-linearity caused by surface oxides on the strands. You also have a similar problem with crystal boundaries too. The long and short of it is small individually isolated solid core strands sound so much better than single thick solid core or stranded wire and it is all explainable in terms of tiny small signal non-linearity.
Now I could easily hear very large changes with this; so since it was easy to hear, we should be able to measure it. So I used 100M of poor grade stranded copper wire and measured it - and could measure no change at all. Later on in my career I was measuring one of my designs - an extremely low distortion digital amp. And I noticed distortion was higher than usual, and it turned out to be the load switch - it had corroded, and the oxides were creating distortion. I also measured improvements when crimped connections were replaced with soldered ones on the load. So loudspeaker cables these effects are measurable - albeit with the worst case. But for interconnects I have failed to measure any kind of distortion, but nonetheless it is very audible.
Now with the Dave design, I started to notice some very odd things - noise shapers were changing the sound quality. Now noise shapers have a low level resolution problem just like cables; small signals are attenuated compared to big signals. Fortunately, we can measure this with simulation. And I could correlate noise shaper performance with depth perception - 220 dB noise shapers have better depth than 200 dB ones - and this effect continues onwards, so one can hear 350 dB noise shapers having better depth than 330 dB. I have repeated these tests many many times, and I always get the same answer - better noise shaper performance equals better depth perception.
What this tells us is that the small signal error, no matter how small, is audible. I do not know why this is, and its a puzzle as to why the ear/brain is so sensitive. But it does provide a magnificent explanation why depth and detail resolution for interconnects is not measurable but audible - we will never be able to measure -330 dB non-linearity.
Rob
JaZZ
Headphoneus Supremus
Fascinating!
Unmeasurable audio penomena! You won't get agreement on the human hearing being more sensitive than the best measuring instruments in the Sound Science forum, but I'm ready to believe it. There's no other way out.
Interestingly I've come to the very same conclusion: Solid-core wires sound better (cleaner, smoother) than stranded wires, but large diameters of them have a tendency to overaccentuated smoothness and a metallic flavor, so the thinner the wire, the better and the more accurate the sound. I ended up with 0.04 mm magnet wires (enamel-coated wires) and used up to 400 per conductor for intercomects and headphone cables.
Unfortunately I never had access to high-quality wire (high-purity, OCC...), so had to accept that some more conventionally designed buyable cables still sounded better in some respects. Also, I seem to prefer silver to copper, it sounds more natural to my ears.
During my hyper-active (hobbyist) speaker-builder years I found fourth-order crossover filters to be preferrable, but they required extremely careful component selection. So I scrupulously measured the AC resistance of the corresponding drivers at the crossover frequencies, assembled capacitors until they had the desired values, the same with inductor coils which I tuned exactly to the desired inductance, the same with resistors for adapting the tweeters to the woofers – all with a precision higher than 0.005%. I guess the acoustic result of such a precision is just as little measurable (in dB) as cable sound. For extreme fine-tuning I finally used rotary capacitors with a few pF, switched in parallel to the main capacitor block (see knob on the black box below the guitar).
The funny thing was that they revealed rhythms within their tuning range: several good sounding areas with nevertheless differing characteristics, and within the good sounding areas there was an even finer rhythm. The less funny aspect of the whole was that the good tuning found with great effort was gone the other day. Apparently temperature and moisture differences had a greater, overriding effect, be it from the contacts or the speaker-driver properties. This was my last speaker project (with an aluminum woofer and a homegrown fiberglass tweeter with extreme rear ventilation for reduced reflections and compression effects). For the mentioned reason I never managed to finish the crossover tuning, let alone building definitive housings. BTW, shifting the tweeter back or forth by one millimeter or even a fraction of it relative to the woofer led to very similar effects.
Interestingly I've come to the very same conclusion: Solid-core wires sound better (cleaner, smoother) than stranded wires, but large diameters of them have a tendency to overaccentuated smoothness and a metallic flavor, so the thinner the wire, the better and the more accurate the sound. I ended up with 0.04 mm magnet wires (enamel-coated wires) and used up to 400 per conductor for intercomects and headphone cables.
Unfortunately I never had access to high-quality wire (high-purity, OCC...), so had to accept that some more conventionally designed buyable cables still sounded better in some respects. Also, I seem to prefer silver to copper, it sounds more natural to my ears.
During my hyper-active (hobbyist) speaker-builder years I found fourth-order crossover filters to be preferrable, but they required extremely careful component selection. So I scrupulously measured the AC resistance of the corresponding drivers at the crossover frequencies, assembled capacitors until they had the desired values, the same with inductor coils which I tuned exactly to the desired inductance, the same with resistors for adapting the tweeters to the woofers – all with a precision higher than 0.005%. I guess the acoustic result of such a precision is just as little measurable (in dB) as cable sound. For extreme fine-tuning I finally used rotary capacitors with a few pF, switched in parallel to the main capacitor block (see knob on the black box below the guitar).
The funny thing was that they revealed rhythms within their tuning range: several good sounding areas with nevertheless differing characteristics, and within the good sounding areas there was an even finer rhythm. The less funny aspect of the whole was that the good tuning found with great effort was gone the other day. Apparently temperature and moisture differences had a greater, overriding effect, be it from the contacts or the speaker-driver properties. This was my last speaker project (with an aluminum woofer and a homegrown fiberglass tweeter with extreme rear ventilation for reduced reflections and compression effects). For the mentioned reason I never managed to finish the crossover tuning, let alone building definitive housings. BTW, shifting the tweeter back or forth by one millimeter or even a fraction of it relative to the woofer led to very similar effects.
DaveRedRef-III
500+ Head-Fier
Fascinating stuff Jazz. A very rewarding project I am sure.
Regarding Silver wires and connections, I find that silver definitely contributes to a smoother audio sound but I find it can be over done and so I prefer some (high quality) copper cable somewhere in the chain. Otherwise I find too much silver seems to create a sheen to the music that isn't natural to my ears. It is something I have experienced regardless of manufacturer. It's like some of the realism is lost with silver overkill. At least that's what I hear. Copper by contrast can be aggressive on 'S's so there is a balance to be had. As I say my preference though is to have some high quality copper in the chain because its sounds more real/organic. Maybe my preference of music has something to do with this observation as I am very picky on how guitar strings sound (both electric and acoustic). Acoustic guitar strings for instance are wound with bronze which in turn is predominantly made from Copper.
Regarding Silver wires and connections, I find that silver definitely contributes to a smoother audio sound but I find it can be over done and so I prefer some (high quality) copper cable somewhere in the chain. Otherwise I find too much silver seems to create a sheen to the music that isn't natural to my ears. It is something I have experienced regardless of manufacturer. It's like some of the realism is lost with silver overkill. At least that's what I hear. Copper by contrast can be aggressive on 'S's so there is a balance to be had. As I say my preference though is to have some high quality copper in the chain because its sounds more real/organic. Maybe my preference of music has something to do with this observation as I am very picky on how guitar strings sound (both electric and acoustic). Acoustic guitar strings for instance are wound with bronze which in turn is predominantly made from Copper.
Here's why never to get silver-plated copper cables (from Ray Kimber of Kimber Cable):
https://www.youtube.com/watch?v=qJ-XeE-2qhI&feature=youtu.be&t=40m14s
https://www.youtube.com/watch?v=qJ-XeE-2qhI&feature=youtu.be&t=40m14s
Rob Watts
Member of the Trade: Chord Electronics
- Joined
- Apr 1, 2014
- Posts
- 3,141
- Likes
- 12,368
Not been my experience - I have silver plated solid core PTFE cables, that are decades old, and when you strip down the insulation it is oxide free perfect mirror finish silver.
These wires are rated at 200 deg C - so its impervious to oxidation.
The only issue you need to be wary off is soldering it - when its soldered the silver is absorbed by the solder and you get an inter-metallic layer that does not sound good - using silver solder eliminates this issue. Also, you can't use crimp connections too; but then you can't use crimp with copper either.
These wires are rated at 200 deg C - so its impervious to oxidation.
The only issue you need to be wary off is soldering it - when its soldered the silver is absorbed by the solder and you get an inter-metallic layer that does not sound good - using silver solder eliminates this issue. Also, you can't use crimp connections too; but then you can't use crimp with copper either.
Triode User
Member of the Trade: WAVE High Fidelity
I thought he was saying there can be very thin oxide layer between the copper and silver after a period of time, not on the outer surface of the silver.
Users who are viewing this thread
