Fas42’s Stereo ‘Magic’

[fas42]

Noting another member on a similar journey to me, @dmance. His angle is stopping RF effects, which is a major part of the picture - has a business, Audiowise, which sells products which aid controlling this factor. His blog on the site uses the same language I do, in describing what one needs to do, and what is possible - he has a solution using a very efficient speaker driver driven directly by the high output DAC circuitry - straight from @ray-dude's songbook! 😁 - which delivers, "magic" sound. ... Yet Again, simplicity is a powerful tool for accessing the meat in this game, 😉.

 

The difficulties typically arise as soon as one tries to extend the possibilities - getting more volume, more powerful bass, etc ... swings and roundabouts. Extra complexity, more power draw, higher currents flowing, breeds problems - and the SQ suffers.

 

How to solve? Attention to detail, experimenting, experimenting, experimenting - as Dan points out, noise is insidious, and gets in any way it can. Whether you wish to call it tweaking, optimising, insightful engineering, etc, the point is that a certain perspective has to be taken on what one is trying to achieve; if there is poor understanding of the end goal, what you're looking for, then the chances are high that the results will not measure up, no matter how much is thrown at the situation using conventional thinking ...

 

 

[ray-dude]

Dan (@dmance) is indeed a fellow traveler!  My last way excessive write up was for his Opto*DX product (with some wide detours into RF/power hygiene).  See:

 

https://www.head-fi.org/showcase/audiowise-opto•dx-optical-isolation-bridge-for-dual-spdif.23757/reviews#review-22155https://www.head-fi.org/showcase/audiowise-opto•dx-optical-isolation-bridge-for-dual-spdif.23757/reviews%23item-review-22155

 

It was quite remarkable what one can hear as one starts to strip away all the things that were keeping you from hearing it.  For me, DAC's like the Chord DAVE are true reference pieces.  Everything else in a chain takes something away from it.  The trick is to eliminate those things or minimize the impact of those things, to get as close to true references as you can.

 

 

[fas42]

22 hours ago, ray-dude said:

 

It was quite remarkable what one can hear as one starts to strip away all the things that were keeping you from hearing it.  For me, DAC's like the Chord DAVE are true reference pieces.  Everything else in a chain takes something away from it.  The trick is to eliminate those things or minimize the impact of those things, to get as close to true references as you can.

 

 

 

Yes, I use the term Subtracting Badness, 😁.

 

From that piece you linked to,

 

Quote

The maddening thing is trying to come up with theories for how things way away from the DAC could be influencing the DAC.

 

Trying to understand this for decades, as I have, doesn't give one full answers, I'm afraid 😉 ... first up, digital playback is incredibly sensitive to all sorts of noise issues, which is exactly what you've covered; and secondly, human hearing can zero in on the anomalies with piercing deftness - it KNOWS something's wrong with the sound ... what one then gets is "fatiguing listening", which @Rexp is currently asking questions about.

 

It would be lovely if a solution was available over the counter, at a budget cost - but I haven't come across such a beast, as yet. Solid DIY troubleshooting, as I do, will knock it over - the more one investigates, and applies measures, the closer one gets to subjectively artifact free SQ. The other benefit is that "unlistenable to" recordings diminish to zero - which for me is a major part of the end game.

[gmgraves]

On 7/26/2020 at 4:32 PM, fas42 said:

Noting another member on a similar journey to me, @dmance. His angle is stopping RF effects, which is a major part of the picture - has a business, Audiowise, which sells products which aid controlling this factor. His blog on the site uses the same language I do, in describing what one needs to do, and what is possible - he has a solution using a very efficient speaker driver driven directly by the high output DAC circuitry - straight from @ray-dude's songbook! 😁 - which delivers, "magic" sound. ... Yet Again, simplicity is a powerful tool for accessing the meat in this game, 😉.

 

The difficulties typically arise as soon as one tries to extend the possibilities - getting more volume, more powerful bass, etc ... swings and roundabouts. Extra complexity, more power draw, higher currents flowing, breeds problems - and the SQ suffers.

 

How to solve? Attention to detail, experimenting, experimenting, experimenting - as Dan points out, noise is insidious, and gets in any way it can. Whether you wish to call it tweaking, optimising, insightful engineering, etc, the point is that a certain perspective has to be taken on what one is trying to achieve; if there is poor understanding of the end goal, what you're looking for, then the chances are high that the results will not measure up, no matter how much is thrown at the situation using conventional thinking ...

 

 

Then how come, on most systems, it doesn’t show-up on an oscilloscope, even when that scope is set to it’s highest sensitivity? On most wide-band ‘scopes, the highest sensitivity is equal to a signal that is more than 130 dB down from the DUT’s maximum signal output, and for most people this “RF interference” still doesn’t show itself. Sure switching transients from air conditioners and refrigerators will show up, but they’re instantaneous, not continuous. Of course, there are exceptions. I know a guy who lives about a mile from an FM broadcast tower. He does have problems with RF interference, but he’s a rare exception. Perhaps, if one’s next door neighbor is a ham radio enthusiast, or one lives close to a commercial radio transmitter, this would be a real-world problem, but for the vast majority of audiophiles, the problem simply won’t exist. Rule-of-thumb, Frank: if the signal can’t be detected at the output of any audio components, then, for all intents and purposes it doesn’t exist!

[fas42]

24 minutes ago, gmgraves said:

Then how come, on most systems, it doesn’t show-up on an oscilloscope, even when that scope is set to it’s highest sensitivity? On most wide-band ‘scopes, the highest sensitivity is equal to a signal that is more than 130 dB down from the DUT’s maximum signal output, and for most people this “RF interference” still doesn’t show itself. Sure switching transients from air conditioners and refrigerators will show up, but they’re instantaneous, not continuous. Of course, there are exceptions. I know a guy who lives about a mile from an FM broadcast tower. He does have problems with RF interference, but he’s a rare exception. Perhaps, if one’s next door neighbor is a ham radio enthusiast, or one lives close to a commercial radio transmitter, this would be a real-world problem, but for the vast majority of audiophiles, the problem simply won’t exist. Rule-of-thumb, Frank: if the signal can’t be detected at the output of any audio components, then, for all intents and purposes it doesn’t exist!

 

Because it's difficult to determine the precise waveforms of the interference signals, and their path within the system - if it was easy to see on a scope, this would have been done decades ago; looking at an output, using conventional ways of hooking up measuring devices, may always miss capturing the anomaly - one may need a very capable instrument, with extensive acquisition memory, and then good software to analyse what's going on in the captured data, to properly understand.

[sandyk]

22 minutes ago, fas42 said:

Because it's difficult to determine the precise waveforms of the interference signals, and their path within the system

 

In the case of PC Audio, most of the interfering signals are generated within the PC itself.

 Even Motherboard controlled PWM fans generate RF/EMI due to the harmonics of their typical 25kHz "square wave" type pulses.

 SSDs also generate considerable RF/EMI and should be shielded .

14A2-B4.pdf

[opus101]

1 hour ago, gmgraves said:

Then how come, on most systems, it doesn’t show-up on an oscilloscope, even when that scope is set to it’s highest sensitivity? On most wide-band ‘scopes, the highest sensitivity is equal to a signal that is more than 130 dB down from the DUT’s maximum signal output, and for most people this “RF interference” still doesn’t show itself. 

 

A bit of back-of-envelope work. A typical mainstream 'scope will have 1mV/div max sensitivity and have an 8bit ADC. This gives 39uV as the lowest level signal it will resolve. 130dB relative to this is 123V - which DUT gives that much output?

[fas42]

Even the simplest setups are subject to the interference 'rules' - with the current active speakers, I have to take absolute care with precisely how every cable is dressed, especially the ones plugged into the mains. And exactly how other electrical devices in the home are plugged in, and switched on. I've found out a "pattern" of doing all that which if everything is exactly as it should be, then the SQ gets into a very good zone. But if I get just one thing wrong - then, well, I get a dose of "fatiguing listening" ... 🤨

 

Now, this is crazy!! So, I need filtering and isolating to work to a high enough standard so that none of this nonsense is necessary - which is what I'm working towards ... what the pattern of electrical noise is, and precisely what it's impacting to cause the loss of SQ, I have very little idea of. Step one is determine a workable solution; at some point, when all is bedded down, then it would be interesting to very carefully monitor all interesting areas in the rig, to see if I can work out exactly what the causal chain is.

[gmgraves]

1 hour ago, fas42 said:

 

Because it's difficult to determine the precise waveforms of the interference signals, and their path within the system - if it was easy to see on a scope, this would have been done decades ago; looking at an output, using conventional ways of hooking up measuring devices, may always miss capturing the anomaly - one may need a very capable instrument, with extensive acquisition memory, and then good software to analyse what's going on in the captured data, to properly understand.

If it isn’t “easy to see on a (100 MHz) scope”, then (“the anomaly”) doesn’t exist, and anything other than that is simply fantasy and wishful thinking.

[gmgraves]

9 minutes ago, opus101 said:

 

A bit of back-of-envelope work. A typical mainstream 'scope will have 1mV/div max sensitivity and have an 8bit ADC. This gives 39uV as the lowest level signal it will resolve. 130dB relative to this is 123V - which DUT gives that much output?

My point exactly!

[opus101]

1 minute ago, gmgraves said:

My point exactly!

 

Tell us the answer to my question - which DUT gives 123V output?

[sandyk]

4 minutes ago, opus101 said:

 

Tell us the answer to my question - which DUT gives 123V output?

 

 A USA mains supply !  😉

[opus101]

2 minutes ago, sandyk said:

 

 A USA mains supply !  😉

 

Ah so the DUT must be a mains supply conditioner, right?

 

But that doesn't fit because @gmgraves clearly called its output a 'signal' and mains supply conditioners only output power, not signal.

[fas42]

13 minutes ago, gmgraves said:

If it isn’t “easy to see on a (100 MHz) scope”, then (“the anomaly”) doesn’t exist, and anything other than that is simply fantasy and wishful thinking.

 

Note the group of members, some who just responded to recent posts here, who echo my sentiments that in fact this is highly audible - and further that it is highly difficult to get under control ... "wishful thinking" may be useful to push the problem away, and pretend it isn't there, but those who are interested in hearing the recording, and only the recording, may benefit from recognising that all this matters ...

[gmgraves]

8 minutes ago, fas42 said:

Even the simplest setups are subject to the interference 'rules' - with the current active speakers, I have to take absolute care with precisely how every cable is dressed, especially the ones plugged into the mains. And exactly how other electrical devices in the home are plugged in, and switched on. I've found out a "pattern" of doing all that which if everything is exactly as it should be, then the SQ gets into a very good zone. But if I get just one thing wrong - then, well, I get a dose of "fatiguing listening" ... 🤨

 

Now, this is crazy!! So, I need filtering and isolating to work to a high enough standard so that none of this nonsense is necessary - which is what I'm working towards ... what the pattern of electrical noise is, and precisely what it's impacting to cause the loss of SQ, I have very little idea of. Step one is determine a workable solution; at some point, when all is bedded down, then it would be interesting to very carefully monitor all interesting areas in the rig, to see if I can work out exactly what the causal chain is.

Your active speakers have what, a couple of inputs, a mains cable, and a speaker cable running from the “active” speaker to the “passive” one? Since active speakers are a closed system with all of the electronics inside one of the speaker cabinets, what can you do? Dress the input cables away from the single mains cable? You could always connect the mains via a commercially available “power conditioner”, but to be honest, without a well-made isolation transformer, those don’t do much. Large transformers with lots of iron, have little to no response above mains frequency (50-60 Hz), making it difficult for high-frequency noise to be coupled across from the primary to the secondary. An LRC filter before and after both windings will complete the isolation. But such isolation transformers are pretty expensive (lots of copper, lots of Iron), hard to find, and the cheap ones are not very effective. I have a 30 Amp hospital-grade one, designed to isolate very sensitive medical test equipment, where absolutely clean mains is often the difference between life and death. Perhaps that’s one reason why, in spite of my use of pretty generic cables, I don’t find my system wanting and in need of constant tweaking.

[gmgraves]

14 minutes ago, opus101 said:

 

Ah so the DUT must be a mains supply conditioner, right?

 

But that doesn't fit because @gmgraves clearly called its output a 'signal' and mains supply conditioners only output power, not signal.

You are correct. I am referring to the output signal found on the RCA or XLR output jacks of any audio component (except for power amps, then it’s the speaker terminals on the amp.).

[opus101]

Just now, gmgraves said:

You are correct. I am referring to the output signal found on the RCA or XLR output jacks of any audio component (except for power amps, then it’s the speaker terminals on the amp.).

 

So which DUT did you use to base your 130dB claim on? This enquiring mind wishes to know.

[fas42]

8 minutes ago, gmgraves said:

Your active speakers have what, a couple of inputs, a mains cable, and a speaker cable running from the “active” speaker to the “passive” one? Since active speakers are a closed system with all of the electronics inside one of the speaker cabinets, what can you do? Dress the input cables away from the single mains cable? You could always connect the mains via a commercially available “power conditioner”, but to be honest, without a well-made isolation transformer, those don’t do much. Large transformers with lots of iron, have little to no response above mains frequency (50-60 Hz), making it difficult for high-frequency noise to be coupled across from the primary to the secondary. An LRC filter before and after both windings will complete the isolation. But such isolation transformers are pretty expensive (lots of copper, lots of Iron), hard to find, and the cheap ones are not very effective. I have a 30 Amp hospital-grade one, designed to isolate very sensitive medical test equipment, where absolutely clean mains is often the difference between life and death. Perhaps that’s one reason why, in spite of my use of pretty generic cables, I don’t find my system wanting and in need of constant tweaking.

 

As I explained on the Edifier thread, it's a bit darned silly to add an isolation transformer, or power conditioner, that costs more than the rig it's being used on! A sledgehammer solution may do the job - but I'm looking for an engineered one, 😉.

 

In a process of refining my ideas using simulation software, right now, on how to get effective filtering, using low cost electronic parts - something that does the job; but no more than necessary, because having such will only cost. Might take a few iterations to hit a sweet spot - stay posted! 🙂

[Rexp]

5 hours ago, gmgraves said:

Then how come, on most systems, it doesn’t show-up on an oscilloscope, even when that scope is set to it’s highest sensitivity? On most wide-band ‘scopes, the highest sensitivity is equal to a signal that is more than 130 dB down from the DUT’s maximum signal output, and for most people this “RF interference” still doesn’t show itself. Sure switching transients from air conditioners and refrigerators will show up, but they’re instantaneous, not continuous. Of course, there are exceptions. I know a guy who lives about a mile from an FM broadcast tower. He does have problems with RF interference, but he’s a rare exception. Perhaps, if one’s next door neighbor is a ham radio enthusiast, or one lives close to a commercial radio transmitter, this would be a real-world problem, but for the vast majority of audiophiles, the problem simply won’t exist. Rule-of-thumb, Frank: if the signal can’t be detected at the output of any audio components, then, for all intents and purposes it doesn’t exist!

Its a known unknown. 

[Rexp]

3 hours ago, opus101 said:

 

A bit of back-of-envelope work. A typical mainstream 'scope will have 1mV/div max sensitivity and have an 8bit ADC. This gives 39uV as the lowest level signal it will resolve. 130dB relative to this is 123V - which DUT gives that much output?

Doesn't sound like a maistream scope would pick up much? 

[opus101]

5 minutes ago, Rexp said:

Doesn't sound like a maistream scope would pick up much? 

 

That's one way to look at it. Another is it'll pick up everything within its bandwidth - then how to tell which part is the signal you're looking for? There's a very good reason why much RF work is done with a spectrum analyser not a 'scope.

[gmgraves]

4 hours ago, fas42 said:

 

Note the group of members, some who just responded to recent posts here, who echo my sentiments that in fact this is highly audible - and further that it is highly difficult to get under control ... "wishful thinking" may be useful to push the problem away, and pretend it isn't there, but those who are interested in hearing the recording, and only the recording, may benefit from recognising that all this matters ...

You’re dreaming. If there is such interference, why would it not be visible on a ‘scope? Are you saying that such instruments are discriminatory, showing some signals and ignoring others? Do you really believe that?

[opus101]

1 minute ago, gmgraves said:

You’re dreaming. 

 

So are you with your 130dB claim. Apparently.

[gmgraves]

3 hours ago, opus101 said:

 

So which DUT did you use to base your 130dB claim on? This enquiring mind wishes to 

I mistyped. I meant -103 dB.

[gmgraves]

44 minutes ago, opus101 said:

 

That's one way to look at it. Another is it'll pick up everything within its bandwidth - then how to tell which part is the signal you're looking for? There's a very good reason why much RF work is done with a spectrum analyser not a 'scope.

While you are right, RF interference with enough amplitude to be a problem would be easy to see on a ‘scope (with no audio signal). Most of the time there is none.