Hi,
there are zillions of threads regarding the performance of voltage regulators. There are suberb papers dealing with the superregulator of walt jung and jan didden, for instance.
What are *your* experiences which parameters make regulator sound good? Let's focus on pre- and lineamps.
I'm interested as well in probably different sonics between regulators with feedback and open loop regs (e.g. emitter followers)
Some parameters that come to my mind are:
- low noise
- low output impedance
- flat output impedance
- stability with rapid load changes (that one might be easier with open loop circuits)
- clear definition of ground (low impedance ground point)
- ...
What is *not* that ultimatly important?
- superproper voltage accuracy
- ...?
I've tried and heard not that few different approaches until now, and my biggest surprise was, how bad batteries (lead gel) sound. Lush, flat and lifeless (yes, they have very smooth and nice trebles, but that's it.)
Nearly all my experiernces where made with various MC-phono stages. And no, no scientific setting or hearing tests were made, just build&hear.
Well, as always, we need to take our choice. I've never heard a more transparent 'stage' and definition of where the instruments and singers are located as with the superreg.
An open loop shunt reg (with a discrete circuit) can make the cymbals sound *right* like nothing else I tried (I'm a bit weighted, because I'm a drummer).
What are you findings and decisions and why?
regards,
Rüdiger
there are zillions of threads regarding the performance of voltage regulators. There are suberb papers dealing with the superregulator of walt jung and jan didden, for instance.
What are *your* experiences which parameters make regulator sound good? Let's focus on pre- and lineamps.
I'm interested as well in probably different sonics between regulators with feedback and open loop regs (e.g. emitter followers)
Some parameters that come to my mind are:
- low noise
- low output impedance
- flat output impedance
- stability with rapid load changes (that one might be easier with open loop circuits)
- clear definition of ground (low impedance ground point)
- ...
What is *not* that ultimatly important?
- superproper voltage accuracy
- ...?
I've tried and heard not that few different approaches until now, and my biggest surprise was, how bad batteries (lead gel) sound. Lush, flat and lifeless (yes, they have very smooth and nice trebles, but that's it.)
Nearly all my experiernces where made with various MC-phono stages. And no, no scientific setting or hearing tests were made, just build&hear.
Well, as always, we need to take our choice. I've never heard a more transparent 'stage' and definition of where the instruments and singers are located as with the superreg.
An open loop shunt reg (with a discrete circuit) can make the cymbals sound *right* like nothing else I tried (I'm a bit weighted, because I'm a drummer).
What are you findings and decisions and why?
regards,
Rüdiger
A constant voltage source can't alter an audio signal. Regulation has to be made extremely loose in order to measure significant signal changes.
On the other hand, the placebo effect has been very well documented and consistently proved over the years, unlike every word written about the subjective "sonics" of electronics.
Considering placebo effect is a must when evaluating things such as new drugs and any other thing that has to be felt by individuals and the result depends entirely on their opinions.
BTW: Only a placebo-controlled test will tell you what you hear and what you don't hear, and with no room for arguments because it just reveals your mistakes. Consider medications for example, would you pretend to know all the effects of all the madications that you have ever taken? Same happens with sound.
On the other hand, the placebo effect has been very well documented and consistently proved over the years, unlike every word written about the subjective "sonics" of electronics.
Considering placebo effect is a must when evaluating things such as new drugs and any other thing that has to be felt by individuals and the result depends entirely on their opinions.
BTW: Only a placebo-controlled test will tell you what you hear and what you don't hear, and with no room for arguments because it just reveals your mistakes. Consider medications for example, would you pretend to know all the effects of all the madications that you have ever taken? Same happens with sound.
Eva's point is not too far off. There are people who are passionate about batteries and think conventional supplies suck. There are people who tout series regulators, others are convinced that series regulators are no good, one must use shunt. Lots of people who think it's all hopeless and won't use active regulation at all.
The absence of any consistent pattern strongly suggests that the differences are more psychological and sociological than electrical. IME, if one gets the criteria outlined in the first post correct, then the power supply will do its job.
The absence of any consistent pattern strongly suggests that the differences are more psychological and sociological than electrical. IME, if one gets the criteria outlined in the first post correct, then the power supply will do its job.
Onvinyl said:
I don't take EVA's post as "telling you what you hear".
My experience is that until you do "scientific setting or hearing tests were made", you can't rule out expectation bias (aka placebo effect)... hence your findings based on "just build&hear" only apply to yourself and anyone else you happen to convince.
Seeing as a battery with sufficient current capacity would seem to represent the ultimate DC voltage source (I'm an electrochemist), your "listening observations" seem to be oxymoronic.
John L.
Hi,
I'm aware what a placebo effect is, how it works and so on. I'm not claiming I'm immune (that would only show I didn't get how it works).
Im just as sure as one can be, that my observations are true in my specific audio chain, because it's consistent since many different tries and circuits.
I would not claim that same results in a sience magazine, beause I just didn't make those double blind test. They have their methodogical issues, but that is not my point here.
Not to seldom, I have listening results that don't fit with my presumptions. Acutally, I was sure that batteries shall sound superior. Wrong, a simple 317 based reg sounded better. The batteries were new and not shy (7.2Ah).
The same with my former tube amp: I was a tube amp believer, but then a relativly simple AB-transitor amp blew it awasy, sonically.
The difference is comparabel in some cases with up- or downgrading your cartridge: the sound is different, and often enough the subjektive rating does not reflect the price tag.
If different circuits would sound the same to me, I wouldn't bother trying diy.
I agree absolutly, that a blind test would be more telling. Sometimes I forgot which mod I exactly performed, when I switch on my system after some days. That can be very revealing.
Rüdiger
I'm aware what a placebo effect is, how it works and so on. I'm not claiming I'm immune (that would only show I didn't get how it works).
Im just as sure as one can be, that my observations are true in my specific audio chain, because it's consistent since many different tries and circuits.
I would not claim that same results in a sience magazine, beause I just didn't make those double blind test. They have their methodogical issues, but that is not my point here.
Not to seldom, I have listening results that don't fit with my presumptions. Acutally, I was sure that batteries shall sound superior. Wrong, a simple 317 based reg sounded better. The batteries were new and not shy (7.2Ah).
The same with my former tube amp: I was a tube amp believer, but then a relativly simple AB-transitor amp blew it awasy, sonically.
The difference is comparabel in some cases with up- or downgrading your cartridge: the sound is different, and often enough the subjektive rating does not reflect the price tag.
If different circuits would sound the same to me, I wouldn't bother trying diy.
I agree absolutly, that a blind test would be more telling. Sometimes I forgot which mod I exactly performed, when I switch on my system after some days. That can be very revealing.
Rüdiger
Thinking a battery to be a good source for audio power is a big mistake, in my experience.
We planned to offer a battery pack as a third power option for our JLTi PhonoPre and a major option for our Reference Master Clock for digital players, but after some measurements, and a listening test, decided against it.
12V Gelcells have a huge amount of LF noise, starting lower in frequency than our spectrum analyser can measure (10 Hz) and extending to above 1kHz. But below 1khz, they are FAR FAR noisier than a good solid state shunt regulator, or even a LM317!
Actuality can be very different to opinion...
Regards, Allen (Vacuum State)
We planned to offer a battery pack as a third power option for our JLTi PhonoPre and a major option for our Reference Master Clock for digital players, but after some measurements, and a listening test, decided against it.
12V Gelcells have a huge amount of LF noise, starting lower in frequency than our spectrum analyser can measure (10 Hz) and extending to above 1kHz. But below 1khz, they are FAR FAR noisier than a good solid state shunt regulator, or even a LM317!
Actuality can be very different to opinion...
Regards, Allen (Vacuum State)
Noise-y batteries
The noise spectrum from a battery may well be a factor, as the capacity (A-Hrs) tells you nothing about the current delivery capacity.
During discharge, the reactions occurring at the electrodes may involve local release of gasses or other non-desirable by-products. These don't contribute to the uniform flow of electrons as expected from the electrode under dissolution that supplies electrons or on the collector, and that disruption of uniform flow may produce microscopic spikes in the local current density and hence the overall delivery spectrum will be less than desirable.
Perhaps a different battery technology would produce better results than a gel cell.
John L.
Allen Wright said:
The noise spectrum from a battery may well be a factor, as the capacity (A-Hrs) tells you nothing about the current delivery capacity.
During discharge, the reactions occurring at the electrodes may involve local release of gasses or other non-desirable by-products. These don't contribute to the uniform flow of electrons as expected from the electrode under dissolution that supplies electrons or on the collector, and that disruption of uniform flow may produce microscopic spikes in the local current density and hence the overall delivery spectrum will be less than desirable.
Perhaps a different battery technology would produce better results than a gel cell.
John L.
IMO, this is going in the direction of generalizations. The power supply and its load are a system and have to be examined as a system. What matters for one load circuit might not matter at all for another. Talking about different topologies of regulator is just as worthless. I can build a lousy shunt regulator or a great one. The LM317 is the same way. Implemented one way, it's not very impressive spec-wise. Add a few parts and get the layout right, and it's extremely good- unless you happen to get one with popcorn noise. Then you might as well use a battery. 
In my particular RIAA preamp, the power supply was very important when the bypassing and layout of the amp circuit wasn't handled correctly. When it was, the power supply became irrelevant so long as it was producing the right voltage at any reasonable noise level. Topology didn't matter. IMO, amplifier circuits that are audibly different with different power supplies should be examined for design flaws before worrying much about power supply subtleties.
In my particular RIAA preamp, the power supply was very important when the bypassing and layout of the amp circuit wasn't handled correctly. When it was, the power supply became irrelevant so long as it was producing the right voltage at any reasonable noise level. Topology didn't matter. IMO, amplifier circuits that are audibly different with different power supplies should be examined for design flaws before worrying much about power supply subtleties.
Onvinyl said:
No one can do that, not even you. "Hearing" is, by it's very nature, subjective. Furthermore, the hearing of any given individual is physically colored by the unique characteristics of that individual's anatomy and physiology. Given that, for anyone to state that Component X or Design Y "sounds better" is pointless unless appended by terms equivalent to, "...to me." As an instrument for objective and accurate evaluation of a particular auditory "experience" (for lack of a better term), we "listeners" are notoriously inaccurate and unreliable. It has been well documented that even when presented with well-crafted A-B comparisons, we often can, and do, miss the mark. How else might one explain this, other than that there is some other-than-physical factor at work here?
This is not to say that we are bad judges of what "sounds good". You are, by so far as to be the only standard that matters, the best judge of what sounds good... to you. But to suggest that there are not psychological factors that can affect that judgment is to ignore a reproducible reality.
This thread is going badly I feel.
I doubt there is a single person on this forum who doesn't have a particular style of design - some preferring the warm lies of tubes and others the precision of solid state. Your choice will almost certainly be governed largely by the style of music you enjoy. It is also easy for audiophiles to go way over the top with some specific element of a design whilst leaving others behind, only to have the latter drag the former down with them. I believe the best designers appreciate both valves and solid state, and can select component such that money is placed where it needs to be, and not arbitarily on big capacitor or heatsinks.
There are the general points you mentioned that will determine the perfection of a supply, then you get to influence that through component performance, selection and layout. If you like 'squishy / smooth' supplies, use valve rectifiers and regulators with minimal capacitance and higher output impedance. If you like faster, dynamic supplies do the opposite. The fine details are far too great in number and personal in choice to specify an ideal design for everyone.
You have to know what it is you're listening for in music to start designing supplies to be pleasing to the ear. For example, replacing the old electrolytic capacitors in my guitar amps supply of coarse not only removes the additional hum that's appeared over the decades but also gives it back it's ability to amplify quick transients. Result, a perfectly audible change in the clarity of the sound.
I doubt there is a single person on this forum who doesn't have a particular style of design - some preferring the warm lies of tubes and others the precision of solid state. Your choice will almost certainly be governed largely by the style of music you enjoy. It is also easy for audiophiles to go way over the top with some specific element of a design whilst leaving others behind, only to have the latter drag the former down with them. I believe the best designers appreciate both valves and solid state, and can select component such that money is placed where it needs to be, and not arbitarily on big capacitor or heatsinks.
There are the general points you mentioned that will determine the perfection of a supply, then you get to influence that through component performance, selection and layout. If you like 'squishy / smooth' supplies, use valve rectifiers and regulators with minimal capacitance and higher output impedance. If you like faster, dynamic supplies do the opposite. The fine details are far too great in number and personal in choice to specify an ideal design for everyone.
You have to know what it is you're listening for in music to start designing supplies to be pleasing to the ear. For example, replacing the old electrolytic capacitors in my guitar amps supply of coarse not only removes the additional hum that's appeared over the decades but also gives it back it's ability to amplify quick transients. Result, a perfectly audible change in the clarity of the sound.
Hi,
well, that's my question: what makes a regulator particulary great or lousy according to your experiences. My Question wasn't ntended to simply judge a certain topology, but possibly how to improve it.
I would be very interested which design issues you fixed to make your desgin immune towards regulator desgin, for instance.
It is not my experience that all power supplies sound the same, even if made the best possible. That's what I know because I heard it. My aim is to discuss the reasons, as long as we can track them down in the circuit design.
Sorry, I don't want to sound rude myself, the thread took both a surprising (to me) and unlucky turn, and I wish we can steer towards a technical discussion.
Regards,
Rüdiger
jrn77478 [/i][B] But to suggest that there are not psychological factors that can affect that judgment is to ignore a reproducible reality. [/B][/QUOTE] I didn't suggest this. Music is a something that happens in your mind. The 'psychological factors' play a main role said:
well, that's my question: what makes a regulator particulary great or lousy according to your experiences. My Question wasn't ntended to simply judge a certain topology, but possibly how to improve it.
I would be very interested which design issues you fixed to make your desgin immune towards regulator desgin, for instance.
It is not my experience that all power supplies sound the same, even if made the best possible. That's what I know because I heard it. My aim is to discuss the reasons, as long as we can track them down in the circuit design.
Sorry, I don't want to sound rude myself, the thread took both a surprising (to me) and unlucky turn, and I wish we can steer towards a technical discussion.
Regards,
Rüdiger
SY said:
Good post, SY.
100 % my own view on power supplies.
Power supply matters, mainly in an electrical sense
.. we can measure some differencies, however small, in audio output.
It is only when a power supply is Really Bad, it will effect sound. In a negative way.
And we do not use or want very bad power supply.
Because this can easily be avoided
To say or claim that anybody can hear these differencies, i.e. the sound anybody hear is effected, is another thing.
Because any claim or hypotesis or subjective listening experience
is only a subjective thinking and personal opinion,
as long as it is not verified in some way.
Using some method that can exclude any possible subjective and mind illusive biasing, deception of your mind.
ABX testing is one such method.
There are few other ways, to really get a good and objective result, I am afraid.
Lineup
-------------------------------------------------------------------------------
now how to build an eletrically ultimate power supply is another question
this question has got not one answer, but many
depending on what your goal is
and what is the situation at hand
There are some common requirements that a descent power supplies should adhere to.....
Power supplies are not "ideal" , they are be NON-LINEAR and therefore this can have an affect on a audio circuit....
Power supply noise floor is critical in all types of applications, I design Military power supplies for RF applications and the requirements can be hundreds of pages long...I can tell my boss that a power supply doesn't affect his RF signals and he will first ask me if i fell and hit me head, then he would hand me a box to start packing my office to look for another job.....
IMHO, the key to good high frequency response in Hi-Fi equipment is PHASE LINEARITY..... I want my -3dB POLE placed as high as possible when I design audio gear...lets say 90kHz or better....this is so I have close to no phase shift down at 15kHz... It is this phase shift in such things as the cymbals that ruins the quality of its sound...
When a power supply is slow and has LAG to high frequency transients, such a found in music, you create a POLE and phase shift in the audio response....
Even the fastest loop in SMPS can't keep up with HUGE fast step loads...
What governs the step response is the ESL/ESR of the output hold-up caps initially before the power supply can catch up and recharge the hold-up caps and supply the load current at the same time...
These small delays cause phase shift in audio response....
This is only one power supply aspect out of many that affect audio performance..
Chris
Power supplies are not "ideal" , they are be NON-LINEAR and therefore this can have an affect on a audio circuit....
Power supply noise floor is critical in all types of applications, I design Military power supplies for RF applications and the requirements can be hundreds of pages long...I can tell my boss that a power supply doesn't affect his RF signals and he will first ask me if i fell and hit me head, then he would hand me a box to start packing my office to look for another job.....
IMHO, the key to good high frequency response in Hi-Fi equipment is PHASE LINEARITY..... I want my -3dB POLE placed as high as possible when I design audio gear...lets say 90kHz or better....this is so I have close to no phase shift down at 15kHz... It is this phase shift in such things as the cymbals that ruins the quality of its sound...
When a power supply is slow and has LAG to high frequency transients, such a found in music, you create a POLE and phase shift in the audio response....
Even the fastest loop in SMPS can't keep up with HUGE fast step loads...
What governs the step response is the ESL/ESR of the output hold-up caps initially before the power supply can catch up and recharge the hold-up caps and supply the load current at the same time...
These small delays cause phase shift in audio response....
This is only one power supply aspect out of many that affect audio performance..
Chris
lineup said:
It is only when a power supply is Really Bad, it will effect sound. In a negative way.
And we do not use or want very bad power supply.
Because this can easily be avoided
I fear this is an oversimplification. Your decision, which parts belong to the reg and which belong to the amp, doesn't affect the electron flow (*if* there is such thing in reality, lol), so I think Conrad is right in claiming to look at the whole ciruit, reg + amp together.
Rüdiger
"Really Bad" is a relative term. If I have a supply with half a volt of ripple and 20mV of hash, that's a superb supply for a push-pull output stage, but a lousy one for an input cascode. But it all gets back to the criteria in your first post- if a supply is quiet enough, low impedance enough, laid out correctly, grounded correctly, and is stable, it will sound no different than a different kind of supply with similar performance (IME). There's nothing mysterious about it, just straightforward engineering.
That said, it's just as important (maybe even more so) to make the circuit itself as insensitive to the supply as practical.
That said, it's just as important (maybe even more so) to make the circuit itself as insensitive to the supply as practical.
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.