Why a linear power supply?

[kisjano]

Finish this upgrade,change z3 pico psu.I use 2pcs lt 1084 linear power (just motherboard)Fantastic upgrade,and sound!!!!around 40 $

this power work well with my audio gd master series system.

[sandyk]

kisjano

I presume that is a simple mains filter to the left of the transformers as used in many SMPS players ?

A suggestion would be to use different colour wires for A.C. wiring, +, - and E to avoid silly mistakes.

Green or green with a yellow trace for example is best for earth/0 volts. Black,white or even dark blue are ideal for -VE rails, and red or orange for +VE rails.

The wires from a short length of mains 7.5A mains cable (light blue, brown, and green with a yellow trace) are ideal for internal mains wiring to on/off switches etc. I normally switch both sides of the mains supply where practical.

I have posted several projects in another forum that were built by many members. Most mistakes and damage were caused when using similar colours for supply rails between PCBs. Almost all of them were avoidable.

Many happened when playing around later, after initially getting the project working.

 

Regards

Alex

[4est]

Ugh...

 

Many happened when playing around later, after initially getting the project working.

[JJJ]

Plus twist all the wiring pairs to minimise EMR - then try to remember which wire is which!

[sandyk]

I should have added that many of the constructors "bread boarded" the various modules first, then perhaps being elated that it actually worked as hoped, they were less careful when wiring them up in their final locations.

[mbain]

Can anyone explain the difference between a regulated power supply and a linear power supply?

[barrows]

Can anyone explain the difference between a regulated power supply and a linear power supply?

 

There is no difference really. A linear power supply refers to one which is not based on switching technology. A linear supply can be regulated, or not.

An unregulated linear power supply will vary in its output voltage, both with AC line variances and with load (current draw) variances.

A regulated linear power supply will include active circuitry which holds the output voltage stabile regardless of load or line conditions (ideally, how well it can do this is limited to some extent).

Additionally, while most switching power supplies are regulated, they do not have to be, so there are also both regulated and unregulated switching supplies.

[mbain]

There is no difference really. A linear power supply refers to one which is not based on switching technology. A linear supply can be regulated, or not.

An unregulated linear power supply will vary in its output voltage, both with AC line variances and with load (current draw) variances.

A regulated linear power supply will include active circuitry which holds the output voltage stabile regardless of load or line conditions (ideally, how well it can do this is limited to some extent).

Additionally, while most switching power supplies are regulated, they do not have to be, so there are also both regulated and unregulated switching supplies.

 

Thanks so much Barrows for your reply. I asked the question as I found a " precision regulated DC power supply" with a 3 amp, 13.8v output. The unit was produced by Tripp Lite, model PR-3a. The scant amount of information I can find says it produces fully regulated, low ripple output. I was wondering if this would be a suitable supply for any of my audio components ie. A supply yielding superior sq ultimately. Any thoughts?

 

Mark

 

P.S.

 

I am also in the market to build a pre amp compatible with the DIY Amp kit momoblocks I recently finished. I have been considering the Pass B1 or possibly a Twisted Pear design. I know you have some experience in this area. Any recommendatios?

[barrows]

Thanks so much Barrows for your reply. I asked the question as I found a " precision regulated DC power supply" with a 3 amp, 13.8v output. The unit was produced by Tripp Lite, model PR-3a. The scant amount of information I can find says it produces fully regulated, low ripple output. I was wondering if this would be a suitable supply for any of my audio components ie. A supply yielding superior sq ultimately. Any thoughts?

 

Mark

 

P.S.

 

I am also in the market to build a pre amp compatible with the DIY Amp kit momoblocks I recently finished. I have been considering the Pass B1 or possibly a Twisted Pear design. I know you have some experience in this area. Any recommendatios?

 

Power supply design is a complicated business. For audio components, you are really listening to the power supply, so the better the power supply, the better the sound. This means you want a no noise, no output impedance supply, ideally, with absolutely stable voltage regardless of load. The demands which high end audio components place on a power supply are beyond all other uses, besides high precision measurement instrumentation, some military hardware (like highly sensitive sonar and radar applications) and some medical tech.

I would not expect any off the shelf industrial power supply to be even adequate for high end audio purposes. When something says "low ripple" but does not specify what that ripple actually is, it is pretty accurate to assume that the ripple is not really that impressive by audio standards.

 

Personally I do not use preamps anymore, so have not investigated them in a number of years. Primary importance is making sure a preamp matches what your amp needs. Discussing all the possibilities of preamp design and component matching here, well, you get the idea. If you are interested in building a preamp, I would recommend doing a lot of research at the diyaudio.com website.

[Superdad]

Power supply design is a complicated business. For audio components, you are really listening to the power supply, so the better the power supply, the better the sound. This means you want a no noise, no output impedance supply, ideally, with absolutely stable voltage regardless of load. The demands which high end audio components place on a power supply are beyond all other uses, besides high precision measurement instrumentation, some military hardware (like highly sensitive sonar and radar applications) and some medical tech.

 

Having used a number of lab/bench/industrial/ham radio linear supplies with older and newer Mac minis, and now having a MUCH cleaner/quieter supply, I can attest that even computers for audio benefit greatly from a really good supply. I was surprised at the difference.

Sadly, there are also a lot of inexpensive supplies in nice looking aluminum cases which don't actually measure much better than the $60-$100 industrial supplies. It is going to take measurements and published scope shots to educate the consumer.

[Paul R]

Okay- so why not just look more at isolation? For example, a computer outputting S/PDIF over an optical connection should be totally decoupled from anything but gross power supply effects, no? If not, how are they coupled? Will simple strategies like ASRC reclocking take care of the S/PDIF transmission issues?

 

I think that optical coupling eliminates the issues with computer power supplies, and means the DAC, Pre/AMP power supplies are where to concentrate effort. And those are pretty well known areas these days, right?

 

-Paul

 

 

Having used a number of lab/bench/industrial/ham radio linear supplies with older and newer Mac minis, and now having a MUCH cleaner/quieter supply, I can attest that even computers for audio benefit greatly from a really good supply. I was surprised at the difference.

Sadly, there are also a lot of inexpensive supplies in nice looking aluminum cases which don't actually measure much better than the $60-$100 industrial supplies. It is going to take measurements and published scope shots to educate the consumer.

[sandyk]

I think that optical coupling eliminates the issues with computer power supplies,

 

Paul

Optical SPDIF is only as good as it's power supply due to the need to be converted to Optical. The same applies at the Receive end, as well as the quality of the Toslink cable where many prefer glass over plastic fibres.

The Coax SPDIF and Optical Out of a soundcard is likely to have the benefit of further PSU improvements in the card itself.

 

Alex

[4est]

Realizing that this is a contentious topic, I am still of the mind that the power supplies may help beyond just noise rejection and transmission. Beyond cost and weight, there is almost never a disadvantage to better/improved/separated power supplies.

Okay- so why not just look more at isolation? For example, a computer outputting S/PDIF over an optical connection should be totally decoupled from anything but gross power supply effects, no? If not, how are they coupled? Will simple strategies like ASRC reclocking take care of the S/PDIF transmission issues?

 

I think that optical coupling eliminates the issues with computer power supplies, and means the DAC, Pre/AMP power supplies are where to concentrate effort. And those are pretty well known areas these days, right?

 

-Paul

[Paul R]

I totally disagree here - if the quality of the computer power supply is sufficient to modulate the laser correctly, I am aware of no further effects that the computer power supply can propagate over the optical connection. No static, no noise, no electrical connections of any type.

 

Optical fibre does not conduct anything save light. There is no electrical connection whatsoever between the server and DAC.

 

ASRC is a buffer based re-clocking protocol and should serve to control jitter about as well as synch USB.

 

The only thing I can see that degrades optical connections and often makes them appear to sound worse than coaxial connections is the quality of the components used in the optical receivers. Those are, to understate the case, very cheap.

 

When you listen over a higher quality optical connection, I personally think it compares with the best of Async USB connections. But that has nothing at all to do with the power supply in the computer.

 

In the same way, arguments about whether optical cables made of glass or plastic sound better are irrelevant to whether the computer power supply can make a difference.

 

Paul

Optical SPDIF is only as good as it's power supply due to the need to be converted to Optical. The same applies at the Receive end, as well as the quality of the Toslink cable where many prefer glass over plastic fibres.

The Coax SPDIF and Optical Out of a soundcard is likely to have the benefit of further PSU improvements in the card itself.

 

Alex

[Paul R]

I am narrowing the focus to the computer power supply, not all the other power supplies down the line. Do you think the computer power supply would make a difference with an optical transmission?

 

Realizing that this is a contentious topic, I am still of the mind that the power supplies may help beyond just noise rejection and transmission. Beyond cost and weight, there is almost never a disadvantage to better/improved/separated power supplies.

[4est]

What I am suggesting is that independent power supplies may help the computer components from interacting, and low Z may help internal latency or at least help control gHz transmission and read/writes or something. I do not know how much any of this really effects things in the first place, yet it seems to. In lieu of that, I feel I must remain open minded and thorough until it is understood better. As I said, improved power supplies are unlikely to hinder.

I am narrowing the focus to the computer power supply, not all the other power supplies down the line. Do you think the computer power supply would make a difference with an optical transmission?

[Superdad]

Okay- so why not just look more at isolation? For example, a computer outputting S/PDIF over an optical connection should be totally decoupled from anything but gross power supply effects, no? If not, how are they coupled? Will simple strategies like ASRC reclocking take care of the S/PDIF transmission issues?

 

I think that optical coupling eliminates the issues with computer power supplies, and means the DAC, Pre/AMP power supplies are where to concentrate effort. And those are pretty well known areas these days, right?

 

-Paul

 

Sorry Paul, there is a WHOLE LOT more wrong with S/PDIF (both on the send and receive sides) than can be cured just with better optical couplers at each end. Solving receiver and impedance issues and doing really good ASRC takes care. I believe there are but a handful of brands that even come close to getting it really right at either end. (I'll try to get John Swenson to jump in here with the technical facts.)

 

So even if optical connections do isolate well enough to lessen the importance of a computer PS:

a) One still has to get a decent quality TOSLINK signal out of the computer;

b) You have to feed it to a DAC with a more perfect optical S/PDIF input.

 

I happen to think that good USB, fed from a quiet and clean source (CAPS, little ARM Linux module, or tweaked Mac mini) to a well done async-USB input module inside a DAC is currently the most economical way to get really close to SOTA. It also readily allows for sample rates of 352.8/384 and DSD (plastic TOSLINK is only good up to 24/96).

 

One easy experiment that anyone with a TOSLINK-equipped computer and DAC can do is to compare player s/w over that interface. Ignoring for the moment that it may/will sound worse than your standard connection, can you hear a difference between straight iTunes and anything better? If optical was to ameliorate all benefits of a cleaner/quieter computer (including power supply), then nothing you do on the computer should make any SQ difference.

[Melvin]

Sorry Paul, there is a WHOLE LOT more wrong with S/PDIF (both on the send and receive sides) than can be cured just with better optical couplers at each end. Solving receiver and impedance issues and doing really good ASRC takes care. I believe there are but a handful of brands that even come close to getting it really right at either end. (I'll try to get John Swenson to jump in here with the technical facts.)

 

So even if optical connections do isolate well enough to lessen the importance of a computer PS:

a) One still has to get a decent quality TOSLINK signal out of the computer;

b) You have to feed it to a DAC with a more perfect optical S/PDIF input.

 

I happen to think that good USB, fed from a quiet and clean source (CAPS, little ARM Linux module, or tweaked Mac mini) to a well done async-USB input module inside a DAC is currently the most economical way to get really close to SOTA. It also readily allows for sample rates of 352.8/384 and DSD (plastic TOSLINK is only good up to 24/96).

 

One easy experiment that anyone with a TOSLINK-equipped computer and DAC can do is to compare player s/w over that interface. Ignoring for the moment that it may/will sound worse than your standard connection, can you hear a difference between straight iTunes and anything better? If optical was to ameliorate all benefits of a cleaner/quieter computer (including power supply), then nothing you do on the computer should make any SQ difference.

 

For quite some time I used the SOtM dx-USB HD converter (and companion battery supply) which passes 24/192 over toslink. Coupled to my QuteHD, which also accepts 24/192 over toslink, it was a revelation. It sounded fantastic and not the mushy mess many report via toslink. FWIW, toslink jitter was measured at 38ps (not too shabby) in this review of the SOtM converter: USB-to-S/PDIF Convertors.

 

I only bring this up because toslink can and does do 24/192. (Apple limits toslink to 24/96 via its optical out.) I don't think many people have heard well-implemented optical. Just sayin' ..

[Superdad]

For quite some time I used the SOtM dx-USB HD converter (and companion battery supply) which passes 24/192 over toslink. Coupled to my QuteHD, which also accepts 24/192 over toslink, it was a revelation. It sounded fantastic and not the mushy mess many report via toslink. FWIW, toslink jitter was measured at 38ps (not too shabby) in this review of the SOtM converter: USB-to-S/PDIF Convertors.

 

I only bring this up because toslink can and does do 24/192. (Apple limits toslink to 24/96 via its optical out.) I don't think many people have heard well-implemented optical. Just sayin' ..

 

Interesting that you were able to run that speed over plastic TOSLINK cable. But if we follow Paul's logic, you would not have needed that extra $400 battery supply for the $450 SOtM USB>TOSLINK converter. It should sound the same when powered via its USB input since it all gets reclocked at the DAC end. But did that battery supply make a difference?…

 

And I see from your signature that you now use the less expensive Bel Canto mLink--via coax S/PDIF as it does not have optical. What made you switch?

[Melvin]

Interesting that you were able to run that speed over plastic TOSLINK cable. But if we follow Paul's logic, you would not have needed that extra $400 battery supply for the $450 SOtM USB>TOSLINK converter. It should sound the same when powered via its USB input since it all gets reclocked at the DAC end. But did that battery supply make a difference?…

 

And I see from your signature that you now use the less expensive Bel Canto mLink--via coax S/PDIF as it does not have optical. What made you switch?

 

If the battery supply made a difference, it was subtle. I've also used the SOtM converter with my KingRex ps and found the same. After using the converter for six months or so I removed the battery (which needed to be replaced) and thought I would take a hit in SQ .. not so. I was very surprised and listened extensively before replacing the ps. I doubt I could reliably tell the difference in a DBT.

 

I decided to purchase the mLink because I was in a tweaky mood, it was 1/2 price ($187), and I was curious. The SOtM was the only converter I've owned and if the mLink proved to be a disappointment, I would just use it in my secondary system. The mLink did not disappoint. Is it better than the SOtM dX-USB HD? I don't know .. it's different. My preference is the mLink right now.

 

 

edit: My point is there is something positive to be said about well-engineered optical. Removing the electrical connection from the computer has benefits.

[Superdad]

My point is there is something positive to be said about well-engineered optical. Removing the electrical connection from the computer has benefits.

 

I certainly agree with you there. Really good galvanic isolation on USB inputs is possible, but for some reason, very few firms do it properly. While such can and will greatly reduce the ability of pernicious ground noise to get into to the DAC, I think that what happens at the computer end will still make a difference.

If I had a really great optical connection (ideally straight off the computer) and a DAC with a really good TOSLINK input, then I'd love to try that is see if software players and other computer variables (including PS) still made a difference.

[Paul R]

(grin) You might notice I did not reference TOSlink, but instead said optical transmissions. Optical transmissions are not limited to TOSLink, or even USB speeds. not b a very long way.

 

I am not all at convinced you are correct about optical vs. USB Alex, and wonder if we might not be jumping to some conclusions. Of course, I agree power supplies make a hell of a difference with USB, but not so much with optical. S/PDIF's one major failing is timing, and good buffering (or ASRC) can and does address that weak spot. I contend that the best part of optical transmission is that it really isolates the two devices and eliminates any noise or stray RF interference.

 

Truly, I would like to see more development going into that field, as I think it holds great promise.

 

Full disclosure. Optical transmissions are in my professional field, while power supplies are not, so I do have a bit of a fondness for using that particular technology.

 

Sorry Paul, there is a WHOLE LOT more wrong with S/PDIF (both on the send and receive sides) than can be cured just with better optical couplers at each end. Solving receiver and impedance issues and doing really good ASRC takes care. I believe there are but a handful of brands that even come close to getting it really right at either end. (I'll try to get John Swenson to jump in here with the technical facts.)

 

So even if optical connections do isolate well enough to lessen the importance of a computer PS:

a) One still has to get a decent quality TOSLINK signal out of the computer;

b) You have to feed it to a DAC with a more perfect optical S/PDIF input.

 

I happen to think that good USB, fed from a quiet and clean source (CAPS, little ARM Linux module, or tweaked Mac mini) to a well done async-USB input module inside a DAC is currently the most economical way to get really close to SOTA. It also readily allows for sample rates of 352.8/384 and DSD (plastic TOSLINK is only good up to 24/96).

 

One easy experiment that anyone with a TOSLINK-equipped computer and DAC can do is to compare player s/w over that interface. Ignoring for the moment that it may/will sound worse than your standard connection, can you hear a difference between straight iTunes and anything better? If optical was to ameliorate all benefits of a cleaner/quieter computer (including power supply), then nothing you do on the computer should make any SQ difference.

[Paul R]

Not exactly Alex - the converter is in the case, affected by hash on the USB transmission of course. My favorite USB DAC has a battery in it powering the critical sections. USB anything benefits from clean power. I never said anything to dispute that.

 

I am a little surprised you believed that even TOSlink is limited to 24/96. The media can and always has been able to handle higher speeds.

 

-Paul

 

 

Interesting that you were able to run that speed over plastic TOSLINK cable. But if we follow Paul's logic, you would not have needed that extra $400 battery supply for the $450 SOtM USB>TOSLINK converter. It should sound the same when powered via its USB input since it all gets reclocked at the DAC end. But did that battery supply make a difference?…

 

And I see from your signature that you now use the less expensive Bel Canto mLink--via coax S/PDIF as it does not have optical. What made you switch?

[Superdad]

I am a little surprised you believed that even TOSlink is limited to 24/96. The media can and always has been able to handle higher speeds.

 

Of course optical can run at outrageous speeds (where's my fiber from the curb for broadband?!). I just recall reading that the cheap plastic optical cables and the receivers used for most audio gear were limited to 24/96. Melvin corrected me on this as well.

 

I think we are mostly in agreement Paul. It is just that my understanding has been that timing/reclocking are by no means the only challenges to good S/PDIF interfaces, nor are they simple to get right. While you say you are talking about optical transmission in general, present-day supported audio interfaces are thin on the ground.

 

USB versus S/PDIF is a balance of compromises for sure. But since this is a computer audio forum, and since a majority are using operating systems and player software to send audio out the USB port, I like to focus on really good DAC-internal async-USB>I2S interfaces and leave behind the delicate extra steps of converting the ancient S/PDIF format.

 

Now if you want to talk about a custom computer processor board with an optical output right on the PCIe bus… But with what clocking and protocol? How about optical Thunderbolt or USB? You get galvanic isolation with those.

 

Also, Ethernet PHY transformers do a very nice job of isolation. Plus running long cables is no problem and there are ways to improve the rest of the software/hardware aspects there. Lots of room for innovation, and optical--but just not TOSLINK over S/PDIF--could be a welcome part of new solutions. IMHO.

 

Regards,

ALEX

[Paul R]

The very inexpensive optical interface in most computers and audio gear is the limiting factor, yes. But spend a tad bit more and voila- magnificent transfer. But I don't even have problems with TOSLink actually. Especially well implemented TOSLink. In fact, I use TOSlink in my bedroom system, and it sounds very good indeed.

 

S/PDIF isn't really an ancient protocol nor is it actually a bad protocol for streaming data around, but it can obviously be implemented a lot better in some cases than in others. Indeed, some DACs sound better through their S/PDIF interfaces than through USB, or require one to convert USB to S/PDIF before they will accept the music transmission. Berkeley for example.

 

As for interfaces, well - almost every platform in the world can use optical transmissions, and the better ones, Servers, midrange, and mainframe class computers, as well as routers, switches, and so forth, have very good connectivity indeed. Surely, the ira od spending a few hundred dollars on a quality optical interface could be presented in a way to make sense.

 

By the way, just playing devil's advocate, but I am not so sure that the majority of DACs in the world are connected via USB either, though I obviously like and even prefer USB for several reasons. But until just the past few years, the top end DACs - with some notable exceptions - did not even have USB options. And of course, the best USB implementations pay a lot of attention to power as well as timing.

 

But getting back to the point, how do you think an improved power supply would improve optical connections from the computer to the DAC? I grant that it might, and more or less agree with Forrest' logic. But I am really interested in other people's opinion.

 

Of course optical can run at outrageous speeds (where's my fiber from the curb for broadband?!). I just recall reading that the cheap plastic optical cables and the receivers used for most audio gear were limited to 24/96. Melvin corrected me on this as well.

 

I think we are mostly in agreement Paul. It is just that my understanding has been that timing/reclocking are by no means the only challenges to good S/PDIF interfaces, nor are they simple to get right. While you say you are talking about optical transmission in general, present-day supported audio interfaces are thin on the ground.

 

USB versus S/PDIF is a balance of compromises for sure. But since this is a computer audio forum, and since a majority are using operating systems and player software to send audio out the USB port, I like to focus on really good DAC-internal async-USB>I2S interfaces and leave behind the delicate extra steps of converting the ancient S/PDIF format.

 

Now if you want to talk about a custom computer processor board with an optical output right on the PCIe bus… But with what clocking and protocol? How about optical Thunderbolt or USB? You get galvanic isolation with those.

 

Also, Ethernet PHY transformers do a very nice job of isolation. Plus running long cables is no problem and there are ways to improve the rest of the software/hardware aspects there. Lots of room for innovation, and optical--but just not TOSLINK over S/PDIF--could be a welcome part of new solutions. IMHO.

 

Regards,

ALEX