Why a linear power supply?

[Superdad]

Right now I use a Toshiba laptop for that purpose (running on battery when playing music). My goal is to replace the laptop with one of those wireless (Wi-Fi) HDDs that have an available app for android or iOS devices to manage them.

 

Sorry George, I am confused: Your plan is to use a wi-fi hard drive as your music server? How would that connect to your DAC or to your stereo?

[davidbeinct]

Sorry George, I am confused: Your plan is to use a wi-fi hard drive as your music server? How would that connect to your DAC or to your stereo?

 

I assume he'd use AirPlay to get the stream from the iDevice to the stereo.

[Superdad]

You posted the above after Eloise posted:

Quote Originally Posted by Audio_ELF View Post

John, I have to ask this; but were you doing any of the comparisons blind or always knew what PSU you were listening to?

 

Then some other people complained.

 

You put details of what stuff is being done with a power supply, and sing its praises. I don't see a problem with her question. You intend to market this at some point. Seems like a reasonable question to me. So far I wouldn't call what you are doing out of bounds, but you are drumming up interest in your product. I for one am glad to have information and a play by play to some extent of the development process. But to act as if Eloise's question was out of bounds is not at all called for.

 

My own thoughts upon reading Mr. Swenson's post wasn't about blind auditioning. But it immediately was, "I wonder if it makes any differences one can measure in the analog result." My impression thus far is Mr. Swenson does things that any designer of power supplies knows is making for a better quieter cleaner supply. Then listening along with you and perhaps others to the result to evaluate whether it is an improvement or not. But not worrying about measuring the analog results. Seemingly with the often felt assumption that those measurements don't show things that we can hear. So auditioning is the final ultimate arbiter.

 

That is fine, and anyone's choice. But not all agree, and simply asking such questions is not out of place especially when someone is touting how good the results of such experimentation are for sound quality for a product they will eventually sale.

 

Hi Dennis:

You took my comment to Eloise slightly out of context by not acknowledging that I also chided SandyK at the same time for jumping in with a request for me to rip discs onto the computer with and without the better supply.

I took Eloise's question as another attempt to pull a thread into a debate over sighted versus blind listening tests. You know very well how many threads we all participated in on that subject.

 

But what does any of that have to do with whether we are ready to publish comparative test measurements of a PS that is not yet in production? John certainly has measured his prior supplies, and I stated up-thread that I'd like to publish actual tests of this new version (which uses a number of better parts and techniques than his prior designs).

 

Yet I do see that your concerns are twofold:

a) That we are not making measurements of the ANALOG result;

b) That talking about some things we heard in the design process--without it having been a "blind" test--makes some people uncomfortable.

 

I don't think "b)" needs any comment.

As for "a)", can you show me where any hi-fi company has published measurements comparing their power supply's effect on the analog output--even the PS of a preamp or power amp, or of a DAC, let alone the analog effect for a computer PS? I am sure one can measure PS quality in the analog output of a big power amp, but for the rest of devices I am not so sure. Of course the quality of any internal or external PS can be quantified with many relevant measurements, and that is certainly part of John's design process.

 

And while it is not too hard to understand why a quiet and robust supply performs (measurably and audibly) better than a small SMPS built for a few bucks and sandwiched inside next to the motherboard, it is a bit harder for any of us to understand or quantify what we hear between a multi-purpose $90 linear supply and something more refined like this or a Paul Hynes or a Teddy Pardo. I'm not here to defend all the aspects of high-end audio which you are skeptical about. People can decide for themselves what they do/do-not hear and what their priorities are.

 

 

By the way, to those wondering about really good SMPS: They do exist, but are not particularly cheap. Some of the best ones John and I have ever found were the "ultra-low-noise" series from this company: Daitron AC/DC Power Supply - Daitron Incorporated The HFS150A series are $285 each in OEM 10 pc. quantity. Even at higher quantity they are still a bit pricy to built into a DAC or preamp. Part cost for a very well done linear would still be less--though not as compact.

[JohnSwenson]

If talking about this design is making people uncomfortable I can stop talking about it. My original post in this thread was a response to a direct question about it and was not meant as "advertising".

 

As to testing, I DO make measurements. I haven't done any yet since the prototype was assembled at Alex's place and as I mentioned I didn't want to bring any test equipment with me. Since I'm not a big company with a large budget most of my test equipment is 20 year old stuff off of ebay, most of which is big and heavy, so hauling it around is not something I like to do.

 

Since he has the prototype I don't have anything to measure at this point. I haven't had time to build one up for myself. So it is going to be a little while before measurements happen.

 

As to how I do measurements, I built a very low noise wideband differential amplifier that is powered off of batteries, and sits in a metal box. (I like 1/4" thick aluminum, why is a very long story) I usually plug this into a spectrum analyzer rather than a scope, it's usually very difficult to see what is happening with just a scope. (BTW I have a 2467B, off ebay of course) The spectrum analyzer is a behemoth HP, very good, but huge and heavy (off ebay of course)

 

I'm sure I will be able to measure differences in the actual supply rails, but measuring analog on the output of a DAC is a different matter. In order to do this well you have come up with a guess as to what the interaction is, figure out a test waveform for that, figure out the best way to measure what you think is happening, set it up, do the measurements, and repeat 50 times. It CAN be done but takes a LONG time. It's not so simple as play some music and look at it on a scope. The effects from these types of power supply differences are small and you really need to "zoom in" on what they are with specific test signals and measurement techniques. I have done this a few times and it usually takes 3 months or so to figure out what is really happening and figure out how to measure it.

 

John S.

[jtwrace]

If talking about this design is making people uncomfortable I can stop talking about it. My original post in this thread was a response to a direct question about it and was not meant as "advertising".

 

As to testing, I DO make measurements. I haven't done any yet since the prototype was assembled at Alex's place and as I mentioned I didn't want to bring any test equipment with me. Since I'm not a big company with a large budget most of my test equipment is 20 year old stuff off of ebay, most of which is big and heavy, so hauling it around is not something I like to do.

 

Since he has the prototype I don't have anything to measure at this point. I haven't had time to build one up for myself. So it is going to be a little while before measurements happen.

 

As to how I do measurements, I built a very low noise wideband differential amplifier that is powered off of batteries, and sits in a metal box. (I like 1/4" thick aluminum, why is a very long story) I usually plug this into a spectrum analyzer rather than a scope, it's usually very difficult to see what is happening with just a scope. (BTW I have a 2467B, off ebay of course) The spectrum analyzer is a behemoth HP, very good, but huge and heavy (off ebay of course)

 

I'm sure I will be able to measure differences in the actual supply rails, but measuring analog on the output of a DAC is a different matter. In order to do this well you have come up with a guess as to what the interaction is, figure out a test waveform for that, figure out the best way to measure what you think is happening, set it up, do the measurements, and repeat 50 times. It CAN be done but takes a LONG time. It's not so simple as play some music and look at it on a scope. The effects from these types of power supply differences are small and you really need to "zoom in" on what they are with specific test signals and measurement techniques. I have done this a few times and it usually takes 3 months or so to figure out what is really happening and figure out how to measure it.

 

John S.

Keep talking. I look forward to the measurements very much.

[Superdad]

If talking about this design is making people uncomfortable I can stop talking about it. My original post in this thread was a response to a direct question about it and was not meant as "advertising".

 

John: Don't sweat it at all. I'm sure it was my posts--talking about it as our planned product--that ruffled one person's feathers. A lot of people here are really interested in the topic and the various alternatives. I'll call you tonight or this weekend. I've been trying to get Debian to boot on my mini so I could try out Miska's NAA, but I guess I'm just a boob when it come to Linux boot drivers!

Alex

[nige2000]

oh definitely keep talking, interesting topic

 

my post got buried

 

ive done some experimenting with lab bench linear supplies 13.8v (reduced for application),

although the few ive tested have had varying results

generally speaking these linears have had a more than positive effect on the sound quality of the pc's.

 

also experimenting with batteries which seem really good too

 

having multiple power supplies to individual components (usb port, operating system SSD cpu etc), lowering current demand and voltage seem to help sound quality too

 

i dont have a whole lot of technical knowledge, i just do a lot of experimenting

 

Any one got any comments or does any of this make sense?

[esldude]

Hi Dennis:

You took my comment to Eloise slightly out of context by not acknowledging that I also chided SandyK at the same time for jumping in with a request for me to rip discs onto the computer with and without the better supply.

I took Eloise's question as another attempt to pull a thread into a debate over sighted versus blind listening tests. You know very well how many threads we all participated in on that subject.

 

I didn't leave sandyK's comment out to remove context. First I didn't see Eloise's question as an attempt to pull the discussion anywhere. That does not seem to be the way she operates. I took it as a question about how John Swenson goes about design decisions. So her question is about the design process and how decisions are made. Sandyk's is more along the line of "hey, have you used it to do this thing I would like to do". Which isn't about the design/improve/ test sequence. So I considered it irrelevant.

 

But what does any of that have to do with whether we are ready to publish comparative test measurements of a PS that is not yet in production? John certainly has measured his prior supplies, and I stated up-thread that I'd like to publish actual tests of this new version (which uses a number of better parts and techniques than his prior designs).

 

Yet I do see that your concerns are twofold:

a) That we are not making measurements of the ANALOG result;

b) That talking about some things we heard in the design process--without it having been a "blind" test--makes some people uncomfortable.

 

I don't think "b)" needs any comment.

As for "a)", can you show me where any hi-fi company has published measurements comparing their power supply's effect on the analog output--even the PS of a preamp or power amp, or of a DAC, let alone the analog effect for a computer PS? I am sure one can measure PS quality in the analog output of a big power amp, but for the rest of devices I am not so sure. Of course the quality of any internal or external PS can be quantified with many relevant measurements, and that is certainly part of John's design process..

 

Now, I posted that analog output measurements were a question that immediately come into my mind. I didn't post the question however for the very reasons you describe. It is a prototype, it isn't finished etc. etc. I only mentioned it when you seemed to think Eloise's questions weren't valid. And were I asking it that too would be to see how John proceeds with his design out of curiosity.

 

You are right other companies don't published specs on what their upgrade, or heavy duty or optional power supplies do. I think they should. I hope you do when you go to market. You have to admit, if you can provide some honest before and after measurements showing improvements of the resulting music signal it would be a big plus for selling these PSU's.

[elcorso]

Sorry if I sound like a broken record ... but this thread has rather turned into pre-release promotional material for a Superdad / John Swenson commercial product.

 

Eloise

 

I remember from Chris C., "if members are interested they are allowed".

 

And WE ARE...!

 

Roch

[elcorso]

You are right other companies don't published specs on what their upgrade, or heavy duty or optional power supplies do. I think they should. I hope you do when you go to market. You have to admit, if you can provide some honest before and after measurements showing improvements of the resulting music signal it would be a big plus for selling these PSU's.

 

You and your measurements...

 

There is a lot of incredible audio gear products where those measurements don't coincide with what you listen. And the opposite also.

 

The paper accept what you write on it!

 

Roch

[esldude]

If talking about this design is making people uncomfortable I can stop talking about it. My original post in this thread was a response to a direct question about it and was not meant as "advertising".

 

As to testing, I DO make measurements. I haven't done any yet since the prototype was assembled at Alex's place and as I mentioned I didn't want to bring any test equipment with me. Since I'm not a big company with a large budget most of my test equipment is 20 year old stuff off of ebay, most of which is big and heavy, so hauling it around is not something I like to do.

 

Since he has the prototype I don't have anything to measure at this point. I haven't had time to build one up for myself. So it is going to be a little while before measurements happen.

 

As to how I do measurements, I built a very low noise wideband differential amplifier that is powered off of batteries, and sits in a metal box. (I like 1/4" thick aluminum, why is a very long story) I usually plug this into a spectrum analyzer rather than a scope, it's usually very difficult to see what is happening with just a scope. (BTW I have a 2467B, off ebay of course) The spectrum analyzer is a behemoth HP, very good, but huge and heavy (off ebay of course)

 

I'm sure I will be able to measure differences in the actual supply rails, but measuring analog on the output of a DAC is a different matter. In order to do this well you have come up with a guess as to what the interaction is, figure out a test waveform for that, figure out the best way to measure what you think is happening, set it up, do the measurements, and repeat 50 times. It CAN be done but takes a LONG time. It's not so simple as play some music and look at it on a scope. The effects from these types of power supply differences are small and you really need to "zoom in" on what they are with specific test signals and measurement techniques. I have done this a few times and it usually takes 3 months or so to figure out what is really happening and figure out how to measure it.

 

John S.

 

I don't wish for you to stop talking about it. Interested in how you go about such things.

 

I take it you have ideas of what might help make a better PS. You implement them. You take it to Superdad, you both see if you think it sounds better. If so you might make other changes or adapt it for making a product. Then you will perhaps do the measurements to see what is really different at the output or not.

 

Nothing wrong with that. Is interesting to me what your course of action is. In modding equipment in the past, that is also more or less how I went about it.

[Audio_ELF]

I remember from Chris C., "if members are interested they are allowed".

As I recall it it's more along the lines of direct questions can be answered...

 

Anyway it was more when discussion started about cost and how it would be packaged which turned into a pre-release promotion while adding little (IMO) to the original question of "why a Linear PSU".

 

Anyway it's not my site to moderate...

 

Eloise

[acg]

John and Alex, please continue with the technical discussion (maybe leave out the price talk). I don't have a mac but I am interested. Very interested.

 

Anthony

[Superdad]

Anyway it was more when discussion started about cost and how it would be packaged which turned into a pre-release promotion while adding little (IMO) to the original question of "why a Linear PSU".

 

I suppose you are right Eloise. It just that in recent days I have been getting a lot of PM asking for pictures and details, and since John talked about his visit and what we are doing I figured I'd clarify and take care of some curious inquiries at the same time. A two birds with one stone sort of thing.

I promise not to promote further in this thread. As you say, direct questions can be answered…

 

Best,

ALEX

[Jud]

John: Don't sweat it at all. I'm sure it was my posts--talking about it as our planned product--that ruffled one person's feathers. A lot of people here are really interested in the topic and the various alternatives. I'll call you tonight or this weekend. I've been trying to get Debian to boot on my mini so I could try out Miska's NAA, but I guess I'm just a boob when it come to Linux boot drivers!

Alex

 

What's a Linux boot driver?

[sandyk]

John and Alex, please continue with the technical discussion (maybe leave out the price talk). I don't have a mac but I am interested. Very interested.

 

Anthony

I believe that most members are interested, even those Aussie banana benders who love some weird device named a Killerdac. Seriously though, John is a very well respected E.E. who is not prone to making unusual subjective claims.

Why should a highly qualified E.E. be disqualified from saying what he hears, before he gets deeply into a test regimen, that may, or may not, confirm what he firmly believes he hears ?

I personally would like to see any technical details and measurements, including "screen grabs" that John is able to supply when a unit is actually integrated into a Mac Mini, and not via a long external lead. My earlier reply was simply a shot across Eloise's bows, nothing more. To keep the thread from looking anything remotely like a commercial at this time, perhaps those interested in purchasing further down the track should directly PM Alex C. when the PSU is closer to becoming a commercial reality? I expect that Alex has his hands full at the moment with pre production stuff including component purchasing , documentation etc. anyway.

[Jud]

If talking about this design is making people uncomfortable I can stop talking about it. My original post in this thread was a response to a direct question about it and was not meant as "advertising".

 

As to testing, I DO make measurements. I haven't done any yet since the prototype was assembled at Alex's place and as I mentioned I didn't want to bring any test equipment with me. Since I'm not a big company with a large budget most of my test equipment is 20 year old stuff off of ebay, most of which is big and heavy, so hauling it around is not something I like to do.

 

Since he has the prototype I don't have anything to measure at this point. I haven't had time to build one up for myself. So it is going to be a little while before measurements happen.

 

As to how I do measurements, I built a very low noise wideband differential amplifier that is powered off of batteries, and sits in a metal box. (I like 1/4" thick aluminum, why is a very long story) I usually plug this into a spectrum analyzer rather than a scope, it's usually very difficult to see what is happening with just a scope. (BTW I have a 2467B, off ebay of course) The spectrum analyzer is a behemoth HP, very good, but huge and heavy (off ebay of course)

 

I'm sure I will be able to measure differences in the actual supply rails, but measuring analog on the output of a DAC is a different matter. In order to do this well you have come up with a guess as to what the interaction is, figure out a test waveform for that, figure out the best way to measure what you think is happening, set it up, do the measurements, and repeat 50 times. It CAN be done but takes a LONG time. It's not so simple as play some music and look at it on a scope. The effects from these types of power supply differences are small and you really need to "zoom in" on what they are with specific test signals and measurement techniques. I have done this a few times and it usually takes 3 months or so to figure out what is really happening and figure out how to measure it.

 

John S.

 

I'll join everyone else in saying how very curious I am to hear what is going on along the way. Your and Superdad's willingness to be open to this extent gives the rest of us a rare opportunity to understand more about the process and what goes into trying to realize such an idea. I hope you will not just restrict yourself to answering direct questions. About price, I think that can be left out. I hope Chris will be inclined to be tolerant.

 

I actually find the rest of your post even more interesting. This will hopefully give people some idea of the type of equipment, its sensitivity, the thought and planning, and the careful, thorough execution required in order to really measure the behavior of a piece of equipment.

[Superdad]

What's a Linux boot driver?

 

You know EFI and GRUB and all that stuff. Kind of a pain to get EFI-based Macs to boot other OSs. I tried all the usual utilities offered for that purpose. And Miska even sent me two versions of preconfigured/bootable ISOs with Debian and NAA. Those started to boot but got hung up after a few screens of hardware-recognition text. I never got a prompt. Miska may have given up on me as he suggested I get an ARM or Atom board. Maybe you and I should PM about this since I recall you had success with Debian/NAA on Mac.

 

Sorry for the OT folks...

[Jud]

You know EFI and GRUB and all that stuff. Kind of a pain to get EFI-based Macs to boot other OSs. I tried all the usual utilities offered for that purpose. And Miska even sent me two versions of preconfigured/bootable ISOs with Debian and NAA. Those started to boot but got hung up after a few screens of hardware-recognition text. I never got a prompt. Miska may have given up on me as he suggested I get an ARM or Atom board. Maybe you and I should PM about this since I recall you had success with Debian/NAA on Mac.

 

Sorry for the OT folks...

 

Yep.

 

BTW, that's a boot *loader*.

[Superdad]

I actually find the rest of your post even more interesting. This will hopefully give people some idea of the type of equipment, its sensitivity, the thought and planning, and the careful, thorough execution required in order to really measure the behavior of a piece of equipment.

 

BTW, John is particularly tickled with his recent acquisition of the 400MHz HP 2467B scope (to go with a fancy new set of 4 probes he got for Christmas). It is the only non-storage CRT scope ever made with a micro-channel plate CRT. It does extremely high speed writing, making one-shot pulses at nanosecond duration visible in normal room light. It is pretty unique and he says it will let him see much finer details than an ordinary scope. I understand that a comparable modern scopes would cost a lot more than what he paid.

[Superdad]

Yep.

 

BTW, that's a boot *loader*.

 

Yeah, well after all the hours I spent trying, it got my boot, and I need to get loadered!

[esldude]

My earlier reply was simply a shot across Eloise's bows, nothing more.

 

And nothing less.

 

So it is about what it seemed. Eloise asks a question, and you feel the need for a shot across her bow? So I take it you don't believe she should get to ask questions of interest if they upset your sensibilities. That is pretty pathetic.

[sandyk]

So it is about what it seemed. Eloise asks a question, and you feel the need for a shot across her bow? So I take it you don't believe she should get to ask questions of interest if they upset your sensibilities. That is pretty pathetic.

Eloise shouldn't have asked that question at this stage, of a well respected E.E. when it hadn't even reached the measurements stage, and was experimental as he checked it with and without the Kelvin sense. You are the one who is being pretty pathetic by polluting this thread.

[wgscott]

Eloise shouldn't have asked that question ...

 

Why shouldn't she be able to ask any question she wants to ask? Why should she require your permission and approval?

[sandyk]

Ho , hum. Here comes the cavalry.

The question was premature, and it was also made clear by Alex C. that both our replies were not appreciated at so early a stage in proceedings. It would be seen by many as yet another attempt to derail another thread where there is a subjective report , just as your present post will be seen by many.