LPSU choices, new contender.....

[sandyk]

30 minutes ago, Jud said:

 

Didn't see him claim it did.

 

I know you've got plenty of LPSU/DIY knowledge. It would be nice for me personally and I'm sure others to get some tips regarding what components one might expect to see in a quality LPSU, if we're thinking of purchasing one.

 

 Why not also ask members like John Swenson that question, as he has proven  expertise in that area ?

You could always start a thread and ask for contributions from both qualified and DIY members

[Ralf11]

What exactly ARE the difficulties in marrying the Digital area to the Analogue Audio area ??

 

Why does it need to be more of a " black art" where very few designers excel, and many use textbook implementations from the Semiconductor Applications Engineers?

 

 

 You only need to look at compilations of claims not based on proper comparisons to realise that your typical products from competent E.E. designers are often capable of marked  confirmation bias.

[jabbr]

1 hour ago, sandyk said:

I never said it was. The shot was aimed at me, and I simply made it clear that I (along with many other members) only need easily accessible high quality audio without the need to use physical media. The fact that in my case at least, it ALSO sounds way better than simply playing physical media is an added bonus !

 

Nothing personal but we’ve been hearing about the “need” to do all sorts of contorted tweaks to our computers to allow them to produce good audio when interfaced to a DAC. LPSUs on the memory banks, SSDs, special cables etc and a new tweak of the week for now years when all this can be solved using a good cheap isolating network.So no I don’t wrap my speaker cables around toroids etc because it’s just not necessary.

 

(good old fashioned heavy iso transformers to supply the audio area remains a good idea, and avoid really crappy SMPS) 

[Jud]

We all might wish for a little more cross-fertilization.  Would be interesting to see what marce thought of Pavel's implementation of Miska's DSC-1, for example, since that's hardware without proprietary IP in the way.

[Superdad]

1 hour ago, Jud said:

It would be nice for me personally and I'm sure others to get some tips regarding what components one might expect to see in a quality LPSU, if we're thinking of purchasing one.

 

Well ideally something better than $0.50 potted bridge rectifiers and ancient LM317 regs coupled with pass transistors and high wattage resistors.  [That's the design of the PSU that this threads was started with. Many of the Chinese units will be much quieter, those those suffer from not being able to produce their rated output currents.]

[sandyk]

32 minutes ago, jabbr said:

So no I don’t wrap my speaker cables around toroids etc because it’s just not necessary.

 FFS !

 I don't need to do that either, as I have long past moved on from using " Gain Clone"chips like the LM3886 which were all the rage many ago and still apparently favoured by at least one high profile member.

 I was just pointing out the RF level of some nearby Analogue TV transmitters back then. They even had to ban  a few people living nearby from aiming high gain Yagi antennas up at the transmitting towers, which were according to folk lore, then able to light up some Neon Tubes.

 BTW, I don't use LPSUs on memory banks, or special internal cables either. There is nothing technically wrong with regulating the existing internal +12V supply down to a clean and isolated +5V for SSD power as many members have verified. 

 John Swenson has even demonstrated that radiated RF/EMI from PWM control of fans causes audible degradation.

Yet you seem to believe that you can correct everything after the horse has bolted !

 

[jabbr]

2 minutes ago, sandyk said:

Yet you seem to believe that you can correct everything after the horse has bolted ! 

 

Yes, as long as the bits remain identical, any sort of noise on the digital line can be corrected. I don’t care to argue this point, particularly with anecdotes about what someone else heard 30 years ago.

 

The fact that music is now distributed over the Internet dies nothing to reduce the quality of such music (files) in any way. I’m not going to respond to any anecdotal experiences of what people may claim to have heard otherwise.

 

As you say you are not a qualified engineer.

[sandyk]

28 minutes ago, jabbr said:

 

Yes, as long as the bits remain identical, any sort of noise on the digital line can be corrected. I don’t care to argue this point, particularly with anecdotes about what someone else heard 30 years ago.

 

The fact that music is now distributed over the Internet dies nothing to reduce the quality of such music (files) in any way. I’m not going to respond to any anecdotal experiences of what people may claim to have heard otherwise.

 

As you say you are not a qualified engineer.

 

Reports are NOT Anecdotal when they are verified by an E.E. with vastly more experience than you have, by way of 6 separate, and correctly performed, so called "Gold Standard"  DBTs.

Even the S/W used to convert an original 24/192 .aiff file to a .wav file matters as Barry Diament  can verify after trialling 4 different versions of S/W for the May 1st release of his new 24/192 album '"Kay Sa." 

 Yes, all 4 versions of S/W sounded a little different.

 

No , I am not a qualified engineer, and glad of it , as unlike some, I don't need to trust everything I read in dusty old Text books.

[Tomslin]

6 hours ago, jabbr said:

 

Ah but the latency of that card is trashed by the Solarflare, haven’t you measured that!?

 

Oh shit! I didn't think about that 😉

[kumakuma]

1 hour ago, sandyk said:

 

Even the S/W used to convert an original 24/192 .aiff file to a .wav file matters as Barry Diament  can verify after trialling 4 different versions of S/W for the May 1st release of his new 24/192 album '"Kay Sa." 

 Yes, all 4 versions of S/W sounded a little different.

 

 

[Ralf11]

 

Extraordinary Claims Require Extraordinary Proof

[sandyk]

43 minutes ago, kumakuma said:

Nothing for a change !

 Further to my previous post, Barry has since re-encoded all of his previous 5 Hi Res albums using this new S/W for those who wish to have the albums in .wav format instead of the original .aiff format.

 

Incidentally, you can listen to excerpts from this new album on the SoundKeeper Formats Comparison page and compare 24/192 against RBCD

http://www.soundkeeperrecordings.com/format.htm
 

 

[marce]

9 hours ago, sandyk said:

 I was asking what actual equipment you are using in your own Audio system.

Still using something like a SBT ,without even a decent Linear PSU such as designed by John Swenson and used by numerous DIY Audio members ?

 

Fender Jazz bass, Fender 200W rumble amp.

[marce]

9 hours ago, sandyk said:

 I was asking what actual equipment you are using in your own Audio system.

Still using something like a SBT ,without even a decent Linear PSU such as designed by John Swenson and used by numerous DIY Audio members ?

 

More worrying is your way of trying to put someone down so as to discredit what they say... SAD.

[marce]

5 hours ago, sandyk said:

 Jud

 Marc keeps telling us how qualified he is in the area he works in, and I don't dispute his expertise in that area,

however it does not mean that this knowledge automatically transfers added benefits to normal consumer Audio products that were also designed by competent E.E.s to a level often limited by cost factors .

 Marrying the Digital area to the Analogue Audio area is more of a " black art" where very few designers excel, and many use textbook implementations from the Semiconductor Applications Engineers. 

 You only need to look at Rajiv's thread to realise that your typical products from competent E.E. designers are often capable of further marked improvements. 

 

Alex

F Off its getting monotonous, I have to defend my self to counter your endless putdowns, you are the grumpy old man I have mentioned previously. I have posted no put downs just some relevant info, I stopped posting because of endless putdowns, off topic comments aimed at a person, back to my bass, I can't be bothered. Goodbye.

[marce]

4 hours ago, Jud said:

We all might wish for a little more cross-fertilization.  Would be interesting to see what marce thought of Pavel's implementation of Miska's DSC-1, for example, since that's hardware without proprietary IP in the way.

Pavel from DIYA?

Got a link...

[Jud]

9 hours ago, sandyk said:

 

 Why not also ask members like John Swenson that question, as he has proven  expertise in that area ?

You could always start a thread and ask for contributions from both qualified and DIY members

 

Already have a number of replies by him. He was an early popularizer of El Cheapo, then moved on to ultracap supplies, shunting noise from SMPSs to ground, and of course the supplies that bear his initials.  I also used info he provided on this forum and others to avoid transformer ringing when I built a DAC.

[sandyk]

1 hour ago, Jud said:

 

Already have a number of replies by him. He was an early popularizer of El Cheapo, then moved on to ultracap supplies, shunting noise from SMPSs to ground, and of course the supplies that bear his initials.  I also used info he provided on this forum and others to avoid transformer ringing when I built a DAC.

Hi Jud

An  example of good design from John Swenson is attached, however it is not ultra low noise, but is eminently suitable for many applications. It uses a Snubber across the transformer secondary winding to reduce ringing, and Schottky diodes for rectification.

In my 15W class A Amplifier ,I like the Cree  C3D06060A  (Data Sheet attached) 

As Alex C previously said, an R-Core transformer is a good choice, doesn't buzz., and it also has a relatively low capacitance to mains earth.

 However , if used in a USB PSU do not connect the shield wire (if provided) to mains earth for best results.

A choke input filter is good if you have the space. Even smaller value suitably rated 470uH chokes can help with transformer buzzing problems in amplifiers, but they will reduce the supply voltage a little, so take this into account.

 If using low noise voltage regulators such as the LT3045, for most neutral sounding results fit a typical value electrolytic filter capacitor in parallel at it's input, like when it is being used after a bridge rectifier. 2,200uF to 4,700uF are usually suitable values.

In general, if you like a slightly warmer presentation such as with Vacuum tubes and some FET amplifiers, this electrolytic could be an Elna for Audio type for example.

If you like a slightly warmer presentation AND a detailed top end you could use something like an Elna for Audio 2,200uF in parallel with a Panasonic FC 2,200uF, which is a very good all round capacitor for use at that location. 

 I don't like very Low ESR filter capacitors as they may cause a " hardening" of the sound.

When using voltage regulators such as the LM317 and LT1085 do not use tantalum capacitors in the " Adjust" location, or very low ESR capacitors at their output.  

 

 Other people will of course have different ideas on this subject !

 

Regards

Alex

CREE C3D06060A.pdf

[jabbr]

9 hours ago, Tomslin said:

 

Oh shit! I didn't think about that 😉

 

Ha ha! I don’t have experience with the so-called Femto Ethernet cards, but I have extensive experience with Solarflare, Intel & Mellanox ... all of which have “femto” class clock oscillators. In fact the 100Gbe Ethernet NICs are said to have a jitter tolerance of “zero” and the latest NICs are 200 Gbe and 400 Gbe (Mellanox is current leader) 

 

Note that these all have their own on card power distribution networks and are all intended to be supplied by PCI-e / ATX type server supplies and ALL have extremely low jitter so my opinion is that for Ethernet NICs the resources and proven track record of Mellanox, for example, vastly outweighs any audiophile manuf.

 

I think this is what @marce is getting at: non-audio, space, defense etc etc designers are looking at the same low noise, low jitter, low latency parameters that are being discussed for digital audio. At least for Ethernet etc we are decidedly on the high speed digital side of this equation.

[jabbr]

6 hours ago, marce said:

F Off its getting monotonous, I have to defend my self to counter your endless putdowns, you are the grumpy old man I have mentioned previously. I have posted no put downs just some relevant info, I stopped posting because of endless putdowns, off topic comments aimed at a person, back to my bass, I can't be bothered. Goodbye.

 

Don't take it personally. His sole argument style is argumentum ab auctoritate. 

 

He is forced to put down an opposing authority, because otherwise his argument can’t stand on its frail leg.

 

But don’t let him run you off because there are those of us who “only know what they have read in dusty old textbooks” and so we are dependent on the experiences of people like yourself. I know that those waves don’t know whether they are carrying a sound or a code 😉

[marce]

10 hours ago, Jud said:

We all might wish for a little more cross-fertilization.  Would be interesting to see what marce thought of Pavel's implementation of Miska's DSC-1, for example, since that's hardware without proprietary IP in the way.

Found it, its the Ladder DAC. Will have a look, my initial comment would be that I would use resistor packs with matched resistors, easier assembly, thermally connected resistors etc.

[opus101]

DSC-1 isn't a ladder DAC its an analog FIR filter. Resistor values don't need to be particularly closely matched as the output is a summation of time-delayed single bit DACs.

[jabbr]

2 minutes ago, marce said:

Found it, its the Ladder DAC. Will have a look, my initial comment would be that I would use resistor packs with matched resistors, easier assembly, thermally connected resistors etc.

Not a ladder DAC, a pure DSD input DAC described by @Miska DSC-1 and modified as 1) balanced and 2) output transformers as the analog filter.

[Jud]

In fact Miska has talked about the resistor matching problem with ladder DACs, one of the reasons his design isn't.

[Monge]

12 hours ago, Superdad said:

 

Well ideally something better than $0.50 potted bridge rectifiers and ancient LM317 regs coupled with pass transistors and high wattage resistors.  [That's the design of the PSU that this threads was started with. Many of the Chinese units will be much quieter, those those suffer from not being able to produce their rated output currents.]

Hi Superdad,

I wish Uptone could develope a PSU for Intel NUC endpoints like the JS-2 at 19V?

I know many People here are using LPS 1.2 and JS-2 with NUC’s.

I guess Uptone Audio is buzy with Etherregen..

Cheers Monge