|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
173.200.6.182
In Reply to: RE: cMP - the open source high-end Memory Player posted by cics on December 30, 2007 at 05:42:01
those of you who have played with the power supplies I would appreciate your consideration and thoughts.
The more I think about all of this linear power supply stuff the more I wonder if we really need to go to that much trouble. What is it about the SMPS that makes it obnoxious?
All power supplies radiate (other than batteries) so that, alone cannot be ultimate "problem". One could see the high frequency switching garbage as a multiplex signal riding along the DC as a definite problem.
Reading the review of a NAGRA amplifier in 6MOONS today which uses an SMPS supply, along with the realization that LINN seems to successfully use them, also, makes me think we might be going to much trouble and not necessarily for good reason.
So my simple "solution" experiment is to see if a large high current choke on each voltage line. with the thought that the lesser the current needs the more HENRYS we can use (due to availability and size considerations)one would think these could filter away the offending garbage, or at least much of it.
I will try a 100mH 5 amp choke on the 5 volts line tomorrow just to see if the thing will still turn on. It will be interesting to see if the added series resistance (o.67 ohms) will cause a problem. At this point I just want to see if it will work.
Maybe I have missed this idea being discussed and if so my apologies for bringing up a settled discussion.
We'll see ...
Follow Ups:
So while we wait, I offer up some reading on SMPS's...
First, this thread about the sonic effects of SMPSs that burst into the PC Audio Asylum last night:
http://www.audioasylum.com/cgi/t.mpl?f=pcaudio&m=77788
Some snarking & sniping, but some good info too!
Second, this thread questioning replacing the SMPS in an Oppo CDP:
http://www.audioasylum.com/cgi/t.mpl?f=hirez&m=267546
I found the comments about Ed Meitner's use of SMPSs and some of what he does to make them sound good especially interesting.
Third this post by Charles Hansen of Ayre talking about what they do to make an Oppo CDP into an Ayre CDP:
http://www.avsforum.com/avs-vb/showthread.php?p=17402589#post17402589
If the link doesn't take you there, go see post #11.
Honestly, I suspect that an excellently-designed and executed SMPS may beat a standard linear supply. But there are few of them around... and that someone as smart and experienced in electronics as Charles Hansen still choses to use linear supplies tells me it REALLY takes something to do that excellently-designed and executed SMPS... or maybe, I'm wrong and a good linear is ALWAYS better!
Greg in Mississippi
Everything matters!
And then I saw many of those posts, too.
I think what made me reconsider is the problem with the switchers infecting the AC lines.
So, I am back to resigning myself to the noble task of AC powered linear supplies.
I have the stuff on had to make most of them and I intend to use chokes before (or maybe after?) the regulators to control some of the digital noises. There again I am not sure if the choke's filtering may result in ringing if used after. I like the idea of having some element that isolates the circuit in both directions. Wish I had an oscilloscope.
I am thinking of using a large choke in the ground line for this purpose.
Have you considered using a current regulator before the voltage regulator instead of cascaded voltage regulators? Not sure if you were using one or two. I am assuming one can use the LT regulators in this way. They say you can on the data sheet but unlike the LM317 data sheet they show no implementation.
Broke down and ordered one of the BURSON 5 volts shunt regulators yesterday for JULI@. Maybe these circuits have to be regulated electronically?
I will use the LT3080 for the the other rails since it has vastly superior ripple rejection to much higher frequencies. To use on the 5 volts rail would require paralleling and I am not sure I should attempt that.
Now to DO something.
Rick McInnis
Regarding choke-filtered power supplies...
In reading some of the posts by John Swensen, especially his posts on low-DCR in the Tube DIY Asylum, I believe there is a science to designing good choke-filtered supplies and John knows that science... and better yet, listens to what he does. I strongly recommend his choke-filtered 9v & 12v supply as shown here:
http://www.audioasylum.com/cgi/t.mpl?f=pcaudio&m=42853
He takes into account not only filtering noise in both directions, but also sizing & selecting components to minimize noise creation in the supply AND having the supply exhibit a critically-damped response, which he suggests is important to good sound.
My only caution with that supply is that in the few situations I've used the Panasonic FM & FC series caps, I've not liked their 'inherent' sound. I'd check with him for alternatives... Elna Silmec or Cerafine, Nichicon FG or FX, etc.
He can also likely provide alternative components for 5v & 3.3v versions.
My gut feel is that just adding chokes without some amount of design may be a crap-shoot on whether it improves the situation or not... and John has done the hard work here, so this is a good and safe bet.
This is the configuration I'm planning to go to when I put my cMP into a home-built case (with more room for PS stuffs).
Then on regulators, first, I think you're making the right move to add regulators for the sound-card power. While you can often get away with non-regulated supplies for analog tube and solid state circuits (and done right, some say it's better-sounding), logic circuits have a pretty small range of input voltages for proper operation. Hopefully you haven't damaged your Juli@ (a possibility).
On regulator types, my next thing to try is the LT3080 and see if it makes an audible improvement over the LT1083. I'll be trying it first in my 'dirty' supplies and on the 3.3v & ATX 12v regulator positions... These are the easiest for me to change out and easiest to put back if there isn't an improvement... but experience has shown that it will be audible (whether for the good or the bad) in those positions. I hope to be doing that in the next couple of weeks.
If it does, then I'll setup the paralleled regs for the 5v where I'll use 4 regulators. If that works well (and if you want to go that direction), if you are cautious about following in my footsteps, I'd be happy to wire one up for you.
But ultimately I plan to go to current-sourced shunt regulators. I have heard from one of my 'tweak mentors' that cascading series regulators does not sound good (in his humble opinion), so except in one position, I haven't done that. But CCS-sourced shunt regulators, such as the Burson, the Tent, the OPA, the Borberly, and the DIYAudio Salas (and other) ones are ok. This link I posted in an earlier conversation:
http://www.diyaudio.com/forums/power-supplies/168631-5v-high-current-low-noise-regulator.html
Is for a varient on the Salas design for high currents and what I intend to try after the LT3080 for the ATX-20 5v & P4 12v regulators. I expect the only thing that will be better are Paul Hynes' regs, which may be next after that.
I hope this all helps... and let us know what you do.
Later!
Greg in Mississippi
Everything matters!
I have done endless modeling with the DUNCAN PS simulator and, of course, one then has to hear what it sounds like.
I am familiar with Swenson's circuit. It follows very closely what is considered best practice.
I use Hyne's large 12 volts regulator for my DAC. Maybe he would change his mind now that the technology has allowed a 2 ampere 5 volts line but a couple of years ago when I asked him to consider the possibility he said it would require something VERY large. I do wonder if such a thing is necessary. It could be better but, I suspect, only a little bit better, not worth the trouble. In other words, there are other areas that would yield more improvement for far less money, or I hope so.
If the other lines are not as critical, then it seems foolish to go with shunt reg's. The soundcard is a different story to my way of thinking.
I am sure Hyne's design is better, but, I bet, not that much better. The Burson's strike me as being very good engineers.
With the onboard 3.3 volts regulator on JULI@ which is an LDO type, needing less than 1 volt for the current requirements of JULI@ so I cannot imagine any damage being done. I really don't see what you are getting at - maybe with voltage too high for the regulator but when it has been in the circuit it was around 4.5 volts under load.
There is no question that too many of these narrow bandwidth regulators in cascade is the equivalent of a high order low pass filter in a loudspeaker crossover, bad. Which is why I am settling on an LC filter after the rectifier, a current regulator and then the voltage regulator for the 5 volts line. For the others I will just use a single voltage regulator, LT3080's). A capacitor after the rectifier will be used to tune the output voltage, just as Swenson, since I think a "choke input" is not needed here since a current regulator will be used.
My thinking is the choke will still be "working" after the LT1083 is beyond its range. By the Nature of the currents involved the choke will be 100mH which is quite large but with low resistance.
Easy enough to take it out if it seems to be causing a problem, which I think is highly unlikely. This will not be my first LC filtered power supply.
The LT3080 may have enough bandwidth to make this unnecessary. Who knows? I am nervous about paralleling, it always seems to have its own set of drawbacks. Of course, none of us gets to win in audio it is always the compromise one finds least obnoxious.
Oy,ve.
Rick McInnis
I got confused by the menu at PARTS CONNEXION where they listed it as such. With thought it occurred to me there is no way shunt reg could be substituted for an LM317 and sure enough it is a series regulator.
In the meantime I came across this: http://www.belleson.com/index.php
This seems like the "latest" in series regulators and though I would still prefer a shunt regulator I am going to give this a try for JULI@.
The fellow making these is a very nice guy and his shop is about one hundred miles up the road. I have to give it a try!
Didn't know you were such an electronics expert.
But I am an obsessed cheap hobbyist who has no choice but to learn to do things for himself.
When I see something I cannot even DREAM of doing myself I readily defer to the likes of Nelson Pass, John Broskie, Paul Hynes and Dave Davenport, who, along with cics, are my audio gurus. And those magic fingered moving coil cartridge artisans ...
What excited me most of cics's work is that he was making it possible for folks without unlimited budgets to have something good for CD playback.
I have been playing with this since I was fifteen when I got a Dynaco PAT 5 kit. (Why did I not get a PAS 3?) You would think I should know more than I do when time spent is considered. (That was forty years ago. YIKES!)
Hoping I have a working C squared set-up this evening. Can't wait to hear what has happened with the software in the last couple of years and very excited about hearing what the balance control can do.
Bye,
Rick McInnis
You are only in your 50's? You are just a young-in! I remember when my parents let me hook up their old Philco tv speakers to my mono record player. It wasn't stero but it was 2 channel mono---just awesome!
Good luck on your new mobo/cpu set up.
It is funny how in your fifties you can still feel young yet you know you are getting old, for better and for worse ...
Not surprising you were a tinkerer from the beginning. That same desire to make it better obviously started at a young age.
I hope I can actually get music out of it! Just have to solder five wires from the DAC board to the line stage. Just could not make myself finish up last night.
Whether anything resembling music results ... At least, I can always play records while sorting this out.
After two years (or so) LP immersion I hope I can listen to it. At worst, it will be something to play while I work on house projects. You can tell I have a little trepidation.
Are you considering the I3 upgrade?
Bye,
Rick McInnis
Yes I am considering the upgrade probably in Q4 (still talk like I'm working). But I have been busy with other non-audio matters for 2 months. When I get started I'm just wondering if I should buy all new (iow keep the old system intact) or mod my old sys. What approach did you take?
nothing wrong with my HDD, or keyboard, or that kind of thing.
Of course, you have no choice but to replace your memory. so that's it - new CPU, memory "card", and MB an d you're off!
Still trying to get some kind of real reason to go with the SDD for WINDOWS and programs, but nothing definitive and I have no interest in experimenting there since I think it is kind of silly though I would love to be proved (or even persuaded) wrong!
I would think there is no need to rush since there are still a few questions about the new system. And the prices may come down.
I was anxious for a new project and that is why I got the new stuff.
I am concerned by Julien's problems and wonder if there are a bad batch of MB's out there. Might as well hear what happens with his project.
What is most important is that you keep your ears in top shape! You possess a very perceptive pair of those.
Bye,
Rick McInnis
Of course, I'll have an opinion on this... and it may not be the one you expect.
I've had several experiences with SMPS's in audio eq over the last 6 years or so, starting with a Panasonic S47 DVD player modified by one of the serious modifying houses. He did quite a bit to the SMPS in the unit... and having an un-modified unit to compare to, I could hear a significant improvement over the stock player and it competed very well against my fully-outfitted Njoe Tjoeb player with upsampler, beating it in some ways, equaling it in others, and doing so for 1/3 or so of the price (and the Tjoeb was fully linear-supplied).
Couple of years later, I read about the Sony Playstation PS-1 as CDP mini-craze and picked up a couple to play with. They were promising enough to dig into and mod... and because I was deep into my then-business and model airplanes, I didn't try to do custom linear supplies, but instead modified their standard SMPS's (see attached pix). My mods, inspired by the ones on the S-47 and information gleaned from other modifiers, included replacing the rectifiers with supposedly good-sounding soft-recovery diodes, the main cap with a Jensen 4-pole (first time I used one), and most of the other electrolytics with Rubycon's (I hadn't tried Black Gates then... they seemed really expensive to me back then). Very nice improvement, it preserved the liquidity and analog-ness of that game console when used as a CD player while improving the resolution and definition. While it didn't have the level of resolution of the S47, it was very fun and listenable and also a good platform for modifications.
Modified Sony Playstation PS-1 SMPS:
One year later, I get a real job again and celebrate by purchasing a pair of serious amplifiers, built-up by the same modifier who did the S-47. These were based on the B&O 1000ASP IcePower modules (over 500 WPC into 8 ohms, over 1000 WPC into 4 ohms), which of course had SERIOUS SMPS's. They worked very well in my system, surplanting an extremely-modified Counterpoint SA-12 and decisively beating it in every parameter... and worked very well with the S47, which became my main CDP again.
And another year or so later, I picked up a used Oppo 970 modified by the same guy who did the S47... another SMPS-powered piece of gear and it was a good step up from the S47.
So far, SMPS audio gear is holding up well.
Next step in my audio evolution was to build a Shigaclone and feed it into a Peter Daniel NOS DAC. That beat the Oppo 970 (altho it didn't play DVD movies like the Oppo and Panny) AND was all linear supplies. I don't think that was the key reason that setup beat the Oppo, but it contributed... and I got to see how people were making linear supplies nowadays (the last ones I built were back in the early nineties before people started using anything but garden-variety diodes).
Then enter the cMP. I put it together and liked it... sound quality in the same realm as the Shigaclone/PD NOS DAC combo with a bit of convenience thrown in (Remember, that PS-1 had NO screen, just a set of controls... and the Shigaclone was a step up with a screen, but just as rudimentary controls).
As I moved forward in implementing the standard tweaks to my cMP, I looked long and hard at the insides of the Antec SMPS when I was disconnecting the fan and un-needed wires... Hmmm, maybe I could do some mods in here. But then I saw the sub-threads about adding caps at the output of the Antec... and then I saw Peter Daniel's PicoPSU implementation... and started trying these tweaks to my cMP... and the rest is history (for me).
What did I learn from all of this (and also reading about SMPS-powered gear that reviewed and sounded good)?
1. SMPS's can be made to sound good (Heck, I knew that back in the late '80s when I got to hear several times a Berning TF-10 tube preamp with SMPS owned by a friend of mine!).
2. Most SMPS's are not designed and/or built for good sound.
3. You can improved the sound quality produced by those SMPS's with some mods.
4. But it's simpler in most cases just to build good linear supplies than modify a SMPS... and it likely will produce a better result.
IMHO, computer power supplies are EXTREME examples of SMPS's that are not designed and/or built for good sound. The ones in the B&O 1000ASPs are designed for and built for good sound. The ones in the Linn gear are too. The ones in the Nagra amps are too and likely better than either of these (although they seem to be more of a hybrid power supply with that large transformer before the power supply board than a classical SMPS setup). The ones in moderately-good-sounding-when-stock DVD players like the S47 and Oppo 970 (and some of the Toshiba models, one of which is playing a movie for my wife right now) are... somewhat. Ditto with the one in the PS-1... and the SMPS's in these game consoles and DVD players responded pretty well to mods designed to improve their sound quality. This was helped by their simplicity (at least compared to a computer SMPS). On the other hand though, the guy who had modified the S47 and Oppo is now offering a linear supply add-on to his latest DVD mod offering.
Given all of this, I believe that modifying the standard computer supplies, even the 'good' ones like the Antecs, are like putting lipstick on a pit bull. They are so far from what is needed to produce good sound that while you can make them better, you'll never make them good. And my experiences with minor changes to my cMP's power supplies, even the motherboard and 'dirty' (USB, screen, HDD) supplies, making a noticable-to-large difference, has told me that power supplies are critical to getting the best sound out of a cMP-type of source, even when the sound card (and DAC and output stages) has it's own power supply.
But of course, I'm very curious to hear what you report. I hope it'll make a significant improvement in your setup. I hope at least that it works ok... I'm not sure it will. If you removed your fan from your computer supply, you may want to add it back at first... I can see some ways where the chokes may cause the supply to work a lot harder and the fan may be what saves your supply until you can power it down.
Looking forward to your report!
Later!
Greg in Mississippi
P.S. Thanks for keeping this all so interesting!
P.P.S. I forgot to mention I used a Panasonic receiver with SMPS for awhile in the last couple of years as I moved into our house down here and it made a better account of itself than I expected.
Everything matters!
Greg,
I would tend to agree than SMPS is not for me when it comes to analog but one wonders if conversely, it is not perfectly good for digital, especially within the computer?
I guess I still have a hard time thinking about a computer as an audio device. I do consider the sound card an audio device but from there back it is a computer, though I do see the goodness in having as little extraneous anything circulating the MB.
I would love for this semi-simple "idea" to work but I can assure you I am not determined to like it. I have the stuff for linear supplies laying about and will have to make the comparison eventually for it to mean anything at all.
I must admit I had no intention of doing anything to the supply itself, just adding an LC filter to each of the outputs. I have got to say the idea of being able to hear a rectifier within an SMPS is almost frightening!
Of course, I will keep you and any others who might be interested informed since this is my favorite forum and I can lapse into chattiness!
My main idea is that if this is an improvement that could be implemented easily by those not interested in the complexities of making linear supplies though, no question, with three terminal regulators, it does become simple(r).
At this point it will be interesting to see if the supply will accept this extra inductance at all and allow the computer to turn on!
Thanks for your interest, help, and ...
I knew there had to be a new cPlay on the way since it has been so long since cics had SPOKEN and sure enough ...
Bye,
Rick McInnis
If I've learned anything over the past 6 years of getting back into audio tweaking, it's that my signature, 'Everything Matters', is more true than I ever imagined.
That I heard a clear difference between Panasonic FMs, a combination of Panasonic FM's & Elna Silmec, all Elna Silmec, and finally a combo of Jensen 4-poles and Black Gates in the linear power supplies that feed my HDD, Zalman screen, and USB port was a revelation to me. Guess which sounded better?
More recently, hearing the sonic signature of those Silmecs again (which isn't a bad one, just one I don't personally like) when they were used in my linear computer supplies, even when finally the only one in the supply was the one across the output of the -12v supply that I turned off after the system was started, had me concerned that either 'EVERYTHING REALLY MATTERS!' or that I was going batty.
Can I explain it? No, not really, I'm not that technical. But I'm very comfortable in reporting it and stating what I've learned, that if a SMPS isn't designed or modified to be good sounding, it is not good enough in my audio system, even for the digital circuits, even for the things that I can't imagine will make a difference like the Zalman's screen or the HDD!
Waiting expectently for your report!
Greg in Mississippi
Everything matters!
Hi Rick I use dual switching supplies (ala cics recommendations a bit back) and I am happy with them sonically. I also use LiFePo4 batteries for my 3.3 juli@ supply (never a problem) and I do the same for my 3.3 volts digital supply on my Buf32s and its not a problem. I went to these batteries per the European guys on AA (bertel and others who have since gone to non cmp solutions but still use LiFePo4's). I am very happy with the LifEpO4 WHICH ARE LESS HAZARDOUS THAN OTHER LITHIUM ION BATTERIES. So I think at least the way I use them are very safe. I use 3 in parallel and they last 5-7 days for my juli@ and longer for the Buffalo before I need to charge.
that the batteries are safe!
Takes away all of the excitement!!!!
I would like to use the batteries and to only use direct 3.3 volts but I am using Dave Davenports HDMI implementation of I2S (similar to the PS AUDIO approach) and need 5 volts to run the little send card.
I like using batteries for sensitive circuits though I would hope you are using a battery with as large an amp/hour rating as you can get. Little batteries have high series resistance which is not so good. Big batteries have vanishingly low series resistance and they can sound excellent.
Theo, hope all is well for you.
Bye,
Rick McInnis
I use 3.2 V 2.00 amp hour batteries (largest ah rating I could find) and with the small draw on juli@ and especially Buf32s it lasts a long time.
All is well.
Post a Followup:
FAQ |
Post a Message! |
Forgot Password? |
|
||||||||||||||
|
This post is made possible by the generous support of people like you and our sponsors: