 
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
|
|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
87.18.87.46
In Reply to: RE: cMP - the open source high-end Memory Player posted by cics on December 30, 2007 at 05:42:01
Hi
after a complex semilinear approach now I'm planning to assemble a full linear PSU system.
My project was thinked around the AMB lab Sigma 11 power supply (http://www.amb.org/audio/sigma11/) in a Miahylov fashion.
Studying about problems, I need some help:
1)The PSU will be external so I need a long run of cable. How can I try to reject various interferences: Shielded cable? Twisted?How twisted?
2) In the Miahylov project ground star seems to me formed by each negative cable coming from each different power supply. From the star, or better to the star, arrive all the ground cables departing from P 24. In the Miahylov PSU I can see one large cable that near to the connector go to join with all cables coming from the P 24 connector. Is this another "floating ground star"? Or It could be better to arrive to the ground star inside the external PSU with each single cable?
3)Where (if i must) I'll connect the PWR_Ok grey cable?
4)where (if I must)I'll connect the PS_ON# geen cable?
5)+5VSB PSU must be always turn on or could start with the others PSU together?
6) always In the Miahylov project, I can see a relais near a little printed board. How is the sequence of turn on each PSU? How is connected the relais?
It could be possible that these arguments was explained yet but my research was ineffective
THX a lot to everyone could help me
THX a lot to bibo01 for Italian cMP2 site and forum: great job, Bravo!
Daniele
PS: my actual system ( GB H55MUSB3+i3 540) is feeded by:
A) Seasonic x650 ATX PSU: P 4( via a run of 6000 mF of Silmic II)+ SSD+CD/DVD/BR+ external HDD
B) Picopsu 150 XT: P24 only
C) Peaktech 6080 +12V: picopsu + sound board Asus Xonar ( 2,45A in music playback. If you connect P4 too, consumption rise to 2,7-2,8A)
D) Peaktech 6080 +5V: Sound board Asus Xonar + USB (1,5A in music playback)
Four switch to turn on and the Peaktech +12 become very hot after an hour, but sound is gorgeous
Follow Ups:
Hi guys,
Today, just hours ago i made huge step forward in my cMP set up.
Ones again -power supply ! I use low profile computer case and have not enough
space for everething i want to put in , so in time my MoBo died, i change design. Cut off some bottom part of a case to provide access to back side of MoBo.
First, i follow JackWrong way , but just install small electrolit cap on open part of MoBo and move some of my caps on back of moBo . It was posible use short wire 30-35 mm longfor big caps.
Difficalt to say about improvment , becouse I start to use Buffalo DAC and tube Buffer, belived it was also on part of cMP.
Access to back of Mobo ,give me ability to mesure by my basic PC Oscilloscope what
i got . I find difference 2-3 fold in mvolt on deferent end of 35mm wire.
So , using strip of PCB board as a wire (should be square size or so, but not narrow)
I connect my caps to Mobo.
Result- excellent. Sound more clean, bass more controlled,mach better resolution on low level. I still not mesured what I got by Oscilloscope, planned make same approch
for CPU +12volt and added small NOS cap all around, will see what happen.
Gene.
Hi Gene!
It's interesting but photo is small and dark :(
http://cmp2-mihaylov.narod.ru/
will be sent tommorow better, this ones just scaled.
Hi guys,
I starting use PIO caps for fitering with great result. Two 2mf PIO caps reduce
noise on 10 db from 10 khz upward , on 5 volt MoBo line, but still some spike.
Sound has changed , more space , cleaner,more impact ,distant instruments and low level signal sound more smooth ,refined.
More experiment on a way , will report.
Gene.
Hi guys,
Continue install PIO caps for filtration on P24 connected by PCB . Again , success .
Noise level drop on 7-9 db according to oscilloscope reading ( +12 MoBo).For +3v MoBo and +5v MoBo added small caps 0.001mf -improvment on 3-4 db,( it is giong to measurement limit of my oscilloscope -90- 95db.).
Sound cleaner, more space. More caps on a way.
I think this is a way forward. Salas shunt regulator splited on two units, last one should
be installed closely to board. Buffalo DAC use also smallest shunt reg., whith connectors
pin just a 1/2 inch long.
Picture show last addition.
Gene.
BTW. How post bigger picture?
"BTW. How post bigger picture?" - In no way unfortunately :( ("All images over 640x480 resolution will be scaled accordingly to reduce bandwidth and storage space."). It is possible to upload here http://imageshack.us/ (for example)
http://cmp2-mihaylov.narod.ru/
Hi guys,
I installed PIO caps on +12 volt CPU line, ( two 6mf caps) but no any good result this time , even added another 47 mf non-polarised cap , not make improvement. Just for experiment I connected my 45 Amp lead battery by short, 150 mm 8 AWG wire to 60 mm long strip of PCB wich is soldered to MoBo- nothing changed. Only option I see now -replace switching regulators. First I am going to try with memory, as JackWrong
did, but will use LT1083 regulator, feeding from original 5 volt spot.
Memory 1.85 volt spot, I find on MoBo ( maybe it is not memory spot ?) very noisy -same as CPU , you can see on picture .
For 3.3 volt MoBo I added two 6mf PIO caps , this give me litlle bit improvement .
Level of a noise same by analyzer , but reading in mv better ( 1.6 mv vs. 3.0mv).
Adding 47 mf non-polarised cap reduced noise just on -2db. Sound- wise I cant see any difference.
Gene.
Jack is not Wrong but Wong ;)
http://cmp2-mihaylov.narod.ru/
Thanks Mihaylov, sorry Jack !
Hi!
1. "The PSU will be external" - It's the correct approach.
"How can I try to reject various interferences" - I think that no interferences will exist and will enough apply a qualitative power (acoustic) cable, for example, the such http://www.soundstringcable.com/audio_cables_cords/In-Wall%20Cables/tricormaxial_UL-bulk_speaker-cable.html together with Neutrik connectors http://www.neutrik.com/fl/en/audio/204_42012003/XX_Series_productlist.aspx and http://www.neutrik.com/fl/en/audio/204_293581/DL_Series_productlist.aspx.
2. Download the document added with two schemes once again. There can be you receive answers to some questions.
3. PSU -12V it not to be necessary for your system board most likely.
http://cmp2-mihaylov.narod.ru/
Hi Mihaylov,
Would this document replace the current one available for download on cMP² site?
Is there a way to manufacture these linear supplies in limited quantities? How much seed capital would be needed? Paul Heynes seems to offer something but at very high cost.
cics,
Please let us know when Mihalov's new document is on your cMP website... I'm not interested in signing up to Rapidshare and giving them my email address to get this.
THANKS!
Greg in Mississippi
Everything matters!
I normally have the same problem with Rapidshare, but if you use a program like JDownloader it's fine - no signing up.
...
Everything matters!
Hi cics!
"Would this document replace the current one available for download on cMP² site?" - Yes, make it please.
I have not possibility to make this psu even in the limited quantities unfortunately. I'm not the manufacturer, audiophilia is my hobby :).
Paul Hynes's PSU (http://www.hifizine.com/2010/06/paul-hynes-design/) is very qualitative device, but it is necessary to consider that this device was made by request of my compatriot Murat for power supply of Adnaco (http://www.adnaco.com/products/s1/) and consequently has only +3.3V, +5V, +12V and -12V (pwr_ok and 5VSB are absent).
http://cmp2-mihaylov.narod.ru/
.
OK.
http://cmp2-mihaylov.narod.ru/
IMHO, a cleaner solution for any ATX case would be:
- an external box with the 3 PSUs
- 3 ext. cables
- a metal plate with connectors, for PC back in place of ATX PS
- 3 internal cables for P24, P4 and Peripherals.
This is how I am building it!
Good Bibo!
"3 internal cables for P24, P4 and Peripherals." - I would make so: first cable (6 wires) for P24 (+3.3, +5, +12, 5VSB, ground - separate low-power psu for 5VSB and pwr_ok, delay board for pwr_ok is in the PC case, for example near the P24 socket), second cable (same, 6 wires) for P4 (+12, ground) and Peripherals (+3.3, +12, ground) and third cable for Juli/Lynx AES16/lynx Two (22) w/o analog (+5, ground).
Connectors are Neutrik.
http://cmp2-mihaylov.narod.ru/
Edits: 12/12/10 12/12/10
Mihaylov,
I am not sure I can agree with your suggestion.
I do not have a Juli@ and my Lynx card would require triple voltage (+5V, +12V, -12V), so I am not going to power for it separetly but through the board.
My idea was to have:
Cable no.3 - (+5V, +12V, ground) for HD, USB, int. LCD + extra device
Cable no.2 - (+12V, ground) for CPU/P4
Cable no.1 - (+3.3V, +5V, +12V, -12V, 5VSB, ground) for mainboard/P24
You suggest to have "separate low-power psu for 5VSB and pwr_ok, delay board for pwr_ok is in the PC case, for example near the P24 socket", any particular reason?
I was thinking to keep everything together, inside the power box.
You also suggest Neutrik connectors. I think I am going to use different power connectors with screw rings which are often used for double frame high-end audio geare.
All transformers are custom made.
OK. But I think it would be acceptable to unite cables no.3 and no.2.
Delay circuit for pwr_ok works from system board's PS_ON input signal and so it's useful to place it near P24 socket.
http://cmp2-mihaylov.narod.ru/
Mihaylov,
You suggested "separate low-power psu for 5VSB and pwr_ok".
Do you mean, "separate from your main +5V linear supply?
Did you mention this in your PDF design plan?
Could you please tell why and what kind of low-power psu?
I would need your info to implement it.
Thanks
-Do you mean, "separate from your main +5V linear supply?- Yes.
-Did you mention this in your PDF design plan?- No. It's a new option offered Greg and improving a sound.
I use such linear psu for 5VSB and pwr_ok in my alternative cMP2's linear PSU (see Photos from my site).
http://cmp2-mihaylov.narod.ru/
Using your delay circuit does Pin 16 PS_ON (Green) get connected on P24?
Thanks
Yes, certainly. Pin 1 DD1 is connected to pin 16 socket P24 of the system board. The system board is started by pressing the button on Zalman's front panel.
http://cmp2-mihaylov.narod.ru/
Thanks.
I wanted to be sure becaus in your PDF design plan, on the last page, Pin 16 is set to No Connection.
"...in your PDF design plan, on the last page, Pin 16 is set to No Connection." - It's correct because delay circuit isn't use in initial variant of linear PSU for GA-G31M-ES2L system board.
http://cmp2-mihaylov.narod.ru/
Thanks a lot.
I will upgrade "my" plan :)
When you say ""separate low-power psu for 5VSB and pwr_ok", was this in your PDF design plan?
What kind of low-power psu?
Thanks
Hi BiBo01,
I’m considering such solution too.
If you have finished them, could you please post some pics here on the forum or on Picasa for some inspiration too others who also consider a separate enclosure for there PSU’s?
Mark
fully AOB optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
Of course I will as soon as it is finished.
No need to ask :)
This is a community, isn't it?!
Your suggested design is what I had in mind. Its also very important to have 2 AC inlets (one for P24 and the other for P4 & peripherals).
How do we get this manufactured?
cics,
I was actually planning to have three power cords, but to make a simpler design I am going to have only two.
dear forum,
getting a linear power supply into manufacture is not the problem...
there are numerous specialist companies here in germany doing all sorts of customized PSUs.
-even in small numbers.
the big question is how many confirmed orders we can get together.
I would be happy to forward a request on a one-box solution, but we seriously need to fix a specific design on such a project.
if the experts can formulate and write the "receipt", and we can do an estimate on the numbers, I shall be very happy to request a price from different manufacturers.
kind regards
cMP2 Computer (XP minlogon) Intel E7400 cPlay039>Allocator>Lynx Two B. /192kHz throughout. 2x AcousticReality Ref. 202 & 2x AcousticReality Ref. 601´s ICEpower. Magnepan MG3.3R beechwood frames & custom stands. Miller chokes on Ribbons.
Hi Playmate,
That would be great idea.
But do we know what too order? What is needed? I think it is a pitfall when telling a PSU manufacturer how too put together a PSU in terms of: how to build it and what too put in.
May be is it better too only identify what parameters accurately describe the quality of the power that is needed for an cMP setup.
I came too this approach after reading this article from mr Rob Reeder (senior converter applications engineer at Analog Devices). He argues: “ Engineers commonly feel that a switching power supply can degrade the performance of an A/D converter, leading them to choose a low dropout (LDO) linear regulator over a switcher, but this percep¬tion isn’t entirely true.”
See: http://www.analog.com/static/imported-files/rarely_asked_questions/RAQ_tamingADC.pdf
So, do we absolutely need an all linear PSU? May be a mixed-mode PSU with Linear Final Regulation is also sufficient for a cMP setup. Who knows?
What specs should a PSU for a cMP setup have?
fully AOB optimized cMP2 PC -> ESI Juli@ -> Van den Hul Optocoupler MkII-> Lavry Black DA10 -> XLR Mogami Gold -> Klein & Hummel O300
Hi all
finally I've found the thread with the responses to my questions
http://www.audioasylum.com/cgi/t.mpl?f=pcaudio&m=65451.
Research an argument in this fantastic forum sometimes is not so fantastic, very slow and demanding. Here is present near everything, but to find it...... it's another story!
First, I think AMB's Sigma regulated PSU looks like a very good choice as the main building block of a home-built cMP power supply. I think that it will be an upgrade over the garden variety LM317-type regulators (I use the better & higher current LT1085 in some places and the better LT3080 in others) and may approach a good shunt supply like the Salas high-current shunts and possibly even Paul Hynes regs.
That brings up a question... You said you'd be doing "an external box with the 3 PSUs... 3 ext. cables... 3 internal cables for P24, P4 and Peripherals."
There are at a minimum 6 different voltages for a cMP using the Zalman case:
- ATX-20/24 (P24) 3.3v/5v (both power and control voltages)/12v
- P4 12v
- peripherals HDD & USB 5v / screen 12v
Are you planning to use a separate transformer for each voltage, or combine some using transformers with multiple secondaries? Also are you planning to use a Sigma for each voltage or use lesser regulators for the less-critical voltages?
My setup uses separate power supplies with separate transformers for each voltage, only sharing one between the 12v screen and the 5v USB. If I were trying to minimize the number of tranformers while maximizing the sonic quality & benefits, from my experience, I'd suggest the following combination of PSU's & regulators:
8v-11v PSU -> 5v Sigma regulator -> ATX-20/24 (P24) 5v (both power & control voltages)
14v-19v PSU -> 12v LT1085 regulator -> ATX-20/24 (P24) 12v -> 3.3v LT1085 regulator -> ATX-20/24 (P24) 3.3v (large voltage drop here, need a significant heatsink)
14v-19v PSU -> 12v Sigma regulator -> P4 12v
8v-11v PSU -> 5v LT1085 regulator -> HDD
14v-19v PSU -> 12v LT1085 regulator -> Zalman screen 12v -> 5v LT1085 regulator -> USB header power (for mouse & touchscreen)
If you wanted to use multiple-secondary transformers for the supplies, the minimum number I'd suggest you consider is three... one for the ATX-20/24 (P24), one for the P4, and another for the peripherals. But I believe that using 5 transformers as I suggested above will sound better (and using 7 as I currently do will be even better!).
Then to your questions about grounding and wiring, each PSU will have a ground point and that should be taken to the output connector for that power line so that the motherboard becomes the star ground point for everything. (I did just realized that I deviated from that in my 5v HDD supply... It reaches ground through the data connection for the HDD. I'll have to try grounding it to the motherboard & seeing if that makes a sonic difference!)
That also provides a method to help 'reject various interferences'. One common way to do this is by pairing a voltage connection with it's corresponding ground in a twisted pair or a braid. I did this as follows with my power supplies (note that I only used the 20-pin version of the ATX-20/24 connector and don't see any reason or recommend connecting up the other 4 pins):
ATX-20 5v - One set of 5v lines from the regulator to pins 19 & 20 in a 4-cross braid with ground line pins 16 & 17, another set of 5v lines from the regulator to pins 4 & 6 in a separate 4-cross braid with ground line pins 5 & 15
ATX-20 12v - 12v line from the regulator to pin 10 in a twisted pair with ground line pin 7
ATX-20 3.3v - 3.3v line from the regulator to pins 1, 2, & 11 (tied together at the plug) in a twisted pair with ground line pin 3
ATX-20 5vSB & PWR_OK - 5vSB directly from the 5v regulator, PWR_OK from a switch in a 3-braid with ground pin 13
Note that on the ATX-20, I did not connect pins 12 (-12v) 14 (5v PWR_ON#), and 18 (-5v)
P4 - 4-cross braid using the two 12v lines from the regulator and the two ground lines
On the 5v HDD, 12v screen, and 5v USB, I just used a twisted pair on each.
This will increase the number of connecting wires you have to use, but provide for better suppression and rejection of interference
Another thing to do to reduce interference transmission is to add bypass capacitors at each plug. I have them currently at the ATX-20 & the P4 and need to add them to the 5v HDD, 12v screen, and 5v USB plugs... see the picture of my supply above.
Then see here for additional suggestions for wire and connectors: http://www.altavistaaudio.com/Umbilical.html . While you wouldn't have to use the wires he recommends, these are two examples of wire that would be better than the garden-variety copper in vinyl insulation. Also, the Hirose connectors he uses look very good and are readily available at both Digikey and Mouser in a number of pin configurations. His comments also echo those of some others on this thread who have found sound quality improvements by using good connecting wires. I haven't done this in mine yet, but plan to with my next big-rebuild.
Then to your questions about the 5v PWR_OK, 5v PS_ON#, 5vSB, turn on sequencing & relays. As I said above, you can connect the 5vSB to the output of your 5v regulator and the 5v PWR_OK to the same place through an on-off toggle switch. You don't need to connect the PS_ON# at all. And you didn't ask this, but to make it all work, you do need a power-on switch connected to the power header on the motherboard. This is nothing fancy, I just use a momentary-on switch for this.
Then to turn on the computer, do the following:
1st - plug-in or turn-on the dirty supplies (P4 12v, HDD 5v, Screen 12v, USB 5v)
2nd - plug-in or turn-on the remaining supplies (ATX-20/24/P24 3.3v, 5v, 12v, & 5vSB)
3rd - press the momentary-contact power-on switch connected to the power header on the motherboard
4th - flip the switch to provide 5v to the PWR_OK connection
This will work for the GA-G31M-S2L motherboard.
If you have a GA-G31M-ES2L board, you also need to provide -12v at startup. This can be done with a 9v battery and can be turned off after the motherboard starts booting.
If you have a GA-EG45M-UD2H board, you have to wait 8-12 seconds after flipping the switch for the PWR_OK, then switch off the PWR_OK for 1-2 seconds, then switch it back on.
I am not sure what works for the GA-EG55M-UD2H board as I don't have one yet, but based on other's comments, I suspect that the sequence for the GA-G31M-S2L will work ok on it.
As you can see from this, no relays are needed at all.
Finally, a few more comments...
First, I use a separate 5v supply with a small transformer to provide power to the 5vSB & PWR_ON connections. This was a small, but noticable improvement over taking it from the main 5v connection.
Second, I'd strongly recommend using John Swenson's raw DC supply for computers as mentioned here: http://www.audioasylum.com/cgi/t.mpl?f=pcaudio&m=82140
He uses a choke filter, a snubber, and careful design to reduce the amount of noise generated in the supply and also fed back to the power line. I have not done this on my supply yet as I don't have room because it's designed to fit inside of the Zalman case, but when I re-do it in a larger enclosure, I will do it then.
Finally, cics had repeated his advice to use separate AC cords & circuits for the base ATX-20/24 (P24) supplies and the dirty supplys... P4, HDD, USB, and screen. This does make a difference and I strongly echo his advice.
All I can think to add for now!
Greg in Mississippi
Everything matters!
Greg,
When you say "I use a separate 5v supply with a small transformer to provide power to the 5vSB & PWR_ON connections.", what kind of supply and tranformer are required?
Bibo,
It was a nothing-too-special quick & dirty supply I did when I realized that the reason my system was cutting off at random with the linear supply was due to my use of a 9v battery to supply the 5vSB & PWR_ON voltages and turning them off after the motherboard had booted. It would run for awhile without these voltages, but would always cut out at some point, sometimes even days later.
It was an afternoon project using a transformer (with a center-tapped secondary) scrounged from an old linear wall-wart, a couple of MSR860 diodes, an LM340-5 regulator chip, and a couple of caps (pretty good quality ones... this proved to be audible, to my surprise!).
The 5v output is fed directly to the 5vSB & via a switch to the PWR_ON for the turn-on exercises. I did a 3-wire braid with these two wires and the ground from the supply, which went to pin 13 on the ATX-20 plug.
I hope this helps.
Greg in Mississippi
Everything matters!
WOW!!
This is'nt a response, this is a treatise!
Thx a lot Greg.
I need to study it with care, but my intention is to do a similar thing.
And yes, I think to use a toroidal to each different board.
Now I get 5 sigma but i must to decide how many single voltage I must use and for what.
But I'vnt clear another point, it could be trivial but not so clear for me:
from each board go out a couple of positive/ negative wire.
For P 24 connector where I'll connect the negative ones? With your order? all togheter?
Daniele
Daniele
Welcome. I hope it helps.
With 5 Sigma regulators, I'd suggest using them for the following voltages:
ATX-20 3.3v, 5v, 12v (3 Sigmas)
P4 12v (1 Sigma)
HDD/USB 5v (1 Sigma)
My reason for picking these voltages is that these are ones that will have a high amount of noise fed back to the regulator and the Sigma should do a pretty good job of dealing with this noise.
If you use separate supplies for the 5vSB & PWR_ON and for the 5v USB, no reason to use Sigmas, just go with LT1085/LT1083/LT317/LM317-type regulators.
I suspect, but do not have evidence, that it will sound better to have a separate power supply feeding the 5V USB than to use the same one for both the HDD and USB. But it's easier to start with just one and add another later.
As for where to connect the negative wires from each Sigma, this would be to the ground wire of the corresponding power connector. Let me revise what I said above to try and make it clearer:
ATX-20 5v - Two sets of 4 wires coming from the 5v regulator. One set is made up of a pair of wires taking the 5v positive to ATX-20 connector pins 19 & 20 and these are braided 4-cross with a pair of wires taking the 5v ground/negative to the ATX-20 connector pins 16 & 17. The other set is made up of a pair of wires taking the 5v positive to ATX-20 connector pins 4 & 6 and these are braided 4-cross with a pair of wires taking the 5v ground/negative to the ATX-20 connector pins 5 & 15. While both of these sets connect to the specified various pins on the ATX-20 connector at the motherboard, they all come together and are connected in parallel at the 5v regulator output at either the 5v positive or negative connections.
ATX-20 12v - One twisted pair of wires coming from the 12v regulator, made up of a 12v positive wire from the regulator to ATX-20 connector pin 10 twisted with a ground/negative wire from the regulator to ATX-20 connector pin 7.
ATX-20 3.3v - One twisted pair of wires coming from the 3.3v regulator, made up of a 3.3v positive wire from the regulator to ATX-20 connector pins 1, 2, & 11 (all tied together at the connector) twisted with a ground/negative wire from the regulator to ATX-20 connector pin 3
P4 12v - One set of 4 wires coming from the 12v regulator, made up of a pair of wires taking the 5v regulator's positive to the P4 positive (yellow) wires and these are braided 4-cross with a pair of wires taking the 5v regulator's ground/negative to the P4 ground (black) wires.
I really need to draw a picture... Sorry, I'm generally very bad about that. I'll see what I can do and post this week. I can also post instructions for making the 4-cross braids, if you need that.
Everything matters!
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:
Top of Page ]
[ Contact Us ]
[ Support/Wish List ]
[ Copyright © 2025, Audio Asylum, all rights reserved ]
