|
Audio Asylum Thread Printer Get a view of an entire thread on one page |
For Sale Ads |
84.58.45.168
In Reply to: RE: cMP - the open source high-end Memory Player posted by cics on December 30, 2007 at 05:42:01
Hi,
i found a solution to go directly into the juliboard with 3,3 Volt, which is the only needed tension for the digital section.
On the opposite side from the outpust there is an voltage regulator U1 from 5 to 3.3 volt, who can be unsoldered and we are able to feed 3,3 Volt ( four 6 volt batteries parallel via a thel regulator to 3,3 volt)to the board as You can see on the photo.
Sounds phantastic and is a real big step.
More depht, information and dynamics without any increase of harshness.
Highly recommended.
bye
alfred
my current stuff;
CMP/Cplay, PC with separated PSU for P4, P24, HD and Fans (Batteries for P4)
Julia I2S to buffalo( batterie PSU) to Zapfilter and to Tripath 2021 (batterie PSU)
Follow Ups:
Hi, I was wondering if it's possibile to use the molex of a picopsu to give the proper tensions to the esi.
In the photo you see the voltage of the molex and if you don't unsolder the regulator on the esi, maybe you could use the 5V and 0V of the molex. Do you think it's possible?
hi alfred, i have been looking at your enternal power mods for sometime..
Until today i started to mod like u. i cut all the +5v feeds from the juli@
card then solder 5v to DVCC1. When i turn on the external power of juli@ it
has power light from the optical jack, however when turning on my pc
optical lights goes off until booted in window it couldnt detect my sound
card.
is my external power supply ground have to connect to my computer psu
ground together? then only it will work?
is the +5V ground is separated with the 3.3v ones? tats y it cant work?
I've got my linear supply for powering the Juli@ separately completed and tested out today (see attached picture) and was looking at an alternative way to feed the voltages into the Juli@, by lifting the input pins on each of the regulators and feeding the appropriate voltage to each. My goal was to power the Juli@ separately without having to cut the voltage finger connections... with the revised caps, it doesn't fit into the computer on my header board anymore, so I can't just cut the connections on the header board.But what I found made me think that won't work... and it's also something to know for those who want to power the Juli@ separately when using only it's digital card. Poking around, I confirmed that the +-12v only shows up at the inputs to the appropriate regulator chips (U11, U13, U10) and the 3.3v is only provided by the output of U1, BUT I found the 5v coming from the motorboard going to several chips other than the U1 regulator... specifically U18 (the DAC) on the analog board and U2 and U8 on the digital board.
So there is likely another SQ improvement available (for both those using the Juli@ just as a digital board and those using the whole card) by taking Alfred's mod to another level by cutting the 5v PCI connector feeders and adding another +5v regulator feeding into the DVCC1 point.
And I'm back to cutting the 10 connections on the Juli@ so I can feed it from this separate PS. :(
Greg in Mississippi
P.S. I have very high hopes for this PS. It's a tweak-ish supply... no magnetic hardware, MSR860 diodes, Jensen 4-pole caps, Dexa regulators for the +-12v, 5v, & 3.3v, and minimal wiring.
We'll see if I get time to try it on the Juli@ tomorrow.
Edits: 04/11/09
Hi Alfred,
What cables are you using for the I2S connections? Did you DIY an I2S connector for the Juli@ or simply soldered them to the pins?
Thank you!
alfred you are one of the juli@ ps modifier leaders. I applaud you on your great work. I have a 5 volt battery (rechargeable) that I formerly used on a usb extender. Question do you think if I de-soldered the 5 volt feed from the pc (on juli@ card) and hooked up this battery the U1 regulator would work to provide 3.3 volts? I used the battery for an opticis usb optical cable.
Is there an english version description of your regulator. It looks very robust but I cannot read German (although my kids can).
hi theob,
The optimal specs of the U1 regulator:
MT1117-3.3, 4.3V ≤ VIN ≤ 5.5V ( Max Voltage not more than 6 Volt)
So it works perfect in this reach,
For the reason of low ESR i prefer 3 to 4 batteries parallel to reach the max. SQ , but one works too.be careful , the MT1117 is very sensitive for to much heat during soldering
Edits: 04/01/09
ok good advice I am not an expert solderer so I'll hold off untill I can find a good solderer. Should I unsolder the ground to run to the battery?
You can solder the ground of the batterie direct to the ground pin, the pin can stay on board.
thanks
Very nice, sir. For those using just the Juli@'s digital side, this will be a very easy way to clean up the power with a separate supply.
I guess it was a very good weekend for modifying Juli@ cards... see my posts below. I'm still using the Juli@'s analog outputs, so my mods covered both the analog card and the digital card. I also got some nice improvements with my mods... next for me is doing something similar to what you did, powering it with supplies that are separate from the computer's supplies and later, upgrading the regulators.
And I'll also be implementing a separate I2S-connected DAC, inspired by your post awhile back about connecting the Buffalo DAC to the Juli@.
Thanks for sharing and for the inspiration!
Greg in Mississippi
I've been toying with this idea for a while and if we could take all the innovation from Alfred, Greg and Promethk the result could be very interesting. The idea is to plug in a new PCB onto the digital section with:
- high quality caps providing cleaner power to Juli@'s digital section using power header pins
- would need to look into 3.3V clean power as per Alfred
- new DAC chips would be TI (Burr Brown) PCM1794A (hw control) which offers 132db SNR @ 9V output. this chip can be configured to have its internal reconstruction filter disabled (i.e. use external filter)!
- output stage using miniature tubes
- faraday cage?
- Neutric or Hicon XLR connectors
- card would use 2 PC slots
- DAC's 5V analogue input and 3.3V digital input is highly refined (also cater for battery supply)
- jumpers to set output voltage (9, 4.5 or 2=default) and ground connection
Juli@'s I2S interface is exceptionally good within a cMP² platform. The aim would be to have an analogue add-on PCB with performance rivaling "cost no object" DACs. (Of course at minimal cost including velvet packaging and white gloves).
cics,This idea of you has me very intrigued. I think there are several ways we could go on this...
1. Develop a new board using the PCM1794A that you mentioned above. I do like it's digital filter disable mode, but it does only do 24-bit.
2. Develop a new board using the AK4399 that I'm getting. It is another of the current 'superchips, most notably used in Alex Peychev's $30,000 NWO 3.0-GO modified Esoteric UW-1. It also has the digital filter disable mode, but it is a 32-bit chip... I'm not sure how much that matters, but it's good from a numbers perspective.
3. Develop an easy-to-implement mod of the ESS Sabre DAC, which has gotten a LOT of favorable press. What makes this an attractive option is the availability of Twisted Pear's ready-to-implement boards including an output IV stage. If the pricing for the new boards will be the similar to the last generation, this may be the most expensive of these options... but take the least development. This newest one is 32-bit AND also has the filter disable mode, so it remains a top contender. One good thing... the supporting I/V stage and power supply boards seem very highly developed.
4. A budget upgrade based on modifying the Juli@'s analog out, similar to what I did this last weekend. I don't yet know how good it will be, it hasn't fully run-in yet, but it's sounding very nice at three days in. Bass is especially well defined and strong, a significant jump over the stock card.
In the new-DAC world, I especially like the idea of dis-engaging the on-chip digital filter... I suspect there are significant gains available just doing that.
I am also expecting that it will be important to provide a separate power supply for best sound. As long as nothing goes too awry this next weekend, I hope to try that with the Juli@. I do have all the parts in-house now and started laying out the raw supplies last night. 4-6 hours of work on Saturday and it should be running. I'm also very interested to see how much of a difference it really makes in my setup where I start with the linear/hybrid supplies powering the computer. I still expect a significant bump, but it may not be as much as when separate digital-interface and DAC supplies are added to a computer SMPS-based setup.
Here's my version of your wish-list:
1. Phase I power mod to add additional filtering caps on the Juli@ digital board and on the DAC option chosen. Phase II power mods going to separate supplies for the digital And , either a linear or battery.
2. Phase I DAC mod to upgrade the Juli@. Phase I to implement one of the three DACs above (or any new contenders that emerge).
3. Output stage may have several options to choose from. The Twisted Pear setups offer nice turn-key stages. I do like the idea of a nice tube stage. I also like the idea of your multiple output voltage choices.
4. Location inside the computer case to minimize the I2S line length. Packaging to minimize RF and EMI interference would be important, such as your faraday cage. Mounting the I2S connections directly on the Juli@, maybe the entire DAC board, would be trick.
5. Balanced out via Neutric or Hicon XLR connectors. Un-balanced out also available.
If we want to investigate the PCM1794A or AK4399 options, it's probably worth getting a couple of the evaluation boards on-order.
One thought I have is that this is more of a DIYAudio-type of thread than most of the ones I see on the Asylum (except for the Tube DIY forum here). So I've started a thread for cMP2 mods there:
http://www.diyaudio.com/forums/showthread.php?s=&postid=1788543#post1788543
I don't know if you've seen it there, but Peter Daniel has posted a bit about his version of an AOB and later a cMP2 in the last few pages of his AudioSector-chip amp kits, dacs, chassis thread:
http://www.diyaudio.com/forums/showthread.php?s=&postid=1788548#post1788548
It was his implementation that was half of the inspiration to go to the linear/hybrid supplies... and the other half goes to the Asylum's Ryeland for his linear P4 power supply.
I invite consideration of moving that thread there, I suspect that the DIY nature of that board will get us more input than we'll see here.
My recent 2 cents.
Greg in Mississippi
Edits: 04/01/09
Greg,
interesting thoughts!
Just a brief correcting remark on the ESS Sabre for now: The current version is the ESS Sabre32, it is 32 bit (not 24) and the deemphasis and FIR filters both can be disabled.
Regards
Robert
Robert,
That's good to know and I updated my post above.
Honestly, if Twisted Pear had not be out of the Buffalo DAC at the time I'd decided which one to order, I'd probably have one of them now instead of waiting for the AK4399 DAC board. I don't feel bad going with the AK4399, but I liked the maturity of their design.
Two questions I have for the group...
One, you mention that the deemphasis is defeatable. One question I have is what happens with emphasised CDs? I happen to know (because I once made a CDP based on a CDP650 that had a demph-on LED) that a few of my CDs are emphasized. Will they play wrong. It's not like I stay awake at night worrying about this, but I am curious.
Second, does anyone have any tips on programming these things. It looks like the AK4399 will need programming to set the filter bypass on... and I wouldn't be suprised if the others didn't too. I suspect that the info on how to do this is in the ESS Sabre DAC thread on DIYAudio, but I haven't taken the time to look yet.
Again, thanks!
Greg in Mississippi
Juli@'s ASIO drivers remains the best I've seen to date.
New analogue board plugs onto this using I2S interface which I assume will support 24 bit DACs only (a deal breaker for 32 bit DACs). Alfred uses this interface for the Buffalo / Sabre DAC. Alfred: can you confirm 32 interface?
I'm not religious about 24 or 32 bit: important factors are SNR and the ability to bypass internal reconstruction filter. My AA Prestige SE uses the PCM1792 DAC which is excellent. Both AKM and ESS Sabre are excellent 32 bit DACs and yes the Sabre would be a brilliant choice (even better than 1794) given that we can disable its filter and includes IV stage & PS refinements.
So lets go with the Sabre and add RCA outs as well.
Will look into Peter Daniels' work as well. DIY thread is a great start (PS - my loyalties lie with AA).
Regarding Juli@ i2s out and ESS Sabre32:
Not sure wether Juli@ i2s out will work with ESS Sabre32 "as is" right away, will have to test. If it doesn't, this can likely be fixed by choosing a different PCM audio interface format - ESS Sabre32's default is 32-bit but can be set to 24-bit by programming the registers.
Documentation is light and so far, my readings suggest Sabre 2008 allows for disabling "Demphasis" only. Can't get any info on Sabre 2018.
Just thinking aloud:
The ESS Sabre32 has 32bit PCM audio input stream as default - would it be possible and would it make any sense to have SRC resample to 32/192...? Haven't thought about this yet, but apparently Asus Xonar soundcards can output 32bit (at least Xonar DX can, see test here: http://www.atomicmpc.com.au/Review/116715,asus-xonar-dx-pci-e.aspx) - not sure if Juli@ can.
Anyone any thoughts if 32bit make any sense?
.
32 bit i2s and filter disable confirmed
Edits: 04/04/09
The little buffalo /sabre via I2S IMHO is as good and even better than every cost no object DAC i know.
The julia with 3,3 Volt from batteries and additional low ESR is another beast than the untweaked card, and the combination with the I2S/Buffalo
is in this moment the easiest way for my needs.
Why You want to do everthing in the PC?
Is Your aim a self developed DAC board ?
Three reasons:
- offers lowest DAC cost (say ±$200 without skimping on velvet and gloves). there's no need for XOs, casing, PSU, etc as Juli@ provides I2S interface and power can be sourced from main psu (with extra filtering on DAC's PCB) or preferably a battery option
- can fully exploit cMP² capabilities like no reconstruction filter in DAC
- elegance of having just one box
The buffalo / sabre DAC is a superb choice.
to the first point:
if You put it in the same case and You want best quality for Julia and DAC you need a battery PS and my experience shows ,that You need at least 3 or 4 sealed batteries to provide lowESR and they don't fit in the case.
The I2s is a little problem because quality is decreasing quickly after 20cm length of the cables. My buffalo is direct behind the PC case , but it is not difficult to integrate it in the same case if you want.
Seconfd point i cannot comment; sounds interesting.
Elegance of one case : ok for the WAF, but for DIY-people free-wired open elektronics provide headroom for new developments .
P.S. cics, I have something else I wanted to discuss with you. Please email me via my profile.
Thanks!
Greg in Mississippi
.
LOL on the velvet packaging & white gloves.I'll take a look at that DAC later tonight. I will also take another look at the AK4399 that's used in the add-on DAC card I've got ordered to see if it also has the option to disable the internal filtering... that'd be trick!
My gut feel from adding linear supplies to my setup is that disconnecting the card from the computer's supplies will make a larger difference than anything else we can do. Alfred's comments also suggest this.
Also, given the output voltages available, a tube output stage may not be needed.
Thanks for getting the wheels turning on this.
Greg in Mississippi
Edits: 03/31/09
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: