Experience with this DIY DAC ? - Page 393 - diyAudio
Go Back   Home > Forums > Source & Line > Digital Line Level

Digital Line Level DACs, Digital Crossovers, Equalizers, etc.

Please consider donating to help us continue to serve you.

Ads on/off / Custom Title / More PMs / More album space / Advanced printing & mass image saving
Reply
 
Thread Tools Search this Thread
Old 21st August 2011, 07:19 PM   #3921
gsa is offline gsa  United Kingdom
diyAudio Member
 
Join Date: Nov 2010
Location: West London
Quote:
Originally Posted by i2k92 View Post
Very interesting thread. I plan to buy this kit, but there's 2 version : red PCB and green/blue (unclear) PCB...which one is the new & most updated ? I believe both version support upsampling module, am I right ?

Not sure if this board is available any more, the blue one was the last in the line AFAIK.

Take a look at Sheldons board it will build into a great sounding DAC without needing massive surgery.
  Reply With Quote
Old 21st August 2011, 08:02 PM   #3922
gsa is offline gsa  United Kingdom
diyAudio Member
 
Join Date: Nov 2010
Location: West London
G'loops I have attached a bunch of photographs. I will attempt to explain what I have done.

OK so the power supply rails on the board are shared in a crazy way as standard.

The whole 8416 board is being supplied by a single 3.3v rail with no capacitor decoupling evident.

There is a dedicated 5v supply for the SPDIF toslink socket???

The upsampling board and VA for the CS4398 share a 5v rail.

The USB has a 5v dedicated rail.

The relay has a 12v supply.

Opamps have a bipolar 12v PSU.

I think thats everything.

----------------------------------------------

CS8416

OK I cut traces on both sides of the board under the CS8416 smurf board to isolate VA[3.3v] pin 6. I cut a connecting trace between pin 23 & 21 just outside the socket for the 8416 this isolated VD and VL. Traces on the MAIN BOARD under the smurf board area not actually on the smurf board.

I added 0.01uf X7R caps with 100uf electrolytics directly to the pins of the 8416 smurf board there is a ground pin on the neighbour pins 7 & 23
** note I havnt soldered 2 100uf caps to VD & VL yet.

So 3.3v seperate PSU for pin 6 VA & pin 23VD and 5v on pin 21VL.

The 5v VL PSU is used to power the optical SPDIF and the pullups for selecting the input.

Also put a 47K resistor from VL to pin 20 to change the jitter something as discussed early on in this thread.


CS8421 board.

Cut trace next to the CS4398 chip between the 3.3v rails from the SMD regulators, this seperates the 5v VA for the DAC from the CS8421 board.

Added a 0.01uf X7R cap to the supply pin to ground on the CRYSTAL.
Remeoved the regulator and associated input caps and jumped the supply using a copper wire.

Now supplied by a dedicated PSU using LM317 etc.

CS4398 board.

Supplies have all been seperated now so power this up as required with 3.3 volts and 5 volts.

I added 0.01 X7R with 100uf to VD VL and Vref - for Vref there is ground plane just below on the underside of the smurf board use this and solder side on to the board. The CS4398 board is NOT exactly pin for pin like the CS8416 board so be carefull.

*** Vref is pin 15 on the smurf board but its leg 17 on the chip please be warned.

So I think thats it.......


In total I have 6 PSU lm317 to power the chips and a 15v bipolar for my opa627 output board.

I think its best to do this in stages. Seperate the supplies and get rid of all the ferrite beads they suck the life out of the sound.

Phew.......
Attached Images
File Type: jpg IMAG0111.jpg (860.2 KB, 606 views)
File Type: jpg IMAG0122.jpg (853.6 KB, 562 views)
File Type: jpg IMAG0124.jpg (385.5 KB, 497 views)
File Type: jpg IMAG0112.jpg (616.4 KB, 480 views)
File Type: jpg IMAG0121.jpg (827.6 KB, 469 views)
File Type: jpg IMAG0123.jpg (956.4 KB, 110 views)
File Type: jpg IMAG0125.jpg (736.3 KB, 98 views)
File Type: jpg IMAG0126.jpg (809.6 KB, 121 views)
File Type: jpg IMAG0127.jpg (446.4 KB, 144 views)
File Type: jpg IMAG0128.jpg (897.6 KB, 147 views)

Last edited by gsa; 21st August 2011 at 08:07 PM. Reason: clarify statement
  Reply With Quote
Old 21st August 2011, 08:36 PM   #3923
gsa is offline gsa  United Kingdom
diyAudio Member
 
Join Date: Nov 2010
Location: West London
Can anyone help me understand how to find out what the input impedence of a given filter network would be? eg I am using the CS AN048 multiple feedback filter with a roll at aprox 50Khz.

I want to raise the cut off Fc to aprox 100Khz. This changes the resistor values, what is the resulting input impedance???

I used this utility Multiple Feedback Low-pass Filter Design Tool

CS AN048 values are 13.7K and 3.32K so is the input impedance aprox 13.7K the first resistor?

And so if I have values of
R1 = 3.3kΩ
R2 = 3.6kΩ
R3 = 3.3kΩ
C1 = 0.001uF
C2 = 220pF

using the utility above I get Fc of aprox 100Khz

Do I have an input impedance of aprox 3.3K......this would be OK for the CS4398 to drive; according to the datasheet?

I studied Mech. Eng and not electrical so on a steep learning curve. Trying to dodge the enevitable fact that I will have to hit the books to re-apply earlier learning....its all a bit similar to Dynamics and Control lots of laplace & Z transforms and matrices thrown in for good measure.

Help me dodge this bullet....
  Reply With Quote
Old 21st August 2011, 08:52 PM   #3924
Ryssen is offline Ryssen  Sweden
diyAudio Member
 
Ryssen's Avatar
 
Join Date: Dec 2003
Location: Sweden--> Here
Quote:
In total I have 6 PSU lm317 to power the chips and a 15v bipolar for my opa627 output board.
And if you get rid of the opamps and install a pair of OPT you get to hear the
CS4398´s potential..
__________________
Regards Åke
  Reply With Quote
Old 21st August 2011, 09:40 PM   #3925
gsa is offline gsa  United Kingdom
diyAudio Member
 
Join Date: Nov 2010
Location: West London
Hi Risen,

At some point I will try using Trafo's, I am fully aware of the massive following they have in this thread. I am also curious to hear what they sound like, I have an open mind on the subject.

I have concentrated on the source of the signal going into the output stage working on the "Rubbish In Rubbish Out" principal. Its easy to fall into the trap of concentrating on the glamerous stuff.

I and others have carried out a fair bit of work on this area and it will benifit both opamp and transformer. A good and fast opamp is very good at revealing a poor signal source even if it sounds "electronic" to some ears. Transformer might sound better subjectively to some ears.

You can put a Transformer on the output and leave it at that however you are not even hearing 50% of the boards capability if this is the only thing you do.

George.
  Reply With Quote
Old 21st August 2011, 10:38 PM   #3926
Ryssen is offline Ryssen  Sweden
diyAudio Member
 
Ryssen's Avatar
 
Join Date: Dec 2003
Location: Sweden--> Here
I have tried many opams and I allways end up with the LM4562 as being the best sounuding,and the thing about rubbish in rubbish out if you put a trafo instead of the opamp,then you hear what was making the rubbish..
I have 4 shuntregs on the DAC and other various tweaks,so let hope they make things better for the DAC.
__________________
Regards Åke
  Reply With Quote
Old 21st August 2011, 11:07 PM   #3927
gsa is offline gsa  United Kingdom
diyAudio Member
 
Join Date: Nov 2010
Location: West London
Ryssen, an interesting point you raise is that you like the LM4562. I think its a very good opamp. When I first listened to it I was very pleased with the SQ it was very impactfull and had good pace........etc etc all subjective stuff.

On closer examination of the sound based on vocals sibilance transparency etc [more analytical listening but still subjective I suppose] OPA627 was ahead, however overall LM4562 was still more pleasing to listen to, why?

As the upgrades were made and I kept swapping between NE5532 OPA627 & LM4562 I realised what was happening.

The OPA627 is definately the most revealing out of all the above OPAMPS and for this reason it is the hardest to get a good sound from.

Its only now after all theses mods that I have been able to get the OPA627 to perform. Now when I put the the other OPAMPS in I lose the magic, the detail, not much detail in the case of the 4562 more with the NE5532 gets lost along with transparency and dynamics.

I found placing a 600 ohm resistor in series on the output of the 627 and decopling supply rails properly is essential for fast high bandwidth OPAMPS.

If you are making PSU upgrades its not enough to just change the existing LM317 to better regulator designs. It will improve the SQ for sure, but its limited. There are fundamental design flaws with the power distribution on this board.

The best and most fundamental SQ improvement is only obtained when the supply rails are seperated properly and the VL on the 8416 is raised to 5v. Get rid of the ferrites! Add decoupling and you enter another world.
  Reply With Quote
Old 22nd August 2011, 12:52 AM   #3928
diyAudio Member
 
stokessd's Avatar
 
Join Date: Nov 2002
Location: Grantham, NH
Quote:
Originally Posted by gsa View Post
At some point I will try using Trafo's, I am fully aware of the massive following they have in this thread. I am also curious to hear what they sound like, I have an open mind on the subject.

I was skeptical at first as well, and I was in the middle of a design using the same chips, so I added a set of Lundahl outputs to the design, and I'm amazed at how natural and lifelike they sound. And I'm not a transformer coupling guy at all.

Also the ground plane is as much or more important than the power supplies. I use 7 discrete power supplies just for the three chips on my design. And the chips are down on the ground plane.


Sheldon
  Reply With Quote
Old 22nd August 2011, 01:22 AM   #3929
gsa is offline gsa  United Kingdom
diyAudio Member
 
Join Date: Nov 2010
Location: West London
Hi Sheldon,

I am familiar with your board and I like it. IMHO its correctly designed to extract the best from this trio of chips. The work I am doing with the Gigaworks is simply bringing the design closer to your board.

I will try the transformer output, I just stopped work on the output stage altogether when I realised how much performance was locked away in the chips!!

When I feel that I have experimented and squeezed [or break] the best from the chips I will look at the output stage again, Trafo's and OPAMP filters.
  Reply With Quote
Old 22nd August 2011, 06:11 AM   #3930
data is offline data  Australia
Account disabled at member's request
 
Join Date: Jul 2007
Location: Brisbane
I would love to buy one of your boards, Sheldon, maybe once I fix my car I can see if you still have some

Though I'm a bull in a china shop when it comes to working with SMD stuff
  Reply With Quote

Reply


Hide this!Advertise here!
Thread Tools Search this Thread
Search this Thread:

Advanced Search

Posting Rules
You may not post new threads
You may not post replies
You may not post attachments
You may not edit your posts

BB code is On
Smilies are On
[IMG] code is On
HTML code is Off
Trackbacks are Off
Pingbacks are Off
Refbacks are Off


Similar Threads
Thread Thread Starter Forum Replies Last Post
A.O.S - Any experience? Keltern Multi-Way 16 26th April 2012 11:15 PM
my t-amp experience Eggzy Chip Amps 35 17th December 2005 06:29 PM


New To Site? Need Help?

All times are GMT. The time now is 10:41 PM.


vBulletin Optimisation provided by vB Optimise (Pro) - vBulletin Mods & Addons Copyright © 2014 DragonByte Technologies Ltd.
Copyright ©1999-2014 diyAudio

Content Relevant URLs by vBSEO 3.3.2