[stokessd]

Sheldon, could you elaborate? What would this entail?

Sure thing.  All oscillators are very sensitive to the power being fed to them.  And as thanks, they toss all sorts of crap back on that power line. 

So you see in the schematic and layout in my latest DAC design (http://getinthewoodchipper.com/?p=302, I've got a ceramic cap as near the oscillator terminals as possible, and a big tantalum (100uF) cap right next to that.  The power feeding the oscillator has a resistor in the line, forming an RC filter right at the oscillator.  For my prototype, which you see in the link above, I used a 25ppm oscillator and that filter/decoupling method and called it good. 

But after I finished the prototype, I realized how easy it would be to add a little inverting amp as a noise eater.  I got this idea from Wenzel (the precision oscillator folks), and it's brilliant in areas where you don't need hugely stiff supply rails.  The idea is that a standard three terminal regulator will still have like 100uV of broadband noise on it under the best of conditions, plus the noise from whatever else is barfing into that same line.  If you build a little very low noise amplifier that takes the AC noise component as an input and outputs an equal and opposite signal, the two cancel and you are left with clean power.  The full story is here:  http://www.wenzel.com/documents/finesse.html


So in the schematics and layout of my DAC in the first link, you see the oscillator is fed from the 8421 logic supply and decoupled from it with a resistor and the caps right at the oscillator, then there is a little single transistor NPN amp which is the noise eater.  Sadly my home measurement gear isn't sensitive enough to tell me how good it gets. 

I'm not sure there is an easy way to implement this technique on the little upsampler board, but it certainly is possible.  If memory serves me right, the oscillator in the gigaworks DAC already has a resistor in the power line, so you'd have to add a transistor, a cap and four resistors.  You could build it up as a little daughterboard that is attached behind the oscillator. 

Honestly, I'm not sure how big a return on investment it really is.  I haven't tried it on my prototype DAC.  I've only used it for other projects which aren't audio related.  I added it to the layout because it's so quick and easy and I don't have to populate it if I don't want to.


Sheldon

[Kwikas]

The input chip in the gigaworks DAC can take up to 4 inputs (in hardware mode), but that is 4 discrete SPDIF streams you choose from like an input selector, not simultaneously.  SPDIF is a two channel protocol, there is no provision in SPDIF for doing more than two AFAIK.  So you would need to run four streams for eight channels.  Thus four DAC's etc. 

Sheldon

Thanks Sheldon. Yes, I'd have to run 4 streams simultaneously. I''ve got that part happening out of my transceiver board. But the DAC part is a lot more difficult.. I'll see whether I can tackle this another way. Cheers

[Atlplasma]

Hi guys:

After three weeks of waiting, I finally received this DAC today. Unfortunately, the cap at C23 has come loose. It dangling by one of the leads. Can I heat of a drop of solder and reattach it, or does it need to go back? I hate to wait another 3 or 6 weeks (depending on the return policy) to get it up and running.

[stokessd]

I don't have the most recent schematic in front of me, but the older one shows that the cap in question is a bypass caps for the Analog voltages in the receiver chip.  You should try to solder it back in place, but even with nothing there, you should still be fine.


Sheldon

[Atlplasma]

Thanks Sheldon. Know an easy was to solder a SMD miniature capacitor? Solder paste and a heat gun seems like a possibility. It kind of looks like I should clean up the connection point with a bit of wick before doing anything else.

[stc4life]

So I am totally stumped.  I have the DAC completely hooked up, using Edcor 10k transformers, but I cannot get any sound.  I don't know what could be the problem

The only real mod is that I removed the caps after the 4398 and soldered leads from the top of the board to the transformers.  Using 13k resistor and 680uf cap across the primaries and have the standard resistor and cap combo on the secondaries.  I am getting a fairly uniform voltage at the primaries and at the top of the 4398 board - they do not really seem to be jumping with the music, but are different for different tracks (usually between 1.5v-2.5v, which seems right).  I am getting nothing at the secondaries and no sound.  Also, no sound if I bypass the transformer entirely and hook it directly to the rca.  I get a strange descending sweep of the spectrum when I power down, but otherwise no sound. 

I am getting about 5v at pin 22, 4.5v at pin 27, and 3.30 at pin 7 on the 4398 board.  I am getting 5v at the caps following each of the 317 regs.  Another strange thing is that I am getting opposite voltage readings at the same points if I use the ground on the IEC or the ground on the board.

I am getting to my wits end...  Anybody have any ideas?  I am going to post on diyaudio as well, but that thread is more focused on complex issues.

[wushuliu]

So I am totally stumped.  I have the DAC completely hooked up, using Edcor 10k transformers, but I cannot get any sound.  I don't know what could be the problem

The only real mod is that I removed the caps after the 4398 and soldered leads from the top of the board to the transformers.  Using 13k resistor and 680uf cap across the primaries and have the standard resistor and cap combo on the secondaries.  I am getting a fairly uniform voltage at the primaries and at the top of the 4398 board - they do not really seem to be jumping with the music, but are different for different tracks (usually between 1.5v-2.5v, which seems right).  I am getting nothing at the secondaries and no sound.  Also, no sound if I bypass the transformer entirely and hook it directly to the rca.  I get a strange descending sweep of the spectrum when I power down, but otherwise no sound. 

I am getting about 5v at pin 22, 4.5v at pin 27, and 3.30 at pin 7 on the 4398 board.  I am getting 5v at the caps following each of the 317 regs.  Another strange thing is that I am getting opposite voltage readings at the same points if I use the ground on the IEC or the ground on the board.

I am getting to my wits end...  Anybody have any ideas?  I am going to post on diyaudio as well, but that thread is more focused on complex issues.

re-reading your post I'm assuming when you say primaries for the 13k/620pf (not uf, !!!), you really mean secondaries and vice versa... if not, then that may be the problem  .

[wushuliu]

AD1865 DAC Update:

I have replaced the diodes by the power supply w/ fast 1A Schottkys (thanks Dave!), swapped out half the resistors in the I/V stage for PRPS and dales and half the caps for Elna Silmic II.

Here's the deal: At this moment, I have glimpsed audio nirvana. For these few moments I know why some of us audiophiles just keep on searching, keep on striving, spinning around in what seems like circles like dervishes. Because when everything falls into place and the music opens itself to you, it is a kind of bliss.

I reached that tonight w/ this dac. I would never have thought I would. Even this weekend I was pretty set on sticking w/ the Gigawork with the tranny mod. It just had a more open and engaging sound overall even though the 1865 sounded really, really good.

After swapping out those parts tonight, it's a whole other story. I've been wondering why the Lampizator article waxed so effusive about this dac kit. I mean, I was impressed but not overly so. But now I get it. If his output stage is indeed better than the kit's, I get it.

I'm going to give it a day or two. Maybe I'm just in a mood, maybe the full moon is out. Maybe it's the new xo in my speakers or A/C cleaning stuff I've been messing with.

Right now... 


 

[stc4life]

re-reading your post I'm assuming when you say primaries for the 13k/620pf (not uf, !!!), you really mean secondaries and vice versa... if not, then that may be the problem  .

Thanks Wushuliu.  I did mean 680pf.  I do have the filters on the primaries and secondaries switched, so I will switch those and try everything again.  I will report back.

[wushuliu]

Update: Gave a listen this morning and a slight grunge has crept in, probably all the new stuff settling along w/ the typical L.A. daytime dirty AC (sound is always best after 11pm here, no lie). Nevertheless the dac still shines. Here's what I'm experiencing.

Lower noise: The sucky thing about noise in gear is that you never know how much you have til it's gone. The lowered noise floor of this modified dac now allows music to approach 3 dimensionality. I have a Chesky Audio Demo Disk and at one point the narrator mentions that some vocalist should sound 'palpably real'. I always chuckle at that. What the heck does that mean? Well, now I kind of understand. Vocals now hover about a foot or so front and center. Miles Davis' trumpet sounds like it leaning towards me. And I'm not talking SACD Kind of Blue, where you'd expect that - or at least hope. I'm talking mp3 on my WDTV/hard drive.

Soundscape/stage/whatever: is much wider. Like when you go to a good movie theater like the Egyptian and they're about to show Lawrence of Arabia and the screen expands and expands and expands to glorious 2:35:1 ratio. This also means I find myself turning up the volume for certain tracks I previously would have left alone.

Transparency: Up to this point, the AD1865 has been more or less on par w/ the Gig Tranny as far as transparency, with the Gig maybe having a slight edge due to less of a roll off in the HF like the stock AD1865. Now not only have the tables turned, they've been thrown out the window. The AD1865 now excels in transparency. I have never felt closer to my music.

Tonality: This is a little tougher to gauge as of yet because I have just changed my speaker crossovers, incl. some PIO caps so it's very probable that what I am hearing as far as gauging tonality right is a mix of the DAC and the xo's. I think it's safe to say that the AD1865 has more of a relaxed tone in general, the 'natural' adjective that has been used to describe it is apt. I will eventually put the Gig back in to compare for sure, but from day 1 the AD1865 has displayed less of a 'high fidelity' sound than the the cs4398, but without loss of detail.

I have ordered more Dale resistors and Elna caps to replace as much of the rest of the I/V stage as I can. Hopefully this will improve things even more.

I'm havin' fun

[jamesdb]

Hi Wushuliu, I am interested in the AD1865N-K DAC. What value (uF & V) were the 4 panasonic FC caps you replaced with oscons that removed the dark sound? Can you tell if it has a digital input transformer on the SPDIF input? Thanks, Jim B

[wushuliu]

Hi Wushuliu, I am interested in the AD1865N-K DAC. What value (uF & V) were the 4 panasonic FC caps you replaced with oscons that removed the dark sound? Can you tell if it has a digital input transformer on the SPDIF input? Thanks, Jim B

Hi Jim: 47uf, 16(?)v. 10v and up is fine for those caps.
There is a sumlink transformer. I took that off per the Lampizator tutorial - I believe
I provided a link to that in an earlier post.

[jamesdb]

I noticed that the same ebay seller of the AD1865N-K DAC also sells 2 USB to SPDIF converter modules. Through other google searches and ebay searches I found a number of other plug-and-play converters ranging from about $20 to $100 or more. Another option to removing the digital input transformer is to replace it with a higher quality one. Several posts recommended Newava Technology Inc S22083 at Digikey, a Lundahl amorphous core, and an Audio Note one with nickel core and copper or optional silver wiring, ranging from $9 to $99.

[jamesdb]

Wushuliu, reading your posts and the Lampizator article on the AD1865N-K DAC inspired me to buy one. It just arrived (pretty fast and packed very well) and I have a couple of questions since no instructions came with this other than the schematics.
On the power toroid which has dual primaries, the two red primaries should connect together for HOT and the two black primaries should connect together for NEUTRAL. Did you connect the single green primary to the safety ground coming in at the IEC? Did you make any other ground connections? It looks like the 2-terminal blue connector +5V & GND on the side by the DAC chip is not used, is that correct? I was not sure if that is for Toslink connections or not. And next to that are 2 connection points labeled G and ER , did you leave those unconnected as well? The Lampizator photos don't seem to show them in use.
I also tried to identify the caps you replaced that removed the dark sound, but I don't see any Panasonic FC caps on this. Were these the same six caps (silver 15uF 20V)

Lampizator replaced with brown dipped tantalum 47uF 25V caps in his photo , or some others?  I noticed my board does not have the 2 large SIC-SAFCO 4700uF caps in the digital receiving power section, it has Elna For Audio there, so that is a nice free upgrade.

[wushuliu]

Wushuliu, reading your posts and the Lampizator article on the AD1865N-K DAC inspired me to buy one. It just arrived (pretty fast and packed very well) and I have a couple of questions since no instructions came with this other than the schematics.
On the power toroid which has dual primaries, the two red primaries should connect together for HOT and the two black primaries should connect together for NEUTRAL. Did you connect the single green primary to the safety ground coming in at the IEC? Did you make any other ground connections? It looks like the 2-terminal blue connector +5V & GND on the side by the DAC chip is not used, is that correct? I was not sure if that is for Toslink connections or not. And next to that are 2 connection points labeled G and ER , did you leave those unconnected as well? The Lampizator photos don't seem to show them in use.
I also tried to identify the caps you replaced that removed the dark sound, but I don't see any Panasonic FC caps on this. Were these the same six caps (silver 15uF 20V)

Lampizator replaced with brown dipped tantalum 47uF 25V caps in his photo , or some others?  I noticed my board does not have the 2 large SIC-SAFCO 4700uF caps in the digital receiving power section, it has Elna For Audio there, so that is a nice free upgrade.

Hi James, my dac was in a temp wood enclosure up until recently so did not wire the ground, but that's probably a good idea.

for 120V, yes wire black together for neutral, red for line.

Don't know about that +5v and G/ER. That is a good question about the toslink. Maybe ask the seller? He's pretty responsive. He sells a separate multi-input toslink/coax/usb/bnc module too which would be perfect for that +5v btw if thats the case.

The silver caps removed are the ones replaced by the tantalums. I used Oscons instead per Dbe's recommendation and they work wonderfully.

You will also want to replace some of the output resistors in the jfet stage right by the L/R output.

[wushuliu]

Hm. $2500 AD1865-based DAC.

http://www.mojo-audio.com/AD1865NOS-D-A-Converter-AD1865NOS.htm

If you enlarge the interior photos, that board suuure does look familiar!

[etcarroll]

25 hundred, and its backordered. I may have to do this project after my SACD project.

What kind of enclosure are people using for these?

Gene

Hm. $2500 AD1865-based DAC.

http://www.mojo-audio.com/AD1865NOS-D-A-Converter-AD1865NOS.htm

If you enlarge the interior photos, that board suuure does look familiar!

[wushuliu]

25 hundred, and its backordered. I may have to do this project after my SACD project.

What kind of enclosure are people using for these?

Gene

I use an Antek 12*12*3 gold alodine.

[jamesdb]

Sure does look familiar. Some of the capacitors they replaced with OSCON SEPC caps on the stock board went into positions where the stock caps were 20V and some maybe 25V, however the data sheets for the OSCON SEPC series shows a max rating of only 16V. Wonder if they will last long, unless the stock caps are over-rated to begin with. Wow, they also sell "quantum purified" digital cables.

[srb]

Hm. $2500 AD1865-based DAC.
http://www.mojo-audio.com/AD1865NOS-D-A-Converter-AD1865NOS.htm
If you enlarge the interior photos, that board suuure does look familiar!

Every part would have to be "quantum purified", kissed by a virgin and blessed by a priest, and even then, $500 max.  Of course it probably would have synergy with your $3500 Lexicon/Oppo Blu-ray player.
 
Steve