Continuation of R2R DAC Discussion From Stax Thread

[joehpj]

http://chihhanlin.weebly.com/2447825286272313552736215.html

Though this was written in Chinese, it was still clear Denafrips suffered from glitch and THD was not good when using USB input. 

The author said switching to coaxial or SPDIF would solve the glitch and cold sound.

I assume the architecture was almost the same between the tested and the one here.

[spritzer]

7 hours ago, wink said:

Just get the YGGDRASIL.....   

Kevin and I are thinking about acquiring one as the measurements for it don't all add up. 

6 hours ago, joehpj said:

http://chihhanlin.weebly.com/2447825286272313552736215.html

Though this was written in Chinese, it was still clear Denafrips suffered from glitch and THD was not good when using USB input. 

The author said switching to coaxial or SPDIF would solve the glitch and cold sound.

I assume the architecture was almost the same between the tested and the one here.

I've been looking at the USB input and I don't really see why it would be increasing the THD and I'm frankly more worried about the RCA input.  Optical should just be thrown in a fire as it is garbage on everything but the coax has just a 75R resistor to ground and then AC coupled to the AKM chip.  No transformers or anything like that and if you are using some odd cables, this could be compromised quite heavily.  I'll try it but I might also go and get some 75ohm BNC's and do this properly.  

[joehpj]

42 minutes ago, spritzer said:

I've been looking at the USB input and I don't really see why it would be increasing the THD and I'm frankly more worried about the RCA input.  Optical should just be thrown in a fire as it is garbage on everything but the coax has just a 75R resistor to ground and then AC coupled to the AKM chip.  No transformers or anything like that and if you are using some odd cables, this could be compromised quite heavily.  I'll try it but I might also go and get some 75ohm BNC's and do this properly.  

The USB XMOS driver was said to deliver 88.2KHz noise into DA module. I don't know if the newest driver has already solved the bug.

Edited February 16, 2017 by joehpj

[spritzer]

I would have to check for that but given that this an off the shelf design, any driver issue would have been quickly resolved. 

Also the Denafrips dac tested there is clearly different and substantially so.  Rather different string setup. 

[gepardcv]

11 hours ago, spritzer said:

Kevin and I are thinking about acquiring [an Yggdrasil] as the measurements for it don't all add up. 

Interesting.  Got any more info?

PS: I have a Gungnir Multibit I might be talked into lending to you and Kevin for experiments (though it'd be nice to get it back in one piece eventually)...

[nopants]

i can donate mine to the cause if someone wants to impedance match the bnc input haha

[mypasswordis]

Tried the kill the clock mod and what it kills for me is any lock onto the signal (no sound) so I removed it. Thankfully I didn't kill the receiver and the DAC still works. Anyone have experience swapping YM3623 out for CS8414?

[spritzer]

I was pulling the MSB out of my system and here is inside one of the dac modules.  The cover was simply soldered in place which the Metcal simply laughed at...   

[nattonrice]

Mushed spider ground topology?

[kevin gilmore]

looks like 20 bit plus 7 segment

supposed to be sign/magnitude but not sure of that

[alvin1118]

On 2/16/2017 at 0:30 PM, joehpj said:

http://chihhanlin.weebly.com/2447825286272313552736215.html

Though this was written in Chinese, it was still clear Denafrips suffered from glitch and THD was not good when using USB input. 

The author said switching to coaxial or SPDIF would solve the glitch and cold sound.

I assume the architecture was almost the same between the tested and the one here.

Just checked this out again, it's the old Denafrips DAC-2014 R2R. Not the Ares that @spritzer bought.

[Pars]

Anyone heard or looked at ECDesign's Mosaic UV Dac? John Brown has been quite active on DiyA with NOS TDA1541A designs, but apparently decided that the chip had too many design problems to go further than he had. He designed and implemented a discrete Multibit Dac design which has gone thru several iterations. I think they discuss it some on SBAF, but haven't read it.

Not sure if this guy and disciples are of the "mystic" variety ala lampizator, etc. though...


Sent from my iPhone using Tapatalk

[mypasswordis]

I assume one of the design problems is the chip hasn't been made in like 20 years and his stock was running low. Any posts on SBAF should be taken with a massive grain of salt, but the DAC looks interesting. Will go through his diyaudio posts to find out more, thanks.

[Pars]

Yeah, TDA1541A availability certainly is an issue, particularly when you need to parallel 8 of them or so...
And agreed on grain of salt regarding SBAF...


Sent from my iPhone using Tapatalk

[TMoney]

3 minutes ago, mypasswordis said:

Any posts on SBAF should be taken with a massive grain of salt

Amen!

[kevin gilmore]

so I bought a bunch of different 10mhz master reference oscillators to see if there was any difference.

and yes there is a big difference, this is my favorite. a rubidium disciplined ocxo

 

[luvdunhill]

Generally, is there enough “drive” in these “scientific” oscillators to input directly into consumer-grade HiFi? I also would need 10MHz for my Esoteric and would love to try one of these out, maybe you have an extra?

[kevin gilmore]

this one has a 5 volt 50 ohm output, so yes. sorry don't have a spare, a lot of the ones on ebay are used and the lightbulb has a fixed lifetime, have been searching for a new one for a while

[Craig Sawyers]

The current generation Rb units put a whole lot more Rb in to give a much higher lifetime. For example Stanford Research's PRS10 has 1mg load rather than 100ug http://www.thinksrs.com/products/PRS10.htm and 20 year lamp life.

Leapsecond have done their usual comprehensive tests http://www.leapsecond.com/museum/prs10/

But the ultimate in Allen Variance is the BVA version of an SC cut crystal; the only currently available one is the RA-14 from Rakon http://www.rakon.com/products/families/ocxo-ocso . It was previously available from Oscilloquartz until it was bought out (spec attached)

But you need to be determined to get one - they are around ten grand new.

 

Oscilloquartz BVA.pdf

Edited October 22, 2017 by Craig Sawyers

[wink]

We used to have a rubidium and also a caesium standard when I worked in the cal lab.

From what I could gather , as stand-alone units, ther caesium standard was the better in the long term, but the rubidium was yhr\e more accurate in the short term.

That was, until the rubidium was connected to a GPS unit that updated it every day.

https://www.vectron.com/products/rubidium/rubidium_index.htm

https://spectracom.com/products-services/precision-timing/frequency-standards

 

Forum chat.

http://www.eham.net/ehamforum/smf/index.php?topic=66315.0

Interesting read on a super-accurate ytterbium clock

http://theconversation.com/sharper-gps-needs-even-more-accurate-atomic-clocks-38109

[anorphirith]

any opinions here on the Audio GD R2R 1 ? 

it seems like a killer deal for the price ? 

http://www.audio-gd.com/R2R/R2R1/R2R1EN.htm

[qqiao]

On 2/26/2017 at 8:43 AM, spritzer said:

I was pulling the MSB out of my system and here is inside one of the dac modules.  The cover was simply soldered in place which the Metcal simply laughed at...   

@spritzer

Just now, qqiao said:

@spritzer

Hi spritzer,

one of my dac module is broken, it looks like transistor q2 is broken. I measured b-c which is shorted. But I don't understand the purpose of this pnp transistor, b-c-e are connected and I don't see any resistor around it. Could you give me some hints? Thank you.