Geek Pulse: Geek desktop DAC/AMP by Light Harmonics
Jun 21, 2015 at 10:50 AM Post #8,087 of 13,800
32/768 PCM with R-2R? I thought that's not possible with the R-2R architecture and that the maximum is technically true 21-bit. Getting precision resistors for true 32-bit audio would be quite a feat....and extremely expensive!


Why would you only be able to go up to 21-bit? Can't you just make the ladder as long as you want?
 
Jun 21, 2015 at 11:29 AM Post #8,088 of 13,800
If you can list just one true 24-bit R-2R DAC outside of LH, I'd be surprised. XD

Getting enough precision for 24-bits in an affordable price range would be quite a feat. The Yggdrasil is only 21-bits for example and they're using a DAC chip that's only ever been used in military and medical applications.
 
Jun 21, 2015 at 11:48 AM Post #8,089 of 13,800
If you can list just one true 24-bit R-2R DAC outside of LH, I'd be surprised. XD

Getting enough precision for 24-bits in an affordable price range would be quite a feat. The Yggdrasil is only 21-bits for example and they're using a DAC chip that's only ever been used in military and medical applications.

 
 
https://hifiduino.wordpress.com/2014/10/12/r2r-for-the-rest-of-us/
 
You're welcome
 
Jun 21, 2015 at 12:13 PM Post #8,090 of 13,800
If you can list just one true 24-bit R-2R DAC outside of LH, I'd be surprised. XD

Getting enough precision for 24-bits in an affordable price range would be quite a feat. The Yggdrasil is only 21-bits for example and they're using a DAC chip that's only ever been used in military and medical applications.


Hmmmmm
 
Jun 21, 2015 at 12:17 PM Post #8,091 of 13,800
If you can list just one true 24-bit R-2R DAC outside of LH, I'd be surprised. XD

Getting enough precision for 24-bits in an affordable price range would be quite a feat. The Yggdrasil is only 21-bits for example and they're using a DAC chip that's only ever been used in military and medical applications.


Just because nobody makes a 32-bit r2r dac doesn't mean it is not possible. As far as I understand it, there are no technical limitations to making 32-bit ladder.
 
Jun 21, 2015 at 12:18 PM Post #8,092 of 13,800
I did some listening tests today with my brand new Pulse Sfi.
 
I compared
DX90->Coaxial Out to SPDIF1 Pulse Sfi -> Shure SE846
Dx90 -> Shure SE846
 
Since the Headphone out and the Coaxial out of the DX90 can run simultaneously, volume level matching was easy.
I was quiet shocked by the result, it sounded the same, I have difficulties to make out a difference from first impression.
 
On the other hand, the Geek Out sounds different then the DX90. Could this be due to missing burn in? Am I hallucinating? Is the DX90 that good? :)
 
Jun 21, 2015 at 1:27 PM Post #8,095 of 13,800
Just because nobody makes a 32-bit r2r dac doesn't mean it is not possible. As far as I understand it, there are no technical limitations to making 32-bit ladder.

THD is one of the major limitations of R-2R DACs because precision resistors, even 0.01% tolerance ones are erroneous after 20-bits.






Except the PCM1704 is a 24-bit DAC, not a 32-bit one. It just has a 32-bit interface, which isn't the DAC. The DAC chips themselves are only 23-bit.
http://www.mouser.com/ds/2/405/pcm1704-442330.pdf
 
Jun 21, 2015 at 2:08 PM Post #8,096 of 13,800
Sorry dude but the PCM1704 is a 24 bit DAC chip. The Audio GD may accept 32 bit input and may do 32 bit signal processing but the D/A converter is still only 24 bit.

J.P.

Except the PCM1704 is a 24-bit DAC, not a 32-bit one. It just has a 32-bit interface, which isn't the DAC. The DAC chips themselves are only 23-bit.
http://www.mouser.com/ds/2/405/pcm1704-442330.pdf

 
Huh, I stand corrected. I thought the PCM1704 was 32-bit but looking at the info from TI it isn't. Carry on.
 
http://www.ti.com/lit/ds/symlink/pcm1704.pdf
 
Jun 21, 2015 at 2:50 PM Post #8,097 of 13,800
I did some listening tests today with my brand new Pulse Sfi.

I compared
DX90->Coaxial Out to SPDIF1 Pulse Sfi -> Shure SE846
Dx90 -> Shure SE846

Since the Headphone out and the Coaxial out of the DX90 can run simultaneously, volume level matching was easy.
I was quiet shocked by the result, it sounded the same, I have difficulties to make out a difference from first impression.

On the other hand, the Geek Out sounds different then the DX90. Could this be due to missing burn in? Am I hallucinating? Is the DX90 that good? :)

How you like the Sfi as a dac only? Compared to other dacs?
Thanks.
 
Jun 22, 2015 at 12:00 AM Post #8,098 of 13,800
I thought the DX90 was good, but my Pono sounds better. I'm not sure the "line out" is really line out, if I remember you can still change the volume when using that jack. So you may just be double among, hearing the DX90 both times...
 
Jun 22, 2015 at 5:05 AM Post #8,099 of 13,800
I thought the DX90 was good, but my Pono sounds better. I'm not sure the "line out" is really line out, if I remember you can still change the volume when using that jack. So you may just be double among, hearing the DX90 both times...

 
But he is not using line out, he is using the digital out. It may be the case that the digital out is not bitperfect and there is dsp applied to the digital out. 
 
One thing that is annoying is we cannot see the samplng rate for tbe spdif input on the geeks.
 
Jun 22, 2015 at 5:37 AM Post #8,100 of 13,800
Burn it in.  When I opened mine. Right out its seemed a bit like my GO. With little bass. 
 
Now its night and day in comparison.
 

Users who are viewing this thread

Back
Top