My DAC for the Raspberry Pi

Jump to Latest
#21
usul27 said:
@smallangryboy: There are 2 eval boards of the prototype, these are already reserved for 2 guys. There might be another 2 eval boards of version 2. The phono sockets will be parts of the kist, however, you do not have to solder them on the board.

@Turbon: This is not a USB DAC, it connects directly to the I2S port of the SoC.
Click to expand...

Ok, rephrase - so we have a solution to get good sound outta the PI? I want one! Sign me up and let me see the bill!

Brgds
 
#22
Any idea what the - 60 dB distortion is ? It seldom is better than 4% . I have been trying to devise a test for this . Perhaps record via an excellent CD burner like TEAC at - 60 db and use it for comparison tests . It should mean something to test the whole chain . An early Denon test disc was of a piano at 0 db - 20 dB and - 40 dB . Doubtless the disc as much as anything at fault . 0 db ( seldom used ) strong and clear . - 20 dB simlar but bleached out detail . - 40 dB like a Kazoo . Denon said construct an attenuator to test at identical levels . I did and didn't have to . From that I conclude pre 1990 CD had a real dynamic range of 20 db at best . Worse still is engineers were told not to go into the 0 to - 20 d range . To an extent dither added 20 dB to that ( 30 % THD became 3 % with dither ) . 4 % THD espeailly if not the higher harmonics at - 60 dB would be excellent . As someone recently said , analog recording fades to noise . Digital if dithered does also . If not one has to ask extra questions as noise is a full register of harmonics if white noise , 7 th and 9 th harmonics are not a great replacement no matter how small . One 24 bit chips maker said " no need for dither " . I wonder ?

Proof that - 20 db was typical is the input sensitive of tape deck inputs works fine with CD . In principle it is about 15 dB too high . Even knowing tape decks peak at + 6 dB doesn't account for it all . Build to the 2 v spec and CD will be too quiet . Not true of a pick up amp built with 6 dB spare gain . It is obviously loud .
 
Last edited:
#24
That's great to know . I will do the chain one day as like vinyl one needs benchmarks . I said to a German guy . " Why do we insist on getting RIAA correct to 0.05 dB when the pick up might be + 7 dB 14 kHz " ? " We get right what we can " was his answer . It sounds right but is mathematical nonsense . The big picture often brings surprises . Doing the detailed study says sometimes why despite evidence one thing sounds different in a way where 0.05dB could not matter . Worse still if 0.03 dB becomes a quest . I get people tell me the NE 5532 is a rubbish op amp . It can be and often is . It is a great op amp with a few problems that need considering . It can go unstable due mostly because it is a high current device . Put in an OP 2604 will often sound better . The reason is it is more tolerant . DAC's might almost be op amps when saying this .
 
#29
hi,

just stopping by to say that i like what you do and your website.

if i were a computer engineer, i d be all over the Pi. What is not to like from such a tiny cheap computer?

Really cool that you successfully tapped from the I2s for the HifiBerry. Makes me want to buy a Pi, upgrade the regulators and try it with my upcoming 1794 dac just to see what quality it outputs...maybe it could replace my hiface,who knows!

keep up the good work!
 
#31
I see you mentioned that the SPDIF output would be limited to 48kHz, why is this? 🙁 If you have I2S going to a DAC, you can just hang a chip like the WM8804 off the same I2S signals and it will convert it to SPDIF. IIRC it should work up to 192kHz.

To make a digital-only version of this board you would probably need nothing but the WM8804.
 
#32
You're talking about the digital output board - right? This is a completely different board. The electrical output on it can be used up to 192kHz, but the TOSLink output is limited to 48kHz. This is a general limit of the optical TOSLink optical transmitters and receivers.
I will start another thread for it when the prototype is available (most likely in december)
 
#35
Hi all,

today i had a little bit time and this was the first attempt to try my HiFiBerry Mini from Daniel (Crazy Audio), about some hours.

After a little mistake from me it plays and play and plays really nice (the P5 con on the Pi). 🙂

What should i say, try it! So looks it for now. Next Step i will change the output connectors. When Daniel is ready for SPDIF, i will integrate it to my Pass D1 Clone! Then we should see again ...
The outputs are not last idea, then i will integrate it to the Pass.

Thanks for your support!
 

Attachments

  • HiFiBerry-Mini-003_kl.jpg
    HiFiBerry-Mini-003_kl.jpg
    112.5 KB · Views: 900
  • HiFiBerry-Mini-004_kl.jpg
    HiFiBerry-Mini-004_kl.jpg
    103.2 KB · Views: 770
#40
@Linuxworks: Not sure were this bug came from. I did a lot of tests also with 0db signals and never had a problem. Did you use the PCM5012A or the older PCM5102?
I really like the sound of this chip. But as always, the sound of different chips is a bit subjective. Some people like one chip more than another. That's ok.

@ajst2duk: Yes, I was also thinking about ESS chip. However I don't really like them. ESS is the only company that does not provide full data sheets on their web site. There are also harder to source than chips from other companies. I might have a look at it in the future, but that's not my first priority.
 
You must log in or register to reply here.
Top Bottom