Good morning Miro,
would a board with both, USB and SpdiF, switchable by hand or auto, be an option. Could save a lot of i2s cable.
+1 to this all in one and switchable
Congrats Miro for this next episode season 2 of the GoT DAC
The little DC traffo spliter sucks a lot of information. I know it's in all on the datasheet, but if one has a RCA with a dc blocking cap, it worths to try to remove it : the sounding difference is huge according the experience I made on an old AD1865 DAC. Also true AESBU with a simple coax TV cable all at the right impedance sounds much better than a RCA, but it means the streaming device to have one with the according circuitry (lesson learned from Jean-Paul and his fellow cool little ESS 9023 DAC which is sounding very correct for a Sigma Delta)
For R7 to R10, resistor choice is critical and not always mandatory, here again if no bouncing, a transparency sucker. I would avoid all not goo enough resistor here as the too much often seen in this thread Dale RN50 (the little brown one) : prefer through hole Yageo 1% precison : much better noise floor ; if SMD that are easy enough to solder : thin film from SUSSUMU has passed subjective listening test here as measurements : it's also easier to remove for zeo ohms than a THT with vias than are not easy here for tweaking (though boards are cheap at the famous JPLCB, and take not golded traces : it 's sound better
You rocks Miro !
The little DC traffo spliter sucks a lot of information. I know it's in all on the datasheet, but if one has a RCA with a dc blocking cap, it worths to try to remove it : the sounding difference is huge according the experience I made on an old AD1865 DAC. Also true AESBU with a simple coax TV cable all at the right impedance sounds much better than a RCA, but it means the streaming device to have one with the according circuitry (lesson learned from Jean-Paul and his fellow cool little ESS 9023 DAC which is sounding very correct for a Sigma Delta)
For R7 to R10, resistor choice is critical and not always mandatory, here again if no bouncing, a transparency sucker. I would avoid all not goo enough resistor here as the too much often seen in this thread Dale RN50 (the little brown one) : prefer through hole Yageo 1% precison : much better noise floor ; if SMD that are easy enough to solder : thin film from SUSSUMU has passed subjective listening test here as measurements : it's also easier to remove for zeo ohms than a THT with vias than are not easy here for tweaking (though boards are cheap at the famous JPLCB, and take not golded traces : it 's sound better
You rocks Miro !
Last edited:
There are a lot of good chips for this purpose : they all works better async : i.e. when the chip is slaved by a local crystal. Here the quality of the crystal still matters : NDK, Fox, Kyocera, etc...
I would populate all the board with Panasonic FC cap or the violett metal can SFP.
Edit, I don't know if a good isolator chip the size the shift register we already soldered exists ? Maybe it will be better than a not perfect Nenawa traffo splitter. the isolator chips for digital circuitry are nowadays very few jittered, less better than 5 or 10 years ago...
Well I assume as usual, proof is in the pudding, while it deserves or not icing has to be tested for real !
Last edited:
It is just an ISO7640FM isolator IC. One side of the IC is the "dirty" input side, the other is the "clean" output side. You need a good quality 3.3V or 5V supply to the clean side.
Attachments
I bought both type. I think maybe 6N23P with Voskhod (rocket logo) is more warm, and fuller bass, and the Reflector 6N23P-EV (or EB) is as I know "Military" type have less coloration.
I preferred Voskhod type.
What is the recommended capacitor type in oap stage? Nichicon FG is better in this position than the Oscon?
USB/SPDIF/Optical to I2S
Simple USB/SPDIF/Optical to I2S
USB up to 16-B/48kHz.
SPDIF/Optical up to 24-B/192kHz.
Jitter is 50ps, so it can be also used with modern delta-sigma DACs.
Transfer is asynchronous (clocks are generated directly in the WM8804).
It is almost THT except for the WM8804 and PCM2706, which are soldered from the bottom.
How to solder the smd IC: youtube IC soldering
Untested at the time of writing this post, so i can't guarantee the functionality right away
(The first post will be updated after testing.)
Components example, mouser:
IC4:595-PCM2706CPJTR
IC1:238-WM8804GEDS
R7, R8, R9, R10, R13, R14:594-MBB02070C2209FCT
R2, R4:594-MBB02070C7509FCT
R12, R15:594-MBB02070C1501FC1
R1, R3, R5, R6:594-MBB02070C1002FC1
R11:594-B0207C1M000F5T
Q1, Q2:520-ECS120-18-4X-CKM
C13, C14, C16, C17:81-RCE5C2A180J0A2H3B
C20, C21:81-RCER71H104K0DBH3A
C1, C3, C5, C7, C9, C10, C11:505-MKP2D031001FJO00
C18, C19, C22:810-FA16X8R1H105KRU6
C12:505-MKS21.0/63/5
C15:667-EEU-FR1H100B
C23, C24, C25:667-EEU-FS1H270
C2, C4, C6, C8:667-EEU-FR1E101
TR1:580-786013C
IC2, IC3:511-LF33ABV-DG
JP1, JP2, JP4, JP5, X5:571-5-146284-2
JP3:571-5-146285-3
X4:571-5-146268-4
X3:502-PJRAN1X1U02AUX
X2:GP1FAV50RK0F Original and New SHARP Sensor | eBay
X6:649-61729-1011BLF
X1:651-1729128
(JP1, JP2, JP3/5, JP4):855-M7582-05
Simple USB/SPDIF/Optical to I2S
USB up to 16-B/48kHz.
SPDIF/Optical up to 24-B/192kHz.
Jitter is 50ps, so it can be also used with modern delta-sigma DACs.
Transfer is asynchronous (clocks are generated directly in the WM8804).
It is almost THT except for the WM8804 and PCM2706, which are soldered from the bottom.
How to solder the smd IC: youtube IC soldering
Untested at the time of writing this post, so i can't guarantee the functionality right away
(The first post will be updated after testing.)
Components example, mouser:
IC4:595-PCM2706CPJTR
IC1:238-WM8804GEDS
R7, R8, R9, R10, R13, R14:594-MBB02070C2209FCT
R2, R4:594-MBB02070C7509FCT
R12, R15:594-MBB02070C1501FC1
R1, R3, R5, R6:594-MBB02070C1002FC1
R11:594-B0207C1M000F5T
Q1, Q2:520-ECS120-18-4X-CKM
C13, C14, C16, C17:81-RCE5C2A180J0A2H3B
C20, C21:81-RCER71H104K0DBH3A
C1, C3, C5, C7, C9, C10, C11:505-MKP2D031001FJO00
C18, C19, C22:810-FA16X8R1H105KRU6
C12:505-MKS21.0/63/5
C15:667-EEU-FR1H100B
C23, C24, C25:667-EEU-FS1H270
C2, C4, C6, C8:667-EEU-FR1E101
TR1:580-786013C
IC2, IC3:511-LF33ABV-DG
JP1, JP2, JP4, JP5, X5:571-5-146284-2
JP3:571-5-146285-3
X4:571-5-146268-4
X3:502-PJRAN1X1U02AUX
X2:GP1FAV50RK0F Original and New SHARP Sensor | eBay
X6:649-61729-1011BLF
X1:651-1729128
(JP1, JP2, JP3/5, JP4):855-M7582-05
Attachments
Thanks, I will try the EB then.
For the caps : I very do NOT like the Nichicon FG. The Oscon gives often a coldy mat sounds on the strange side in analog stages I found.
Then caps also depends on your systems, the chosen oaps and listening tests have to be made. It is important to respect the leads spacing of the cap of the design.
Very different character, but all good :
Pan FC, sometimes FM in small values (both gives a different result, the Fm is the less good imho and sometimes have a fatiguing mid-treble, sort of smoothed harshness, depends on how it is used as well)
Nichicon KG
Nichicon KZ, though the lead spacing will not fit
There is also an official BOM page 1.
Last edited:
WM8804/5 or DIR9001 have lower jitter (good internal PLL, only 50ps jitter), consider it
are they genuine chip anf if on a board a good pcb layout ? You never know with AlieExpress. The only two AD1862 I bought there had one with a broken leg that could not be fixed... monney lost, so it's always a bet. I'm fed up with the no quality mentality of China, we know where it has driven us...
ahaha RED, nah it's good if it has the X4 conector still (in my head both should be on board for everybody , the uf-l could be on the top side, with traces towards X4 till on the right side of the pcb for the stacking with the main DAC !
The stacking X4 connector for overlapping on the main dac is a great idea ! Imagine a Miro board with the UTFL96 streamer board from ECDESIGN via Toslink, with Diyiggy parts BOM and the output stage of your choice (Fran-Wood) : it will beat several USD grand SIGMADELTA AK and ESS, hey ! SOme can also uses sota SD Trans via UF-L; a NAS via USB, a toslink towards soundcards of computers, eetc VERSATILITY
, the uf-l could be on the top side, with traces towards X4 till on the right side of the pcb for the stacking with the main DAC !The stacking X4 connector for overlapping on the main dac is a great idea ! Imagine a Miro board with the UTFL96 streamer board from ECDESIGN via Toslink, with Diyiggy parts BOM and the output stage of your choice (Fran-Wood) : it will beat several USD grand SIGMADELTA AK and ESS, hey ! SOme can also uses sota SD Trans via UF-L; a NAS via USB, a toslink towards soundcards of computers, eetc VERSATILITY
Last edited: