I just finished a new build of an RPI+Kali+Volumio feeding a DDDAC via I2S. It sort of works, but I get a high-pitched noise / hiss at the DAC output. It's not very loud, but it's bad enough to screw up silent music. From the tests I did, it looks like it has to do either with the interfacing of the RPI+Kali to the DDDAC via I2S, or it's happening in the RPI+Kali somewhere (see my other detailed post in the DDDAC thread). Can anybody help?
I had the same experience this weekend. I put my Kali in the I2S line between my hifiberry dac+pro I2S out and the DAC. Without Kali its dead silent even with no shielded wires. With kali in place I also note a hizz noise when pausing music or very slightly on silent music parts..I tried several wires and power feeds from external, feeding raspi from Kali or not, connecting just pin39 as ground or the closest of each i2s lines for coax shields..finally all I2S wires are coax with ground connected at ends. Even no diff. to unshielded short wires. I had the impression it became slightly better changing from Rasp3 to 2 without Bluetooth. And a Alu-board below the DAC -Kali combi (I attached the Kali with a std. printboard to the DAC, but there are wires at in and out of Kali) made it better, but a steel metal cover over the kali made it even worse..Any help would be fine ( how to shield kali best)?
Interesting! How and where did you connect the GND of the RPI?
I found that when I set the RPI/Volumio to upsample the audio stream to something higher than the standard 44.1/48 kHz rate the noise/hiss goes away during audio playback. Is it possible the I2S signal interferes with the analog signal somehow (which would not be audible if the I2S signal is beyond audible frequency)?
I did not have any such problems with the Kali running from a WaveIO board instead of the RPI. The main difference I can see is that the WaveIO provides "isolated" I2S output, whereas the RPI does not.
The RASPI is not grounded at all. The PSU board that feeds the Raspi (and the KALI) is a sigma11 board from amb. it is just connected to the transformer with 2 lines.
Yesterday I was not able to start the Kali.. usually it starts blinking like crazy with blue LED when powering up, but I could not get it work this time. I tried at least 20 times in different orders, once with I2S input signal already active or not - no way. I checked it on my Raspi piano setup and it powered from the first trial..
Nevertheless the hiss was also audible in this status and it could be reduced when I connected the Minus of sigma11 output to the Minus of the DAC digital side psu - as recommanded in the other thread you linked above.
KALI seems to be quite sensitive to its environment..
Yesterday I was not able to start the Kali.. usually it starts blinking like crazy with blue LED when powering up, but I could not get it work this time. I tried at least 20 times in different orders, once with I2S input signal already active or not - no way. I checked it on my Raspi piano setup and it powered from the first trial..
Nevertheless the hiss was also audible in this status and it could be reduced when I connected the Minus of sigma11 output to the Minus of the DAC digital side psu - as recommanded in the other thread you linked above.
KALI seems to be quite sensitive to its environment..
We changed again the PCB (this is version 1.6 I think) , found more problems on power for analog stage (we need 2.5W)
Except for that, we also cannot wait for samples to be delivered and tested. We designed this DAC to sound (very) good, so you can imagine we are also eager to listen to it.
Anyway our full team is on it. Will give an update when new PCB design will be sent to the manufacturing house.
Except for that, we also cannot wait for samples to be delivered and tested. We designed this DAC to sound (very) good, so you can imagine we are also eager to listen to it.
Anyway our full team is on it. Will give an update when new PCB design will be sent to the manufacturing house.
What about the GND connections of the I2S interface? Did you connect them between Kali and DDDAC? All of them? Just one?
on In and out side each I2S wire shield is connected to the closest GND pin on Kali GPIO. The other Ends of the I2S wire shields are connected to the GND Pins of the I2S out of RASPI (Hifiberry dac+pro) and on the other side to the I2S GND PIN of DAC board Input. No seperate GND wires (the shields are used as GND wires).
But I also tested with unshielded wires and separate GND lines - makes no difference.
But I also tested with unshielded wires and separate GND lines - makes no difference.
Exactly like my current setup. The RPI/Kali combo is grounded to the DDDAC GND via the I2S GND wires. Is this not good enough? Does this inject some of the I2S signal into the audio GND, making the audio noisy?
I'll try different groundibg schemes once I find some time.
Interesting to see Codaca inductor in place.Will there be PFFB and an external SE2Diff stage implemented as well? From our measurement, we have seen that using an external SE2Diff stage (even without PFFB) gives an advantage of nearly 20dB regarding k1/k2 and k1/k3 margin. PFFB then pulls another 10-20dB better performance. With an overall system gain of 32dB THD+N hits 0.0077% while pure THD is at 0.0022%. At 1kHz k1/k2 is 108dB, k1/k3 is 120dB. These values are for 10W into 10R.
Not bad , but I think pursuit of THD+N only might not be the right approach. Dont forget that those numbers while very important, are just a part of sound quality . I Still suggest to read all literature on analog and improve based on it.
noise
the problem was not the kali or dac but the bluetooth and and wifi transmitter on the RPI 3, in my opinion the RPI 3 is not suitable for audio.
what ik did is covered the wifi bluetoothtraqnsmitter of teh RPI 3 with a piece of Lead and did wrap all my I2S wires with copper foil and now it is dead quiet.
But i can tell you my dac is sounding far more better and musical without the kali with a old RPI B and is also a lot more responsive. often audio is about inproving things and specs, i can tell you less is more, even a I2S isolator is adding jitter to you`re signal.
just use good liniar power supply`s, sounds far more better and gives more inproment then buffers/isolators and other things added in the signal path. i am using a liniar power supply for my RPI and 2 others for my dac one for the dac and a Second one for the I/V stage and teh resuls are better then adding a kali
I had the same with my own build TDA 1543 Nos dac with 2sk170 fet i/v.
the problem was not the kali or dac but the bluetooth and and wifi transmitter on the RPI 3, in my opinion the RPI 3 is not suitable for audio.
what ik did is covered the wifi bluetoothtraqnsmitter of teh RPI 3 with a piece of Lead and did wrap all my I2S wires with copper foil and now it is dead quiet.
But i can tell you my dac is sounding far more better and musical without the kali with a old RPI B and is also a lot more responsive. often audio is about inproving things and specs, i can tell you less is more, even a I2S isolator is adding jitter to you`re signal.
just use good liniar power supply`s, sounds far more better and gives more inproment then buffers/isolators and other things added in the signal path. i am using a liniar power supply for my RPI and 2 others for my dac one for the dac and a Second one for the I/V stage and teh resuls are better then adding a kali
I also noted improvement from Raspi3 -> Raspi2 (without BT), especially on one set-up where I even had drop outs of signal before shielding the I2S wires with AL-film to protect them from BT (I guess this was the reason).
Did you put the CU film to GND (if yes, on on which side?)?
Nevertheless the power-up issue with Kali and hiss noise is still there with Raspi2 and with shielded coax wires for each i2S line I would like to know where it is coming from.
Did you put the CU film to GND (if yes, on on which side?)?
Nevertheless the power-up issue with Kali and hiss noise is still there with Raspi2 and with shielded coax wires for each i2S line I would like to know where it is coming from.
Last edited: