The Gibbs problem doesn't show up with an 0db sine wave. Try an 0dB square wave. It should look something like this File:Gibbs phenomenon 50.svg - Wikipedia, the free encyclopedia and the tips of the oscillations shouldn't be clipped.
Really this is the fault of the music industry for selling us "music" with 0dB square waves in it. 🙄 With a 24-bit DAC you can easily fix it by digitally attenuating the signal say 6dB: you're probably not going to notice the difference between 23 bits of dynamic range and 24, especially since most 24-bit gear only has about 19 bits of DR to start with.
Usul27, how hard do you think it would be to make a version with just a SPDIF transmitter on it? Could we run it with the same driver, or is it expecting to be able to talk to the DAC over I2C to set it up?
Really this is the fault of the music industry for selling us "music" with 0dB square waves in it. 🙄 With a 24-bit DAC you can easily fix it by digitally attenuating the signal say 6dB: you're probably not going to notice the difference between 23 bits of dynamic range and 24, especially since most 24-bit gear only has about 19 bits of DR to start with.
Usul27, how hard do you think it would be to make a version with just a SPDIF transmitter on it? Could we run it with the same driver, or is it expecting to be able to talk to the DAC over I2C to set it up?
Last edited:
@scopeboy: I'm working on a SPDIF version, but this will use a completely different chipset. The DAC used here does not support SPDIF.
Yeah, I know it doesn't, but you can get a separate chip that converts I2S to SPDIF, such as the WM8804 or CS8406. I'm asking whether your design would still work if the DAC were replaced by one of these chips.
I've had good results by just connecting the WM8804 in parallel with the I2S lines to the DAC. It is smart enough to figure out the MCLK division ratio and sample rate by itself. The question is whether I could do this to your design, and then throw the DAC away. 🙂
I've had good results by just connecting the WM8804 in parallel with the I2S lines to the DAC. It is smart enough to figure out the MCLK division ratio and sample rate by itself. The question is whether I could do this to your design, and then throw the DAC away. 🙂
Last edited:
Yes, a design with the WM8804 is in development. I hope to have the first prototype running in some weeks. It will still need a different driver. Therefore, the digital interface and the DAC will be two separate PCBs. They will only share some basic ideas.
I wonder if this rumor is true: I've heard the wm8804/5 is being discontinued. anyone know if this is true?
it might affect how people design; no point in making boards for parts that are EOL...
it might affect how people design; no point in making boards for parts that are EOL...
Hi Daniel
I am waiting for my first raspberry pi to be delivered and I know: I must have a DAC board from you!
Very good Project, go on!
Greetings from Bern
Franz
I am waiting for my first raspberry pi to be delivered and I know: I must have a DAC board from you!
Very good Project, go on!
Greetings from Bern
Franz
Smile
As long as you offer the DAC for a very reasonable price, you will have the first Swiss customer for sure 🙂
Franz
As long as you offer the DAC for a very reasonable price, you will have the first Swiss customer for sure 🙂
Franz
Sorry to ask if already answered . What is the rms output of the DAC ? I assume about 4 V pk to pk or 1.414 Vrms ? That is about what a 5 V rail might optimistically offer ?
I think, it hasn't been answered before ;-)
The DAC output stage is running with +/- 3.3V rails. It creates the negative power supply internally with a charge pump (frequency depending on sample rate, but at least >800kHz). With this, max Vpp measured with 0db sine waves was 6.2Vpp, 2.2Vrms.
The DAC output stage is running with +/- 3.3V rails. It creates the negative power supply internally with a charge pump (frequency depending on sample rate, but at least >800kHz). With this, max Vpp measured with 0db sine waves was 6.2Vpp, 2.2Vrms.
That's the right side . Any idea what typical maximum output is these days ? We said in the past 250 mV rms input for a standard 2 V DAC . My input is about 400 mV for full output . I think it would be OK ? My valve amp of 600 mV sensitivity seemed to go to clipping on some CD's via a Marantz CD 63 .
I have +/- 6V available . Could it be used ?
I have +/- 6V available . Could it be used ?
Not sure if there is really a "typical" output voltage these days. I've seen lots of different values from 0.5V up to more than 2Vrms.
What do you mean with "could it be used"? You have +/-6V power supply and want to use it for the DAC? Or should this be used for your preamp?
What do you mean with "could it be used"? You have +/-6V power supply and want to use it for the DAC? Or should this be used for your preamp?
I have +/- 6V for part of my pre amp . Just wondering if it would be a better power source for part of the DAC ?As you seem to have taken a lot of trouble over the performance I doubt it would be an improvement .
For this specific DAC is would not work. The DAC always generates the negative power supply internally and is is also only specified up to 3.6V.
RASPBERRY PI WOLFSON AUDIO BOARD
Hi,
This may be a significant event:-
Community: Wolfson Audio Card | element14
It is due shortly (next few days?), I have seen $30 quoted, but there are NO tech specs yet
Patrick
Hi,
This may be a significant event:-
Community: Wolfson Audio Card | element14
It is due shortly (next few days?), I have seen $30 quoted, but there are NO tech specs yet
Patrick
Tech specs are:
- WM5102 CODEC, code named ARIZONA- same as on Samsung Galaxy S4 (GT-I9500 model)
- Audio Hubs (CODECs) | WM5102 | Wolfson Microelectronics
- ASoC driver: http://lxr.free-electrons.com/source/sound/soc/codecs/wm5102.c?v=3.6
Last edited:
Doesn't it have compatible pinout with PCM5142? If so it would be interesting to use it (after swapping the PCM5122 with PCM5142) for digital filtering (with or without RPi), meaning for speaker crossovers and like (room correction EQ maybe). If just two cards could share the same data and clocks it would be possible to implement a 2-way stereo crossover (but if both are master then it isn't possible).
Btw. The distortion on PCM514x can be avoided by switching to Vcom mode (don't know if other PCM51xx have that option).
Last edited:
This would work. However, the DSP on these chips (the 5122 also has the DSP included) is quite limited. For a more powerful version, have a look on my other project:
http://www.diyaudio.com/forums/digital-line-level/245228-dsp-raspberry-pi.html
I think PCM5142 is as powerful as the ADAU1701 part, haven't compared them though but TI data sheet tells "The miniDSP can run up to 1024 instructions on every audio sample at a 48kHz sample rate." That's exactly the same 50MIPS as with AD part, right? Only difference is that audio performance for the DAC is better on the PCM5142. PCM5142 can also be fully programmed so it can do FIR as well as the typical biquads. Also it's got several digital outputs (I2S) so you could add another 5142 or just the 5102, or put just the headers on the board.
Regarding extensibility it would be time to design and agree on a common standard I2S header layout which all DIYers would use for their devices. It should supply I2S signals + master clock and I2C at least.
I didn't notice yet, that the DSP of the 5142 is fully programmable. I will have a look into this. However, this will take some time, as my DSP software is still in development.
Regarding an I2S standard connector. This will be quite difficult, because different people will have different requirements. For this project, the space is limited, therefore I do not want a very large connector. I will think about it an present a proposal in another thread. Let's see, if we find a common standard.
Regarding an I2S standard connector. This will be quite difficult, because different people will have different requirements. For this project, the space is limited, therefore I do not want a very large connector. I will think about it an present a proposal in another thread. Let's see, if we find a common standard.