Hi
I have never purchased an oscillator / clock / crystal and any assistance would be greatly appreciated. I need to replace a 50mhz oscillator with 11.2896mhz on an es9023 DAC. I would like to use something with low jitter/high performance but not over priced.
My DAC is a basic es9023 the same as this; AUDIOPHONICS I-Sabre LTE DAC ES9023 Raspberry Pi A+ B+ 2.0 /3.0 I2S - Audiophonics
It has a universal footprint for oscillator placement.
What would you suggest??
Reasons for replacing:
I wish to connect to a Wiimu/Linkplay A31 media streamer module. The A31 operates in slave mode requiring LRLCK and BCLK. It however will only send data via 44.1khz 256 (16 bit; bit clock 2.8224 MHz). The es9023 data sheet which has very little information indicates I would need a clock of 11.2896 MHz.
Similarly, the A31 has been adapted to the ADAU1401 by replacing the oscillator with a 11.2896 MHz however I wish to use my es9023 DAC; I assume it will work?
Wiimu A31 module in combination with ADAU1401/1701 DSP
I have never purchased an oscillator / clock / crystal and any assistance would be greatly appreciated. I need to replace a 50mhz oscillator with 11.2896mhz on an es9023 DAC. I would like to use something with low jitter/high performance but not over priced.
My DAC is a basic es9023 the same as this; AUDIOPHONICS I-Sabre LTE DAC ES9023 Raspberry Pi A+ B+ 2.0 /3.0 I2S - Audiophonics
It has a universal footprint for oscillator placement.
What would you suggest??
Reasons for replacing:
I wish to connect to a Wiimu/Linkplay A31 media streamer module. The A31 operates in slave mode requiring LRLCK and BCLK. It however will only send data via 44.1khz 256 (16 bit; bit clock 2.8224 MHz). The es9023 data sheet which has very little information indicates I would need a clock of 11.2896 MHz.
Similarly, the A31 has been adapted to the ADAU1401 by replacing the oscillator with a 11.2896 MHz however I wish to use my es9023 DAC; I assume it will work?
Wiimu A31 module in combination with ADAU1401/1701 DSP
Last edited:
Not sure if an external clock is suitable but this clock is amazing
The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator
The Well Tempered Master Clock - Building a low phase noise/jitter crystal oscillator
You are recommending a clock that costs more than 20x what his entire dac costs to implement (minimum) and takes up about that many times more space and consumes many times more power. It doesnt have a BNC, or ufl connector for the XO.
Are you really recommending that clock for a 9023 dac? its only got space for one crystal frequency too, so would be run in async mode, with that clock? yeah-nah
well, I suppose not if it will only ever be sent 44.1khz content.
its a great clock clearly, but entirely unsuitable for this purpose IMO.
Are you really recommending that clock for a 9023 dac? its only got space for one crystal frequency too, so would be run in async mode, with that clock? yeah-nah
well, I suppose not if it will only ever be sent 44.1khz content.
its a great clock clearly, but entirely unsuitable for this purpose IMO.
Last edited:
Its unfortunate you can only accept that frequency. the NDK NZ2520SD clocks (which would be my recommendation and only cost $8) only come in at 22.5792MHz and up (aside from the 12MHz). Unfortunately I dont have any good recommendations (that I have used personally) for this price range and that frequency.
Perhaps this TXC is suitable
Perhaps this TXC is suitable
Last edited:
You are recommending a clock that costs more than 20x what his entire dac costs to implement (minimum) and takes up about that many times more space and consumes many times more power. It doesnt have a BNC, or ufl connector for the XO.
Are you really recommending that clock for a 9023 dac? its only got space for one crystal frequency too, so would be run in async mode, with that clock? yeah-nah
well, I suppose not if it will only ever be sent 44.1khz content.
its a great clock clearly, but entirely unsuitable for this purpose IMO.
I hear you loud and clear
was just sharing an idea and obviously not ideal, apologies
Thanks for looking.
I NDK and Cystek seem popular but it couldn't find 11.2896 with the latter either.
SiTime has a number of MEMS oscillators with low phase jitter; namely:
Sit3808:
https://www.sitime.com/products/voltage-controlled-oscillators/sit3808
Sit8208:
https://www.sitime.com/products/lvcmos-oscillators/sit8208
Sit8008:
https://www.sitime.com/products/lvcmos-oscillators/sit8008
I believe they all come in 11.2896 and various package types including something more easily hand soldered.
I'm not sure I'm entering the correct parameters in the parametric search however the sit8208/8209 is used in khadas DACs which measures reasonable well.
That said I can't find anything in stock.
I don't mind waiting but it seems there's no lead time for so many items.
Last edited:
ES9023 has internal ASRC. Its clock does not need to be synchronized to the incoming audio clock. You just feed it some I2S and it does the job.
If you use 11.2 MHz MCLK instead of 50MHz on ES9023, it will underclock the S-D modulator, you will get higher noise, lower performance, etc.
Basically there is no reason to do what you ask.
If the I2S source requires a 11.2MHz clock, you can put one there, no need to mess with the DAC. If it already has a clock and outputs I2S, then job done, just connect that I2S output to ES9023, it will work perfectly fine.
If you use 11.2 MHz MCLK instead of 50MHz on ES9023, it will underclock the S-D modulator, you will get higher noise, lower performance, etc.
Basically there is no reason to do what you ask.
If the I2S source requires a 11.2MHz clock, you can put one there, no need to mess with the DAC. If it already has a clock and outputs I2S, then job done, just connect that I2S output to ES9023, it will work perfectly fine.
I think the reason why a clock cannot be simply added to the a31 (source) is because it acts as a slave module attached to a daughter board with a microcontroller that generates the LRCLK and BCLK. The daughter board has i2s out that can be tied into a DAC but the aim is to use the source without the daughter board which means generating LRCLK and BCLK.
I thought by using a 11.2 oscillator on the es9023 it's internal PLL would generate LRCLK and BCLK which could be fed back to the source a31? I'm not sure if that's possible but you've indicated it would result in poor audio quality.
An adau1401 was used in this thread to generate LRCLK and BCLK. And might explain better than me.
Wiimu A31 module in combination with ADAU1401/1701 DSP
I would like to interface the source a31 directly to a DAC; if that's possible. If not is a DSP 1401 etc the next best solution.
ES9023 doesn't have a master mode, its I2S pins are input only, so it can't generate your bitclock and wordclock. In fact it doesn't have any configuration pins to set modes. It's just plug and play.
Also it doesn't have a PLL, it's a purely digital ASRC.
So... I guess you need something to generate the signals you need. You could use a WM8805, you get a SPDIF input with that, it can act as a master, but you'll have to write some registers via I2C. Or you could use the ADAU board, you get a free tone control with that if you use the DSP. Or use some 74HC dividers... You could even use the existing DAC anyway, don't waste your time trying to have ES9023 work in master mode, it can't.
Why not use whatever board came with the A31 since it already generates those signals?
Also it doesn't have a PLL, it's a purely digital ASRC.
So... I guess you need something to generate the signals you need. You could use a WM8805, you get a SPDIF input with that, it can act as a master, but you'll have to write some registers via I2C. Or you could use the ADAU board, you get a free tone control with that if you use the DSP. Or use some 74HC dividers... You could even use the existing DAC anyway, don't waste your time trying to have ES9023 work in master mode, it can't.
Why not use whatever board came with the A31 since it already generates those signals?
Last edited:
Thank you that makes sense!
You can purchase the wiimu a31 modules separately and usually cheaper than a bundle. The bundled offerings (which I have) include a poor quality DAC, very basic power supply, and I read on another forum the i2s generated onboard was noisy; although no evidence from tests were provided.
However they work flawlessly with streaming services and multiroom audio so there's incentive to build a better DAC, power supply etc
Looks like the ADAU might be the easiest option. Which brings me back to 11.2 clock shortage
Yeah, but the 50M on ES9023 board matters a lot more wrt jitter than the one on the I2S board... since ES9023's ASRC will reject jitter from the I2S clock but not jitter on its own MCLK.
Winner of my $2 clock roundup was Vectron VCC1 but they have since gone unobtainium and price has increased ridiculously.
Winner of my $2 clock roundup was Vectron VCC1 but they have since gone unobtainium and price has increased ridiculously.
These are good but ... they're not SMD...They do fit on the board of the OP though! They perform less good on 3.3V than on 5V.
Last edited:
Unfortunately no, and what's left is just a bunch of notes on the bottom of the hard drive...
The A31 datasheet provides examples for the WM8919/8960 but I'm not sure why they refer to it as 'master mode'. I assumed the codecs would be master in the example? The datasheets for the WM's indicate they can outputs the required clocks.
Wiimu Information Technology A31 Wireless Audio Module User Manual 15 A31 UserMan
Also noticed a few of the Ti DACs might be able to output the required clocks eg PCM5122 / 5121, pcm5252...
I think at least with the pcm512x you don't need onboard MCU to control them. Unfortunately no cheap generic boards available.
Wiimu Information Technology A31 Wireless Audio Module User Manual 15 A31 UserMan
Also noticed a few of the Ti DACs might be able to output the required clocks eg PCM5122 / 5121, pcm5252...
I think at least with the pcm512x you don't need onboard MCU to control them. Unfortunately no cheap generic boards available.
