Open-source USB interface: Audio Widget
- Thread starter borges
- Start date
Hi Gioz,
Here is an example of someone who has mated it with a Buffalo DAC. He has used the one with XOs mounted on the board. I don't know if the Buffalo is able to produce 11.2896 / 12.288MHz clocks.
Innmat.. Hva du vil!!! - Side 82
Børge
Here is an example of someone who has mated it with a Buffalo DAC. He has used the one with XOs mounted on the board. I don't know if the Buffalo is able to produce 11.2896 / 12.288MHz clocks.
Innmat.. Hva du vil!!! - Side 82
Børge
I thought that that board could be used as-is without any modification for DAC with ASRC input such as the stock BII/BIII. Isn't it like that?
doesn't the USB-I2S module (version with on-board clocks pre-mounted) have its own divider & clock switch logic on board?
unfortunately it does require registration to see attachments/images.
Last edited:
It does, yes. The divider is there. The switch logic is implemented by enabling one XO at the time and shorting their outputs.
So it seems, yes. I don't own a Buffalo DAC. There is some talk about an 80MHz clock. If that is the case, and the MCLK from the XO-equipped module is not used, the DAC chips are operating asynchronous to the sample rate of the music. All my experience is with synchronous systems, but that doesn't mean async is bad.
Børge
yes. IIRC newer versions have 100MHz clock. According to many (including the chip designer), the higher the clock the better it works (and sounds). This is true also in synchronous mode: for best performance you'd need two ~90MHz clocks.
(and it is true for all Sabre DACs, including the ES9023. Unfortunately, the ~20MHz clock of the AB-1.1 gives a bad performance penalty, specially with source material at higher sample rates).
You can start here and here. Consider this and that from Guilio; he might be able to help further.
Please keep us posted and take some pictures of what you do. Also I for one am very interested in the effects of your capacitor changes, how would you describe the effects on the sound?
if you have set a non-default number of post per page those links will not work. If you have this problem, try removing the page number from the URL, like this:
post2840845
post2841397
post2958268
post2960104
(the system will redirect you automatically to the right page).
http://www.diyaudio.com/forums/digi...e-usb-interface-audio-widget.html#post2960104
Finally, I'm able to test the AK4430 properly. It's running through my Graham Slee HP amp and a pair of AKG K701s. I'm using ASIO+Win7-64+JRiver, source is a good Vivaldi CD.
Compared to the ES9023 I think it packs more punch in the bass. The high notes sound kinder.
That may be due to the DAC, or it may be due to the caps. I've no more OS-CONs left, but I did put in the 10µF EEF-CD0J100R at pretty much every point of load I could think about.
I tested some tones with Realtime Signal Generator for windows UAC1, 44.1, 441Hz square wave, attenuation variable down to 0dB.
A full-scale square wave has dV~=4.6V between the lower DC level and upper DC level. This correspons to DC gain. But the DAC has more headroom, allowing the overshoot of the digital filter to run its course.
The ES9023, on the other hand, has dV~=4.8V between DC levels of the square wave. But it overflows heavily due to what I believe is the DAC's digital input range being too small for the internal digital filters's digital output range.
So at this moment the AK4430 clearly leads by a better design. And if it isn't due to the caps, it leads with more punch as well.
Cheers,
Børge
Compared to the ES9023 I think it packs more punch in the bass. The high notes sound kinder.
That may be due to the DAC, or it may be due to the caps. I've no more OS-CONs left, but I did put in the 10µF EEF-CD0J100R at pretty much every point of load I could think about.
I tested some tones with Realtime Signal Generator for windows UAC1, 44.1, 441Hz square wave, attenuation variable down to 0dB.
A full-scale square wave has dV~=4.6V between the lower DC level and upper DC level. This correspons to DC gain. But the DAC has more headroom, allowing the overshoot of the digital filter to run its course.
The ES9023, on the other hand, has dV~=4.8V between DC levels of the square wave. But it overflows heavily due to what I believe is the DAC's digital input range being too small for the internal digital filters's digital output range.
So at this moment the AK4430 clearly leads by a better design. And if it isn't due to the caps, it leads with more punch as well.
Cheers,
Børge
Last edited:
Hi,
Just as I was asking about my little issue with my Imac... has any body got to try this DAC (QNKTC) with a windows based tablet? There are some tablets like the EXOPC or the wintab which have 2 usb inputs. That would be interesting because you could switch at the same time a DAC and a hard drive with all the FLAC's. Also because there is a big issue with this android 4 and usb audio which seems not to go any better...
Just as I was asking about my little issue with my Imac... has any body got to try this DAC (QNKTC) with a windows based tablet? There are some tablets like the EXOPC or the wintab which have 2 usb inputs. That would be interesting because you could switch at the same time a DAC and a hard drive with all the FLAC's. Also because there is a big issue with this android 4 and usb audio which seems not to go any better...
I know nothing about Android and USB audio. Care to enlighten us? Perhaps the issue we suspect in the Linux kernel is holding things back?
Are there other USB devices which act strangely on that tablet? which OS do you use? How is the cpu load? Is it performing better with Nik's ASIO driver and UAC2? It's hard to say if what you are experiencing is in the tablet or the DAC.
Børge
Are there other USB devices which act strangely on that tablet? which OS do you use? How is the cpu load? Is it performing better with Nik's ASIO driver and UAC2? It's hard to say if what you are experiencing is in the tablet or the DAC.
Børge
Thank you ElEsido. I ended cutting the USB cable since i read your message too late but is working fine so far.
I made a external source with lt1086 , some Mundorf caps and i can tell you , changing the original os-cons got me a lot of details in sound + bass extension, but now after external source a whole new world has opened full of micro-details, texture and flavor.
Slightly off topic, open source Bach:
The Open Goldberg Variations, by Kimiko Ishizaka | Open Goldberg Variations
Looking forward to playing this on some proper equipment at home. I'm on the road with work right now with not-so-cool speakers in the laptop.
Børge
The Open Goldberg Variations, by Kimiko Ishizaka | Open Goldberg Variations
Looking forward to playing this on some proper equipment at home. I'm on the road with work right now with not-so-cool speakers in the laptop.
Børge
That looks like a cool chip! With one we have I2S and MCLK. In addition a few GPIO lines are needed, but they should be able to go over cheaper and slower isolators.
The GPIO lines I think of are:
- Filter select
- VBUS_EN
- MCLK_P48_N441
- UART TX & RX (it's nice to have some comms link)
- Maybe an interrupt line
Any more lines needed? I imagine making a go-between board between the USB-I2S module and the analog board where the bare minimum of IO lines are isolated and power is inserted.
Børge
The GPIO lines I think of are:
- Filter select
- VBUS_EN
- MCLK_P48_N441
- UART TX & RX (it's nice to have some comms link)
- Maybe an interrupt line
Any more lines needed? I imagine making a go-between board between the USB-I2S module and the analog board where the bare minimum of IO lines are isolated and power is inserted.
Børge
Help with firmware programming
Hi everybody, I need some help with my AB 1.1.
Today I flashed the latest firmware, and it was all OK, working.
Them I tried again the same operation .... and since then I'm stuck in this situation:
Running batchisp 1.2.5 on Wed May 30 12:44:55 2012
AT32UC3A3256 - USB - USB/DFU
Device selection....................... PASS
Hardware selection..................... PASS
Opening port........................... PASS
Reading Bootloader version............. PASS 1.0.3
Erasing................................ FAIL Could not read from USB device.
(A)bort, (R)etry, (I)gnore ?
Looks like it's impossible to erase the chip.
I found something about such problems on the net, and tried some batchisp
command that demonstrates that the chip is alive, with 1.0.3 bootloader ... but it's impossible to reprogram.
Now the chip is constantly in DFU mode, with led OFF. When I launch the usual batchisp command (that includes a reset), the result is as above, and the LED becomes red ... but the board is actually unrecognized.
If i hit reset, or unplug, I go back to the initial situation (led OFF, board in DFU mode ). What can I do?
Some help would be appreciated!
Hi everybody, I need some help with my AB 1.1.
Today I flashed the latest firmware, and it was all OK, working.
Them I tried again the same operation .... and since then I'm stuck in this situation:
Running batchisp 1.2.5 on Wed May 30 12:44:55 2012
AT32UC3A3256 - USB - USB/DFU
Device selection....................... PASS
Hardware selection..................... PASS
Opening port........................... PASS
Reading Bootloader version............. PASS 1.0.3
Erasing................................ FAIL Could not read from USB device.
(A)bort, (R)etry, (I)gnore ?
Looks like it's impossible to erase the chip.
I found something about such problems on the net, and tried some batchisp
command that demonstrates that the chip is alive, with 1.0.3 bootloader ... but it's impossible to reprogram.
Now the chip is constantly in DFU mode, with led OFF. When I launch the usual batchisp command (that includes a reset), the result is as above, and the LED becomes red ... but the board is actually unrecognized.
If i hit reset, or unplug, I go back to the initial situation (led OFF, board in DFU mode ). What can I do?
Some help would be appreciated!
SOLVED!
I have solved the issue. My compiler (due to wrong switches) was generating a 73 MB elf file ... batchisp was probably writing over wrong portions of the memory.
I solved by removing the "erase" part from the command for the first time ... without erasing I was able to reprogram the chip. Then also the complete command with "erase" was working again.
I have solved the issue. My compiler (due to wrong switches) was generating a 73 MB elf file ... batchisp was probably writing over wrong portions of the memory.
I solved by removing the "erase" part from the command for the first time ... without erasing I was able to reprogram the chip. Then also the complete command with "erase" was working again.