Laptop on battery still allows the digital noise generated by the laptop itself to be transmitted across.
How about powering the isolated parts with battery, and charging the battery with USB power when the unit is not playing music?
Alex
reminds me of an old camelot DAC I bought in the 90's. it had 2 6v lead acid cells that would charge during 'rehearsal' (switch setting) and be the main power source during 'performance' (lol, that's the 2 words they used on that toggle switch).
lots of extra room needed in the chassis, charging circuit costs, batteries that need replacing, shipping weight. blech! not a good return for what you have to do to get it all working.
lots of extra room needed in the chassis, charging circuit costs, batteries that need replacing, shipping weight. blech! not a good return for what you have to do to get it all working.
HID support
Hi guys,
I could need some help with USB descriptors in the latest experimental firmware.
I have been toying a bit with HID support lately. Changes are checked into the audio-widget-experimental branch. All my code changes are marked "BSB" and the date for easy grepping.
I've got HID working on UAC1. It works by sending a simple ASCII formatted HEX command over UART to the audio-widget. This was quite straightforward with the advice Alex gave me.
However, on UAC2 both HID and audio are broken. I know very little about USB descriptors. Could you give a hand and help find out what is wrong here:
widget_awx-bsb_20120720.elf
Don't worry about breaking your audio-widget. Functional firmware is here:
audio-widget_20120721.elf
Cheers,
Børge
Hi guys,
I could need some help with USB descriptors in the latest experimental firmware.
I have been toying a bit with HID support lately. Changes are checked into the audio-widget-experimental branch. All my code changes are marked "BSB" and the date for easy grepping.
I've got HID working on UAC1. It works by sending a simple ASCII formatted HEX command over UART to the audio-widget. This was quite straightforward with the advice Alex gave me.
However, on UAC2 both HID and audio are broken. I know very little about USB descriptors. Could you give a hand and help find out what is wrong here:
widget_awx-bsb_20120720.elf
Don't worry about breaking your audio-widget. Functional firmware is here:
audio-widget_20120721.elf
Cheers,
Børge
Last edited:
One of our ES9022 users told me that this excellent project also uses the ES9022/23.
And that you might be interested in this :
http://www.diyaudio.com/forums/group-buys/216669-joachim-gerhard-filter-buffer-es9022.html
It is a nice addition to the 9022/23 from Joachim Gerhard.
And he is generous enough to make it available to DIYA subscribers.
Best regards,
Patrick
And that you might be interested in this :
http://www.diyaudio.com/forums/group-buys/216669-joachim-gerhard-filter-buffer-es9022.html
It is a nice addition to the 9022/23 from Joachim Gerhard.
And he is generous enough to make it available to DIYA subscribers.
Best regards,
Patrick
Inductors in output filters can work very well. They are not well understood by audiophiles so its uncharted territory.
Unfortunately the audio widget doesn't have a negative supply to the buffer would require an external supply. The problem can escalate quickly as the external supply brings more connections to the table. If you were feeling adventurous you could steal the negative supply from the charge pump. It might work if the draw is not too great and the circuit would work on +/- 4V approx.
A passive version with similar characteristics could be built pretty easily. I prefer inductor input for several reasons. Finding suitable inductors is the hard part. The chip doesn't have a lot of drive capability so dropping the load impedance to reduce the inductor size is not an option either.
The link talks about the HF output above the audio band. If my audio widget weren't on the other side of the world I would look at that. I hope Borge has the next run ready soon, I need several. They will be using the AK4430 which I found sounded a lot better.
Unfortunately the audio widget doesn't have a negative supply to the buffer would require an external supply. The problem can escalate quickly as the external supply brings more connections to the table. If you were feeling adventurous you could steal the negative supply from the charge pump. It might work if the draw is not too great and the circuit would work on +/- 4V approx.
A passive version with similar characteristics could be built pretty easily. I prefer inductor input for several reasons. Finding suitable inductors is the hard part. The chip doesn't have a lot of drive capability so dropping the load impedance to reduce the inductor size is not an option either.
The link talks about the HF output above the audio band. If my audio widget weren't on the other side of the world I would look at that. I hope Borge has the next run ready soon, I need several. They will be using the AK4430 which I found sounded a lot better.
Things have taken a while with the next batch including AK4430. I did receive pictures of the board testing, and the boards were shipped from the manufacturer today. I'm pretty eager to get my hands on them and test them myself.
Demian, when I tried the AB-1.13 with AK4430 I really noticed a lot more punch in the bass. That may be due to the DAC chip, and to the 10µF Panasonic chip caps that I put in instead of the OS-CONs. Do you have any similar experience?
Børge
Demian, when I tried the AB-1.13 with AK4430 I really noticed a lot more punch in the bass. That may be due to the DAC chip, and to the 10µF Panasonic chip caps that I put in instead of the OS-CONs. Do you have any similar experience?
Børge
Last edited:
In my ES9022 modules, I have a huge plastic polymer cap, the legs of which were soldered directly at the power supply pins of the DAC.
No lack of punch whatsoever.
So IMHO more to do with layout than DAC I think. But we'll try the AK4330 just to compare, if we can get one.
The AK4330 also use a LP filter at the output (470R 2,2N).
The ES9022 has a built-in Rout of 240R, and 4.7n caps.
So the filtering is the same in both.
And I expect the same benefit wit the JG Filter.
If you look at the JG Filter schematics, the first part is fully passive filtering.
The Cascoded JFET Source follower thereafter is just an impedance buffer.
You can leave it out if you wish, but we do prefer it with the buffer.
Why not just try it ? You can easily breadboard one.
Patrick
No lack of punch whatsoever.
So IMHO more to do with layout than DAC I think. But we'll try the AK4330 just to compare, if we can get one.
The AK4330 also use a LP filter at the output (470R 2,2N).
The ES9022 has a built-in Rout of 240R, and 4.7n caps.
So the filtering is the same in both.
And I expect the same benefit wit the JG Filter.
If you look at the JG Filter schematics, the first part is fully passive filtering.
The Cascoded JFET Source follower thereafter is just an impedance buffer.
You can leave it out if you wish, but we do prefer it with the buffer.
Why not just try it ? You can easily breadboard one.
Patrick
Another first today: The SPDIF output on AB-1.13 is now up and running. It came straight up with no issues. I tried 24/44.1 and 24/192 content on a Burson HA-160d. Next I'll try impulse responses etc. just to check that everything adds up.
Follow my signature for links to AB-1.12/1.13 schematics.
One purpose of adding SPDIF out is to be able to record the digital output of the USB interface and check if everything is bit perfect. Any opinions or suggestions for test methods, please let us know.
Børge
Follow my signature for links to AB-1.12/1.13 schematics.
One purpose of adding SPDIF out is to be able to record the digital output of the USB interface and check if everything is bit perfect. Any opinions or suggestions for test methods, please let us know.
Børge
I want to upgrade my AB-1.2 with the Vanguard 0.3ppm.(24-576MHz & 22-5792MHz)
I found it on eBay.
Vanguard Ultra precision Golden TCXO, 0.3ppm 24.576MHz | eBay
Vanguard Ultra precision Golden TCXO, 0.3ppm 22.5792MHz | eBay
This will be a good idea?
Maybe someone has experience with the 0.3ppm?
I found it on eBay.
Vanguard Ultra precision Golden TCXO, 0.3ppm 24.576MHz | eBay
Vanguard Ultra precision Golden TCXO, 0.3ppm 22.5792MHz | eBay
This will be a good idea?
Maybe someone has experience with the 0.3ppm?
Those XOs look very cool!
What you need in an audio application is low jitter. Long-term stability isn't really that important. There tends to be a correlation between good specs in one area and good specs in another. Feel free to install these babies in the AB-1.2 and tell us what happens!
I'll get the BOM uploaded. At the moment, all parts can be found in the schematics at Q N K T C USB-I2S Module and Analog Board
Børge
What you need in an audio application is low jitter. Long-term stability isn't really that important. There tends to be a correlation between good specs in one area and good specs in another. Feel free to install these babies in the AB-1.2 and tell us what happens!
I'll get the BOM uploaded. At the moment, all parts can be found in the schematics at Q N K T C USB-I2S Module and Analog Board
Børge
The Vanguards are not bad but the Crystek oscillators http://www.crystek.com/crystal/spec-sheets/clock/CCHD-957.pdf are 25 dB better at 1 KHz:
Vanguard: Phace noise: -125dBc/1KHz
Crystek CCH -148 dB @ 1 KHz
This would not be exceptional for exotic oscillators however the Crystek's are available from both Mouser and Digikey for reasonable prices in single piece quantities: http://www.digikey.com/scripts/DkSearch/dksus.dll?keywords=CCHD-957%20crystek Crystek Corporation CCHD-957 Standard Clock Oscillators | Mouser
In the same price range you can get the oscillators from Tent Labs which are similar to the Crystek's but in traditional cans.
Improving on these is a very expensive undertaking and probably not audible. However really good supply isolation and low noise is essential to getting the performance.
Vanguard: Phace noise: -125dBc/1KHz
Crystek CCH -148 dB @ 1 KHz
This would not be exceptional for exotic oscillators however the Crystek's are available from both Mouser and Digikey for reasonable prices in single piece quantities: http://www.digikey.com/scripts/DkSearch/dksus.dll?keywords=CCHD-957%20crystek Crystek Corporation CCHD-957 Standard Clock Oscillators | Mouser
In the same price range you can get the oscillators from Tent Labs which are similar to the Crystek's but in traditional cans.
Improving on these is a very expensive undertaking and probably not audible. However really good supply isolation and low noise is essential to getting the performance.
I actually tried to work out how to fit a CCHD957 onto the existing AB1.12/13 board layouts a few days ago, just looking by eye at photos of the pcb layouts I think they end up fouling some of the bypass caps and surrounding components potentially? Though I think the pad placement for the golledge parts lines up ok with the CCHD957 pads.
Last edited:
Borge would be the best for that. Changing surface mount oscillators looks to be quite difficult in general without some very special tools. You could add some short wires and float the oscillators above the board. Attach a bypass cap beneath them incase the internal bypass is not enough. The isolation from floating them may help since crystals are sensitive to vibration.
That, or you can dead-bug prototype them. Glue or tape them upside-down onto some free boards area and patch the wires onto the SMD pads of the Golledge XOs. Works with both SMD and pinned components. Just check if the case should be GND or float. And remember to put decoupling right onto the dead bug itself. Wires should be insulated and pushed tight to board/XO so that they don't create a loop antenna.
For my CD player I use the PCM1704 and am not bound to 2^n division rates. So I'll be using 13.0548MHz and divide by 34 or 37. It's off by +-92ppm but I can live with that, investing in one cool XO rather than two four the two conventional sample rates. Except Crystek don't make custom frequencies. NEL will do that, but they are very expensive.
Børge
For my CD player I use the PCM1704 and am not bound to 2^n division rates. So I'll be using 13.0548MHz and divide by 34 or 37. It's off by +-92ppm but I can live with that, investing in one cool XO rather than two four the two conventional sample rates. Except Crystek don't make custom frequencies. NEL will do that, but they are very expensive.
Børge
Readme file
Several people find this thread very long and confusing. To sum things up a bit I have started on a readme file for the Audio Widget project. I hope it explains things.
This file is still a work in progress, so feel free to suggest changes to it. This first version is biased towards the AB-1.x hardware implementations I have designed. However, that is not a goal for this file.
Børge
Several people find this thread very long and confusing. To sum things up a bit I have started on a readme file for the Audio Widget project. I hope it explains things.
This file is still a work in progress, so feel free to suggest changes to it. This first version is biased towards the AB-1.x hardware implementations I have designed. However, that is not a goal for this file.
Børge
