freeDSP-aurora - DSP with 8 I/Os, USB Audio, S/P-DIF, ADAT, Bluetooth and Wifi contro

Finally got the redesign of the aurora board assembled and first tests done. So basically I have redesigned the analog buffer part of the board with fully differential opamps, updated the power tree and then used 4-layer board instead of 2-layer. THD+N with the original aurora board was -100…-105dB depending on output, THD -100…-110dB. Now with new board THD+N -108…-109dB, THD -112…-115dB. Noise levels are lower and supply spurs are gone.

Happy to see even better THD+N results than on AK4458 EVM 😊. I still need to check what is causing increase on jitter, that is next on my list. I’ll post some pictures and test results once done.

Btw I’m using E1DA ADC (with REW) for testing, excellent device with low price!
Wow.. Great new. I hope will become available again to sell!!
 

tomik

Member
2019-04-04 5:56 pm
Finally got the jitter problem solved (well there is still some, but I guess it is the nature of ADAU1452 PLL clock and I can live with it). It was caused by too low input voltage from the SMPS (I turned down the supply voltage to make everything run cooler, but went too far..) Here's comparison spectrum with standard 1kHz 0dB signal on same output. In some frequencies the differences were even bigger. This is with optical input.

Standard aurora THD -104dB, THD+N -102.9dB (for comparison I would need to re-test the minidsp 4x10HD with the E1DA ADC, but I already know aurora beats it with some margin). Btw this is with OPA1612 which I tested to give best results.

new THD -115dB, THD+N -109dB (AKM spec -107dB)

Still some work to do, I'm missing the CS2100 clock chip to get the XMOS USB interface to work. Sourcing components is just horrible at the moment... And I would like to try using OPA1637/THP210 instead of OPA1632 to make it run cooler. OPA1632 runs quite hot especially when there are 16 of them on the same board...

comparison_1kHz.png
 
  • Like
Reactions: 1 users

tomik

Member
2019-04-04 5:56 pm
Anyone managed to flash XMOS FW? I first failed on Windows computer to even compile the FW, it just didn't work. The error messages were not very helpful...

Then on Mac I got the FW compiled but now it is failing on flashing. I'm using xTAG 3 debug interface and it is recognized correctly. One green LED lit on board. Aurora board is powered and running normally when I try to flash. xTIME Composer just says "An internal error occurred during: "Launching app_freeDSP-aurora 2i8o8xxxxx_tdm8"." Not very helpful and not sure how to debug...

Any tips?

Thanks.
 

tomik

Member
2019-04-04 5:56 pm
New error messages trying to flash from terminal window instead of GUI. Now I get this error message:

Warning: F03098 Factory image and boot loader cannot be write-protected on flash device on node 0
xflash: Warning: F03148 --quad-spi-clock not given, using default 15.62MHz
s2l-n0-e3b23621:2907:3: error: use of undeclared identifer `quad_spi_qe_location_status_reg_0'
quad_spi_qe_location_status_reg_0,
^
s2l-n0-e3b23621:2908:3: error: use of undeclared identifer `quad_spi_qe_bit_6'

quad_spi_qe_bit_6
^
Error: F03010 Failed to compile second stage bootloader
 

tomik

Member
2019-04-04 5:56 pm
Finally got the XMOS building and flashing working. Few cycles of reinstalling SW/drivers etc. and got the building to work on a Windows 11 machine and got connection with XTAG3 established. Then I deleted the `quad_spi_qe_location_status_reg_0' and `quad_spi_qe_bit_6' declarations and finally the flash got through. Here's the messages I got after deleting those:

xflash: Warning: F03148 --quad-spi-clock not given, using default 15.62MHz
xflash: Warning: F03149 QE_REGISTER and/or QE_BIT locations not found in XN file for this flash device. Using default flash_qe_location_status_reg_0 and flash_qe_bit_6.
Warning: F03150 The use of libquadflash will be deprecated from XFLASH in xTIMEcomposer 15.0.0.
Please add the PageSize, SectorSize and NumPages attributes to your External Device definitions in your target XN file to enable the use of lib_flash.


Anyway it is now recognized as audio device in mac. I still need to test how it works..