I think 6642 is a new compact alternative for 6631, even if you need one more SPDIF receiver IC to get finally I2S. Tiny CM6642 QFN-48 5x6mm + WM8804 SSOT-20 + 12MHz Xtall requires about 45% of the PCB area where was implemented 6631 based USB audio interface i.e. CM6631A+Pm39LV512 + 49.152MHz 3.3V low jitter OSC + 45.184MHz 3.3V low jitter OSC + 12MHz Xtall. Furthermore, WM8804 with an average Xtall probably outperforms CM6631A regarding jitter, the BOM cost will be also 20% less. For instance, I migrated from 6641 to 6642+WM8804 and got more than enough PCB area to implement CC2540 BLE controller on the same PCB size(see attached screenshots) and within the same BOM cost!
Drawbacks:
a) well known SPDIF limit 24/192 vs 32/384 for CM6631A, also, CM6642 doesn't support two Fs 88 & 176.
b) this is a quite new and C-Media's IC, hence near to impossible to get any application info. Actually, this is a complex chip with ADC/DAC, SPDIF and headphone amp+chargepump onboard, a brand new chip configured as headfone+mic set without SPDIF.
Drawbacks:
a) well known SPDIF limit 24/192 vs 32/384 for CM6631A, also, CM6642 doesn't support two Fs 88 & 176.
b) this is a quite new and C-Media's IC, hence near to impossible to get any application info. Actually, this is a complex chip with ADC/DAC, SPDIF and headphone amp+chargepump onboard, a brand new chip configured as headfone+mic set without SPDIF.
Attachments
oooh, I forgot to mention that 6642 doesn't need any driver to work with Windows(and Mac/Linux of course but that's as usual), I consider this fact as pros too. In case of my W7 64, the first time system spends about 1 minute after plugging USB with 6642 to recognize it and show the new audio device. After the recognition, 6642 ready immediately after plugging in USB. Nice, don't need to explain users anything about how to use a USB, just plug it in, that's it 
After a few months using that IC my users found a serious problem of that, actually design mistake. The chip doesn't check the validity of audio data(no CRC) and if USB connection will have some disturbances, that quite common for portable use(shaking cables, dust etc), CM6642 will send to your DAC any kinda corrupted audio data, like a noise, distortions. So I want to inform about that issue anyone who want's to use the IC, I recommend do not use it at all.
Have you found a good alternative? It seems there aren't a lot of options for low cost USB HS to I2S or S/PDIF bridges in QFN packages. There are a couple of Savitech parts like SA9137 that look like they could work. Beyond that, the alternatives are general purpose MCUs that end up pretty large and expensive like STM32F730I8 (~$4, but a BGA), Nuvoton M484SIDAE2U (~$4, QFP), FTDI FT903 (~$5), or XMOS XU208 (~$6).
Actually, Nuvoton NUC505YLA2Y also looks interesting. <$2 QFN with USB HS.
On my Nobsound 8xTDA1387 CM108 chip does the same, it passes corrupted data stream to the DAC, but only in WASAPI push mode. In the WASAPI event mode it behaves the same like XMOS does. In any events of distrupting data transfer, it silences data stream for a half a second. Perhaps you should try.
CM6642 looks like an improved version of CM108 (that only support a maximum 48kHz sample rate).
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.