• WANTED: Happy members who like to discuss audio and other topics related to our interest. Desire to learn and share knowledge of science required as is 20 years of participation in forums (not all true). Come here to have fun, be ready to be teased and not take online life too seriously. We now measure and review equipment for free! Click here for details.

ES9038Q2M using SPDIF 192 input THD problem

IVX

Member
Joined
Mar 3, 2019
Messages
31
Likes
16
#1
Hi everyone, I keep playing with my proto DAC+HPA project and found abnormally high THD(.002%@1kHz 0..-20db pretty much the same) with SPDIF input and default reg#0 = 0x00 i.e. clk_gear = XI/1. If clk_gear = XI/2 or XI/4 either XI/8 the THD becomes better, and XI/8 THD have the lowest value .0004%(without TH_COM cheating or .00025 with). The problem is ES9038Q2M does work with SPDIF Fs=192kHz(96kHz ok with XI/2 and XI/4, 44-48 ok with XI/8) only if reg0=0, hence I've no chance to get decent THD at 192 at all. I quite happy with 44-48 SPDIF input because it does work with XI/8 setting, THD at 96 .0007% is so-so, and only if MCU will dynamically setup reg#0 according to current audio format(if 192 then reg#0 = 0x00, 96 then reg#0 = 0x08, 44-48 reg#0 = 0x0C). Such on the fly update makes muting for 1/2 S, it is really bad if a track-list contains different formats. The clock is SiT8209 100MHz, I've tried 50MHz Xtal with ES9038Q2M internal oscillator and got a bit better THD but SPDIF 192 didn't work at all, hence 100MHz isn't max XI frequency but rather minimal one or ONLY the frequency allowed. I use CM6642 as USB/SPDIF bridge(5x6mm QFN with 3-5pcs 0402 surrounding that allows me to fit all in the USB dongle like case), and I have the same THD for 44.1 either 192 in another my design where WM8804 does convert SPDIF into I2S, hence, the root cause is ES9038Q2M.
Any input, ideas welcome.
 

IVX

Member
Joined
Mar 3, 2019
Messages
31
Likes
16
#2
Finally, I found the reason )) It was fake LDO "TPS79133DBVR" which I bought on taobao intentionally most expensive variant to avoid fake-related problems. That fake has very low loop gain so the PSRR is about [email protected] or so. When I switched clk_gear = XI/1 DAC consumption dropped proportionally and decoupling caps helped to keep 3.3V rail's ripple good enough. Now I tried 1A LDO MP20051with PSRR 63db and 13uV noise, clk_gear setting not affect THD at all, adding 1000u [email protected] to the 3.3V rail also doesn't affect THD performance. Chinese always makes fakes for famous companies - Ti, Linear/AD etc and never for tiwanese like MonolithicPower aka MPS. I bought 20pcs MP20051 just for $0.15/pcs and it is genuine LDO!
 

Similar threads

Top Bottom