fx aUDIO D2160 pURE DigiTAL ampLIFIER

Status
Not open for further replies.
Not sure what you're asking ... you might miss analogue inputs with a full digital amp? - so for that an ADC is needed. This one uses PCM1808 I believe and the clock is 12.2Mhz so I assume it's sampling at 48/24. According to the datasheet, it mostly performs better at 48 rather than its max 96KHz sampling.

Yes I agree, the output filter is super important. Texas Instruments have a fantastic free spreadsheet for calculating this. http://www.ti.com/tool/LCFILTER-CALC-TOOL

My speakers are homemade 3-way using Aperion and Klipsch drivers, Mundorf supreme silver/oil cap, Jantzen cross coil inductor. 4 ohm in the bass rising to 16 ohm in the treble. They are bi-amped but can use a regular stereo amplifier too.
 
Last edited:
haha... yeah. I checked out some photos of it - seems to have a few issues - power supply is one, noisy regulators is another (one of them is the 5V analogue supply for the ADC, another is the 12V supply for the TAS - LM2596 have a low switching freq so they will pollute everything), and there are vias in the MCLK signal trace, the tcxo appears to feed at least two ICs - the ADC (very long trace with vias) and the TAS (shorter with several vias) but has no fan out buffer..... so, China style for sure... a little shop of horrors :) But they were on sale yesterday so I got one anyway for 130 Euros.
 
Last edited:
noisy regulators is another (one of them is the 5V analogue supply for the ADC, another is the 12V supply for the TAS - LM2596 have a low switching freq so they will pollute everything)

Fun, fun - I found precisely the same issue on a digital amp I bought from Taobao (much cheaper though at 300RMB). Dynamics completely lacking on the analog input, an LC filter between the switcher and the ADC power pin fixed this very nicely.
 
Hi abraxalito, 12V for the power stage IC like 5624 is not a critical rail regarding the noise. I using there 2MHz buck-converter 600mA SOT-23 IC with 10uF 16V 0805 X7R, and I got 1db better S/N than Ti promised.
So GVDD, VDD etc are not sensitive to HF noise... cool :) Did you use the 3.3R that Ti have on the GVDD_AB input and again on GVDD_CD?

How about PVDD? I assume that needs to be as good as possible because that couples directly into the output?
 
Hugh Jazz, sorry, I got mistake regarding 2MHz buck-converter 600mA SOT-23 IC with 10uF 16V 0805 X7R. It was about my early version but currently, there is 78L12 with 4.7 ohms resistors and 1uF 0603 to GND. However, earlier version has the same S/N&THD. The main power rail needs to be very clean indeed, and I keep there 8pcs 1000uF/35V and 7pcs 1uF X7R 1210, all of that placed on very compact area, I guess it is about 1/4 of your SMPS_less_amp PCB.
 
TAS5558 layout for reference
 

Attachments

  • tas5558.jpg
    tas5558.jpg
    174.9 KB · Views: 284
The D2160 arrived - only had time for a quick listen with no break-in but my first impressions are positive - lots of definition and space around vocalists and instruments. However, I think there may be an issue with the left channel - sounds like a pumping/hiss overlaying the deep bass notes. I've got no time to investigate right now but I've asked the seller about it.
 
Status
Not open for further replies.