WM8805 upgrade board (cs8414 pins) - dissapointed

[mfrimu]

Hello,

I have finally installed an wm8805 upgrade board - pin compatible with CS8414 and find the results inconclusive.

Ebay:
WM8805 Digital receiver, CS8412 CS8414 converter module | eBay

Other than minor malfunctions of switching circuits, the borad is 'working' but I can't distinguish any remarkable sound improvements.

I happen to have 2 dacs and compared the sound with the stock cs8414 dac. Both sound good. My transport is a cheap LG DVD player, so I would assume there should have been an improvement.

The minor malfunctios of this kit:

1. Led indicating input source (4 of them) don't light up
2. The switch that is supposed to change the outpout mode (i2s 16bit, i2s 24bit, etc) does not have an effect on the associated LED's indication the actual output mode, hence not working
3. Missing error pin on pin 25 of original cs8414 (my dac used this for an led).

It looks like the circuit does not use the crystal option on the wm8805, maybe that would improove performance?

Maybe it would have been easier to build this from scratch, assuming the WM8805 can output i2s 24bit in hardware mode (as most of it's settings are only available in software mode)

Also my Dac has a 'buffer' on the spdif input, not sure that would affect negatively the wm8805.

Any thoughts?

[jean-paul]

Hi, WM8805 needs a 12 MHz clock for operation which includes dejittering the incoming signal. This function is one of the features that make this chip stand out from the rest. It certainly is way better than CS8412 or CS8414. All this together with the issues you mention make me think you have a crappy designed device. It is definitely not the chip that is disappointing I can tell.

Maybe you can ask the seller for a schematic ? BTW except for item #3 I would call the "minor malfunctions" major malfunctions as the device obviously does not work as advertised. It is defective.

[mfrimu]

Thanks for the info. I was expecting improvments even without the external 12MHz crystal as the intristic jitter of the device in 50ps and PLL clock recovery and de-jitter using a buffer can be quite effective. The crystal 8414 has 200ps of intristic jittter, so no matter how good the input or de-jittering pll, the device can contribute its own 200ps.

In terms of starting from scratch my concern is that thw WM8805 requires software register settings for i2s output among other things, and that can get complicated, hence the Atmel cpu on the chineese board. The WM88805 Data Sheet is not clear about this.

[jean-paul]

Send it back to the seller and ask for a working one is what I would do first. If the new one works OK I would add a 12 Mhz clock (if it has none).

It can be simple sometimes You can't expect improvements when the chip is not used correctly or when the design has flaws. Are you sure the WM8805 Xin pin is not connected ?

[abraxalito]

Originally Posted by mfrimu

Other than minor malfunctions of switching circuits, the borad is 'working' but I can't distinguish any remarkable sound improvements.

I happen to have 2 dacs and compared the sound with the stock cs8414 dac. Both sound good. My transport is a cheap LG DVD player, so I would assume there should have been an improvement.

I have quite a few cheap DVD players (unknown Chinese brands). Their SMPSUs are fairly noisy and the noise carried through the SPDIF cable into the DAC is the main reason for sound degradation. If you want to improve the sound you're getting, improve the grounding of the SPDIF input of your DAC rather than swapping out the SPDIF receiver.

[mfrimu]

Yes, ground noise is probably an issue with new cheap players. I have added a Monarchy Audio DIP which has an isolation transformer on the spdif output. (I have tsted without the DIP and with the DIP, both results inconclusive)

[abraxalito]

Isolation transformers only isolate perfectly at DC. Do you know the interwinding capacitance of your transformer? It would still pay to ensure that the ground side of the trafo output doesn't share any common ground impedance with the audio outputs.

<edit> Ooops, misread your post and assumed you'd added the trafo on the input. Sorry about that. A trafo on the SPDIF output can still couple noise, depending on the interwinding capacitance.

[Pano]

Interesting. I had seen these boards on Ebay. Glad I didn't buy one.

[abraxalito]

I recently bought this one, from the same designer:

DAC TDA1543*4+WM8805 NOS USB, Optical, Coaxial input | eBay

Had no problems at all with the WM8805 and software, just the grounding (hence sound) sucked until I rewired it Now it sounds jolly sweet, even fed from a 199rmb DVD player.

[Pano]

Cool! Thanks for the link.

I don't know what Hui means in Chinese, but in Hawaiian it's "group" or "bunch". So raindrop_hui is kinda cool. A downpour?