Caiman dac chip decoupling

[robyonekenoby]

Hello, while this topic has been already discussed, I want to share my own results

After the passive mod, in my mind the next step to improve performance is working around WM8716 ps decoupling caps

I proceeded step by step analizing the result of every single action; my system is so transparent that allows me to catch the most subtle difference in DAC signature

1) replaced the 220uF coupled to the 7805 and the 100uF near the chip with rubycon ZA 220uF 35 V: improved bass depth and weight

2) added a 1uF unbranded film cap (orange drops) in parallel to the tants for AVDDL and AVDDR: smoother and more refined highs

So far so good

3) added a couple of 330nF MKT to bypass each tantalum for AVDD and DVDD: further refined highs, but a noticeable drop in bass and overall volume, I actually have to rotate furhter the knob to get same volume before 3)


I don't really have an explanation for this behaviour, someone else does?

I'm planning to replace both tantalums on AVDDL and AVDDR with 3x3.3uF wima MKS2, somebody did try this solution yet?

Thanks

[UV101]

If you can seperate and analogue and digital rails with seperate regs, I'd expect massive gains.

For me, I always use Rubycon ZA's on an analogue rails and OsCon SPEC or Nichicon PLE on digital. Both of those solid polymer are 2nd to none on the digi rails.

Its definately worth playing around.

Nice approach btw.......I tend to just throw it all in and see what happens (not very scientific eh!! )

[robyonekenoby]

Yes I know should be better having at least 2 separate regs for digital and analog pins, but unfortunately this mods would require very invasive surgery on the PCB, cutting tracks ...

My next dac for sure will have separate ps, and possibly two dac chips

But for the moment I want to squeeze out all the juice from this beast

The dac chip is very sensitive to analog pin decoupling, every suggestion is welcome

[robyonekenoby]

But .... wait a minute

I measured 2.5 V across the cap I supposed to be AVDDL, and makes not sense

This voltage should be found on VMIDR and VMIDL

Is the cap code printed on the pcb respecting the wolfson datasheet?

I mean, the C5 on pcb is the same C5 on datasheet?

Follow tracks from caps lead to chip pin is almost impossible

[UV101]

this mods would require very invasive surgery on the PCB, cutting tracks ...

Go on......cut the track.........you know you want to!!!!!!

[robyonekenoby]

Too late

As I stated in the other thread, my wm8716 is gone

I've probably shorted the VMIDL pin to ground with the scope probe, and now the left channel output voltage is halved

[UV101]

Rather cheeky, but do you want to sell it as is?

[robyonekenoby]

with a fried dac chip?

Who is going to buy a not working dac?

At least I'm quite sure is fried, my only mistake was shorting for a moment the leads of the 10uF tantalum on the VMIDL

Maybe I've just fried the tant?

[UV101]

with a fried dac chip?

Who is going to buy a not working dac?

At least I'm quite sure is fried, my only mistake was shorting for a moment the leads of the 10uF tantalum on the VMIDL

Maybe I've just fried the tant?

You should check for obvious things first. Also, the actual DAC chip is available in the UK for less than £5. If you are unable/ not interested in repair, I'd be happy to make a small offer for the non working unit if you are interested?

[StanleyB]

with a fried dac chip?

About U$5 for the chip, and 15 minutes of a steady hand to replace .