Problem with SRC4392 - WM8741

Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.
Hi, I designed a DAC with SRC4392 and WM8741 and have run into a problem I am unable to solve. The device works and it sounds good but when I decided to do some measurements, things don't look well at all. Take a look at the THD in the attached image. Spikes everywhere and 3rd harmonic at >-70db?

The chain was PC (RMAA software) -> CM8761 USB thingy with coax out -> my device -> Sound Blaster ZxR analog input.

In my experience, this kind of image is usually due to some kind of digital process - filter, resampler etc. But this looks too gross to be coming from any of that. The first thing to suspect would of course be the maligned RMAA and windows and some kind of resampling happening there. However, when I connect another DAC in the same chain, just using optical from the usb thingy instead of coax (they are simultaneous outputs), I get perfectly clean picture. The problem is not the RMAA input either, because I also own Keithley 2015 THD which corraborates THD at about 0.075% with 10+ harmonics (but only 0.001% with just second harmonic, just as the RMAA suggests).

So I obviously suspect I made a mistake in how I designed and wired and configured chips. I feed SRC4392 directly, all subsystems use the board generated master clock 24.576MHz, SRC is fed from DIR and port A is fed from SRC and port A output goes to WM8741. Port A is master mode, generating bit and lrck signal from master clock. My assumption is that SRC's job is to take input from DIR - which is fed master clock but obviously recovers the clock from input signal as well and produces data that is sync'd to recovered clock, not master clock - and resample it, providing resampled data at master clock rate. DAC chip is also fed master clock. In other words, I expect input at say 44.1kHz/24 bit synced to recovered clock coming from USB gadget and output from SRC at 192kHz/24 bit synced to on board master clock. And it appears to be working until you look more closely.

I replaced the WM8741 DAC chip suspecting it got damaged, with no change.

Anyone has any idea what did go wrong here? I can provide more info as needed.
 

Attachments

  • cs4392thd.png
    cs4392thd.png
    27.8 KB · Views: 187
I connected WM8741 directly to USB audio's I2S header and got the same result - -65dB 3rd harmonic. Again, other DAC connected to it is clean. So the problem seems to be Wolfson chip - but I am using datasheet schematics and I double checked values and connections. There are only a few registers to configure and I tried all combinations I could think of. I just can't see what could cause a DAC chip to produce such results (I replaced the chip itself before with no change).
 
This was caused by inappropriate bypassing capacitor for the WM8741 regulator. Apparently the SMD ultralow ESR polymer cap may be too much for the top of the line regulator (/s) and a hacked in aluminum can "normal" low ESR capacitor yields results which are dramatically better and generally acceptable. The regulator wasn't oscillating and monitoring rails on the oscilloscope didn't look bad but there it was. Anyhow hopefully a useful lesson for myself.
 

Attachments

  • cs4392thd-2.png
    cs4392thd-2.png
    28.5 KB · Views: 193
Status
This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.