DSPRelated.com
Forums

Cirrus (Crystal Semi) CS42436 codec noise problem

Started by Roman Rumian September 29, 2008
Hello,

in a multichannel audio mixer I use two CS42436 codecs. Input and 
outputs are differential.
When the input PGA is set to more than +12dB then a audible level of 
white noise can be heard at the outputs (codec is passing through input 
signals). Shorting inputs does not change this situation.

Technical support has not sent the answer for more than one week (they 
promise to give one in a three days), so maybe you can help me ?

Regards

Roman Rumian

Roman Rumian wrote:
> Hello, > > in a multichannel audio mixer I use two CS42436 codecs. Input and > outputs are differential. > When the input PGA is set to more than +12dB then a audible level of > white noise can be heard at the outputs (codec is passing through input > signals). Shorting inputs does not change this situation. > > Technical support has not sent the answer for more than one week (they > promise to give one in a three days), so maybe you can help me ?
I used CS42438 a lot. My advice would be never use any built in features of the codec; do all controls either by DSP or by the external hardware. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
Vladimir Vassilevsky pisze:
(...)
> I used CS42438 a lot. My advice would be never use any built in > features of the codec; do all controls either by DSP or by the external > hardware.
thanks Vladimir. :-) It is too late to throw up internal PGA. Probably, most of high end multichannel audio devices use stereo codecs or even A/D converters only. Do you know how they are designed ? Kind regards Roman Rumian

Roman Rumian wrote:

> Vladimir Vassilevsky pisze: > (...) > >> I used CS42438 a lot. My advice would be never use any built in >> features of the codec; do all controls either by DSP or by the >> external hardware. > > > thanks Vladimir. :-) > > It is too late to throw up internal PGA.
Then you have to reconcile to the performance level you that you are getting. BTW, check the codec errata sheets also; there are few nasty bugs with the built in features.
> Probably, most of high end multichannel audio devices use stereo codecs > or even A/D converters only. Do you know how they are designed ?
I guess I know, since I design them :-) CS42436 is not a high end codec, it is a consumer quality part. The performance designs use the external front/back end amplifiers if the gain matching is required. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com
Hi Vladimir,

Vladimir Vassilevsky pisze:
(...)
> I guess I know, since I design them :-) > CS42436 is not a high end codec, it is a consumer quality part. > The performance designs use the external front/back end amplifiers if > the gain matching is required.
and what is your opinion about CS5364/6/8 multichannel A/D converters (also TI PCM420x) ? Have you used them or know a professional audio device based on these devices ? Kind regards Roman

Roman Rumian wrote:

> Hi Vladimir, > > Vladimir Vassilevsky pisze: > (...) > >> CS42436 is not a high end codec, it is a consumer quality part.
> and what is your opinion about CS5364/6/8 multichannel A/D converters > (also TI PCM420x) ?
The magic number is the dynamic range of 120dB or higher. PCM42xx fits this requirement. Very careful attention should be paid to the schematics and the layout.
> Have you used them or know a professional audio device based on these > devices ?
Many audio folks prefer AKM parts. This is the area of beliefs, and the AKM has the widespread reputation of the unbeatable performance. Vladimir Vassilevsky DSP and Mixed Signal Design Consultant http://www.abvolt.com