Configurator-Example "Wireless Headset" Problems

Hi,

I found the problem.  Needed to remove my own custom files from My Documents (Application Roles and Audio Devices), and do a clean install of Configurator.

However, when using the Example "Wireless Headset", I have been trying to modify the audio device settings by choosing 'custom setup', and find it empty.  When 'Reset All', the different settings for master (M) and slave (S) are lost.  I recall TI saying that the .PPWADD and .PPWARD files had not been properly modified in this example to allow custom changes.

Would TI please supply a full set of files for the Example "Wireless Headset" so that customers like me can use the 'custom setup' to change parameters.  This is needed because I am making some error in developing  files for our own application which gives grossly excessive noise (even when touching the line input terminals), whereas the Example "Wireless Headset" does not.  

If you cannot provide full files, how to you modify the codec settings for Example "Wireless Headset".  Any chance of a fast reply on this one.

Thank You

Geoff Pickford

Hi Geoff,

I don't understand your problem. Every time you enable custom setup you will find the Audio device customization sequences empty. This goes for all examples, not just Wireless Headset. Infact the default setting for the Custom setup label of every audio device is Disabled. When changing this to enabled the Audio device customization panel will allways be empty. At the bottom of the screen you will even see the message: All configuration sequences are empty. Press "Set all sequences to default" to initialize. If you then press "Set all sequences to default" the default I2C transactions will show up. 

I don't understand what you mean with TI to provide a full set of files. 

Best regards

Kristoffer

Hi Kristoffer,

Thanks for responding so quickly.

When I first tried "Set all sequences to default" for Example Wireless Headset, the 'filter sequences' appeared to fail.  On repeating my set-ups, all is now correct, and you are completely right.

When tested, mic input for the Slave (S) and line input for the Master (M) are effectively noiseless.

I then copied the aic3101 file to the MyDocuments/Texas Instruments/PP WirelessConfig/Audio Devices folder, and modified it by changing all "+{mic_input" sequences to be the same as the "-{mic_input" ones - i.e. all line input.

A test from M to S and S to M using line inputs are effectively noiseless, which is excellent.

When I modified the aic3101 file by changing all "-{mic_input" sequences to be the same as the "+{mic_input" ones - i.e. all mic input, a tests from M to S and S to M are effectively noiseless. 

One minor question relating to this Example, when using L channel into mic of one device with a -20 dB 1kHz signal produced a -20 dB signal from the other device. but when the R channel was used, there appears to be cross talk around -60 to -65 dB in the other device, vis around 40 dB or more leakage.  Is this normal?

We have been developing 1 M to 2 x S configuration using the 'filter sequencing' you kindly referred to some time ago.   Unlike the Example above, we have been plagued with significant noise, including unacceptable noise when touching any input terminal with a finger.

Attached is a zip file comprising the project, configuration, codec, application role files we have been using.  We have activated all inputs and outputs so we can explore various options, and then narrow them down to a small number.  Would you mind having a look at it and seeing if there are obvious problems.  My colleague and I have spent many hours trying to find the problem with no success.  The Zip file also contains an Audacity spectrum of the Example Wireless Headset, and also of our configuration - both set with an input signal such that the output signal is around -30 dB.  As you can see no comparison, even though the Example has 30 dB less amplification than ours.

Kind Regards,

Geoff Pickford

Hello TI,

It has now been a month since I posted this enquiry, which to me is disappointing.  Is there any better way that I should present my questions?

In that time, I have pressed on and spent much time finding several simple and now obvious problems, including the fact that the Line2 inputs are not self-shorting and have to be shorted or used to prevent noise for my application.  Further, a 'noisy' power supply of a computer running audacity introduced noise into the test setup.

Elimination of both sources has assisted in reducing the noise to a point that the project is allowed to proceed using the CC85XX and AIC3101.

Still large problems with mic bias which introduces 'packet' noise to an unacceptable degree for my configuration.  Is this a common problem - I would have thought the mic bias power supply would have been filtered?  

At this moment, I am planning on not using the mic bias at all, but providing my own - any advice on this one please?

Thanks,  Geoff