CC8530 Audio Output on Master and switching masters

Hi David,

I'm not sure what you mean when you say that you would stream to the master via another wireless technology, can you elaborate on this? It is not possible to route and input to the chip to an output internally. Only signals received over the network can be output on the device. 

Also, each slave can only be connected to one master/network at a time, so a multiple network slave configuration is unfortunately not possible. 

Regards,
Johan

Johan,

I would be streaming to the master speaker using bluetooth on a CC2560.  Then I would connect a system of 4 speakers using PurePath.  I am asking if you have the master CC8530, do you need an audio codec to output directly from the master CC8530 to a speaker (with a delay to match the slaves)?  If I want a system where every speaker can serve as the master or as a slave, do I need a master and slave cc8530 in each speaker?  Only one master at a time, but I want each speaker to have the ability to serve as the master or as a slave.  What would this take?

Thanks,

David

I am also facing a similar problem as I have been trying to accomplish the exact same thing for about 10+ months since starting to work with the CC85XX chipset.

It seems that if a speaker or audio device needs to be able to "adapt" and become a slave or a master depending on if it's receiving or distributing the source audio, from bluetooth, the chip would need to be re-flashed whenever the role is switched.  This leads to excessive erase/write cycles on the CC85XX chip.  Would anyone at TI know if this is acceptable?

Thank you,

-Zack Lyzen

Lyzen Audio LLC

Zach,

Have you considered using 2 CC8530's in each speaker: one master and one slave.  This will require a little more harware, but it avoids having the reflash the boards continuously.  We are looking into preventing the speakers from permanently pairing master/slave relationships so every time you turn the speaker on in slave mode, it scans for master speakers near it.  I am unsure about the excessive erase/write cycles, but this would be another work around.

Regards,

David

David,

This is my exact predicament.  I would like to avoid laying out 2 circuits due to cost, board size, and size limitations of the product.  Perhaps they can share the same CC2590/antenna since only one will be powered at a time though.  

The idea scenario is that when the audio devices are on standby they are in slave mode looking for a master signal which would be presented when the user powers on one of the speakers.  thus allowing them to "wake-up" their whole set of speakers by only pressing the power button on one device (without this a network of 5 speakers would need to be turned on individually by the user without additional hardware).  

Other than that I would need to have a different firmware/configuration available that allows me to switch between master/slave with only SPI EHIF commands.  That is the "ideal" solution, whether its possible or not would probably depend on new firmware.

Is this something TI is able to provide for a small business/startup?

-Zack

The more I look at this:

- I do not have enough memory on my host processor to store 2 configuration files (I would have to add external memory)

-even if I did, I would not feel comfortable re-flashing this chip so frequently (possibly dozens of times in a day) since it seems like the product would be doomed to premature failure due to memory fatigue (limited erase/write cycles?)

so...

- I will have to either:

1.) use 2 CC85XX circuits (1 master 1 slave)

2.) use 2 CC85XX chips (sharing cc2590/antenna)(if possible)

Unless...

There is a way I can custom-configure a CC85XX to have the functionality I described above (a different firmware/configuration available that allows me to switch between master/slave with only SPI EHIF commands)

Would anyone working on Purepath and the CC85XX chipset at TI be able to advise me on this?

Thank you,

Zack 

Hi Zack, 

This is not possible I'm afraid as there is not enough internal storage to contain configurations for both slave and master. For reference, same applies to USB vs. I2S. 

Regards, 

Kjetil

Thanks for your reply Kjetil,

Is there any danger in leaving the RF_N and RF_P lines as reflective shorts?

I am interfacing 2 CC8520 chips (master and slave) to one CC2590 via a rf switch and when one IC is disconnected from the CC2590 it is a reflective short.  Should I add a 1-2M resistor in parallel to these outputs?

I would only need one of the chips powered on at a time.  However, I need to have both chips powered since the voltage rating on the digital input pins (SPI lines, is VCC+0.3V) and both CC8520 chips are connected in parallel to the SPI lines.    Normally I'd have the chip set to "off" or the deepest standby state I could find, but it seems from the datasheet and family user guide that the RF never completely stops when the chip is connected to VCC.  Is this correct?  If I am able to stop RF transmissions while still applying voltage to the chip then I am good to go but it seems this isn't possible.

Do I have the right idea so far?  or will I be damaging my chip since when in "off" (but still receiving power) the CC8520 will be transmitting data and RF_N and RF_P will be reflective shorts in standby "off" mode.

A little more about my project.  At most I will be streaming to up to 4 slaves with PCM.  I need data, but looking at the configuration application it looks like the RC commands might work.  Which would lead to better bandwidth and better audio robustness, data side or RC commands?

The "master" can control all of its commands locally and any slave that receives a similar command should be able to send it as an RC command.  It also looks like network standby can be initiated via an RC command does the master have a way of implementing this to all of the slaves since I believe RC commands are only Slave->Master.

Also:

I have the headset development kit and I am pleased with the performance, I do imagine however, I will see a drop in link robustness as I add slaves to the network is this correct or do the time slots always account for 4 slaves? (I always plan on using only 2 channels of PCM)

Thank you very much for your help

Zack Lyzen

Hi Zack,

I have actually never tested a system setup like this. I discussed your concerns with one of our hardware experts and our conclusions is that it should be OK, but since we have not done any tests on this here we can't give a more detailed/correct answer that this. One solution would be to use a switch that connects the unused pins to some 50 Ohm load instead of leaving them reflective short.

The master can tell all the slaves in the network to go to network standby (User's guide section 3.7.1.1):

"When the protocol master enters the NETWORK STANDBY state, all connected protocol slaves will follow.
Network standby mode can be controlled directly at the protocol master and/or through the protocol master using remote control commands. For example, in a system with a stereo input base station and two wireless
mono speakers, a “Standby” push-button on one of the speakers can take all three units into and out of the
NETWORK STANDBY state."

Is this what you are looking for?

The autonomy of the time slots depend of the number of slaves that is supported for the network. So if you don't want to use 4 slaves, you can get a better robustness if you configure your system to only support the number of slaves that you actually are using.

Regards,
Johan