• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 30 subscribers
  • Views 1098 views
  • Users 0 members are here
Support feedback
Options
Options
Related

This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

About PurePath Wireless(CC85xx)

steven chuang1
Prodigy 140 points
Other Parts Discussed in Thread: CC8530, CC2590, CC8520

I have a few questions on CC8530:
1) Without a master programming MCU, what are the default data rate (2M/5M) and audio sampling rate?
2) Is uncompressed audio achieved for both 2M and 5M configuration? What is the trade-off in audio quality?
3) What is the typical range achieved when using 2M/5M data rate?
4) When CC2590 is added, will it still comply with FCC part 15? what will be the range improved for 2M/5M operation?

 

Best regards

Steven

  • 0
    TI__Mastermind 21125 points

    Hi Steven, 

    1. Even in autonomous mode this is configured through the PurePath Wireless Configurator. Default in our examples are usually 5M and 48 kHz, but this can be changed to whatever you feel like. 2M/5M mode is set under Advanced Options(both master and slave has to be set to the same)
    2. From the PurePath Wireless COnfigurator tool:

    In most cases the default 5 Mbps data rate should be used. Using the 2 Mbps data rate increases sensitivity (and thereby theoretical range) at the expence of throughput capacity and number of tries per audio slice, and is only recommended for networks with:
    • 1 audio channel using any streaming format or 2 audio channels using the SLAC or PCMLF streaming formats
    • Maximum audio latency (2048 samples)
    • Maximum slave count of 1
     

Trade-off with regards to audio quality for SLAC vs. PCM16, please see appendix "D SLAC Performance Analysis" in the User Guide (http://www.ti.com/lit/ug/swru250l/swru250l.pdf)  

3. This is very dependent on the overall configuration and not only 2M/5M setting. Typical parameters influencing range is:
- number of slaves
- sample rate
- latency
- number of audio channel supported
- direction of audio
- streaming
- target output power and CC2590 range extender
- audio fading
- RF data-rate 

Usually we recommend our users to test a real life scenario with our development kit. Switching from two channels PCM16 at 5M to two channels SLAC at 2M should give a substantial increase in range for a given latency setting - ball-park estimates would indicate roughly twice the range (line of sight) for the higher latency settings under good RF conditions. 

The theoretical answer is that 2M gives ~3dB better sensitivity vs.  5M mode at the expense of throughput capacity and number of re-tries in case of lost packets. 

4. CC2590 will still comply with FCC part 15 (see this FCC report for the Headset kit: https://apps.fcc.gov/oetcf/eas/reports/ViewExhibitReport.cfm?mode=Exhibits&RequestTimeout=500&calledFromFrame=N&application_id=143803&fcc_id=ZAT85XX2590EM). Similar to the above, it is hard to give a good indication of the range improvement. The CC85XXDK contains evaluation modules both with and without CC2590 to enable customer to test this themselves.

Best regards, 
Kjetil 

  • 0 in reply to Kjetil
    Prodigy 155 points

    I understand giving a numeric answer is an over-simplification, but it can be very useful as a starting point. Ti tests ( http://www.ti.com/general/docs/lit/getliterature.tsp?baseLiteratureNumber=swra320&fileType=pdf ) show 20m indoors (non-LOS) without the extender, and 25m with. NB this is on both master and slave, at PCM16 (which is the 5M setting i think?).

    Sadly they do not test SLAC range or the range of a master with extender together with a slave without (nor distances much above 25m), which would be very interesting to know.

  • 0 in reply to M. B.
    TI__Mastermind 21125 points

    M.B, 

    Agree that a numeric answer can be useful to some extent but I do want to stress the importance of testing this to get a feel on the performance yourself - there are so many parameters influencing this to give a complete matrix. The CC85XXDK contains boards with and without the range extender so that users can get a feel of the various combinations. 

    With or without the CC2590 in one or both end of the link and 5Mbit/2Mbit can be simplified looking at link budget only. 6dB increase in link budget gives a theoretical 2x distance - but off course the real world is usually not that simple. The datasheet gives sensitivity and output numbers for the various combinations. 

    SLAC vs. PCM16 affects number of re-transmission attempts within a given latency. Under perfect RF conditions range should not be influenced by this. In real world there are no such thing as perfect RF conditions and changing this will have an effect. SLAC requires less bandwidth and thus the experienced range is usually longer.

    Hope this helps. 
    Kjetil 

  • 0 in reply to Kjetil
    Prodigy 155 points

    Completely agree testing is required too.. but it would be a waste of time if you were expecting way off, so more numbers please, just add the disclaimers!

    Are you saying that there is no benefit to using a CC2590 unless you use it on both the master and slave? AFAICT the datasheet shows nothing about combinations (eg master with, slave without).

    Also am i right in thinking that there is no way to extend the signal other than using the CC2590 at the link ends? -- what i mean is, there is nothing that can act similarly to how a wifi range extender/repeater works??

  • 0 in reply to M. B.
    TI__Mastermind 21125 points

    M.B

    Since the protocol / link always is two way (even in cases where the audio is uni-directional) you will have to take this into account when looking at link budget and with/without CC2590 considerations. 

    From the datasheet:
    Output power CC85xx: 3.5 dBm
    Output power CC85xx+CC2590: 11 dBm

    Sensitivity CC85xx: -83 dBm
    Sensitivity CC85xx+CC2590: - 87 dBm

    This gives:
    Link budget with CC85xx in both ends: 86.5 dB
    Link budget with CC85xx+CC2590 in both ends: 98 dB
    Link budget with CC85xx in one end and CC85xx+CC2590 in the other:  90.5 dB

    Note that in the last case you basically only get the gain of the LNA part of the CC2590 since the link is bi-directional in nature. That being said, if you add the CC2590 in the audio source side of a uni-directional audio link the additional output power do allow for further range as it will ensure longer range for the audio than the acknowledge back. 

    As a reference, 2Mbit gives an additional ~3 dBm sensitivity over standard 5 Mbit link. 

    Depending on your configuration you might have the chance of setting up a 'range-extender'. In a point-to-point scenario you can add a 4 channel bi-directional master to allow for a Slave1 -> Master -> Slave2 scenario where the master is receiving audio from slave 1 and sends it to slave 2 (I2S output is tied to I2S input). 

    Regards, 

    Kjetil

     

  • 0 in reply to Kjetil
    Prodigy 155 points

    Thanks..

    So a link with the CC2590 on only the audio source side (one-way audio only) should provide a 93.5 dB link budget? That is (say we are sending a 5Mbit audio signal) the acknowledge back  signal requires a 2Mbit link, or will the acknowledge back signal require less bandwidth (and therefore possibly a little higher link budget?)

    RE range extender:

     Say you have 3 devices; A, B and C. B is the 'range extender'.

    1. Is this possible: Without B present, A is a master/audio source and C is a slave/audio sink.
      With B present, A becomes a slave (but still audio source), B takes over as master, and C remains as audio sink.
       
    2. In this scenario, does B require 1 or 2 CC85xx devices?, assuming A/C have 1 each.
       
    3. Is it possible to make this process (adding range extender to network) transparent to the end user, given that B would always be a dumb range extender, C would always be a dumb slave/sink, and A (of which there could be multiple of) could be a microprocessor controlled unit that must be able to function as either sink or source. Or would this necessitate some form of manual control?
       
    4. The addition of the range extender would reduce the number of other devices possible by 1, right? (as B would be present on the network but not playback audio, unlike A/C/other members)
  • 0 in reply to M. B.
    TI__Mastermind 21125 points

    Hi M.B

    No, the theoretical link budget is still 90.5dB for the bi-directional.RF link. 
    Lets say you are transmitting audio from master to slave and only the master have the CC2590. The link will work fine within the 90.5dB range, but if the distance increases, the packets from the master to the slave (including audio) will still reach the slave due to the higher output power of the CC2590 while the return data (audio acknowledgement) will no longer reach the master. The master will re-transmit the audio due to lack of ack, but as the protocol is written to maintain a fixed latency it will only do so until it has to transmit the next block of audio. 

    2 Mbit is a change in the physical behavior of the radio and both ends of the link have to be set to the same and the datarate is fixed and can not be changed run-time.

    RE range extender:
    1. No this is not possible. A device can not change role from master to slave or vice versa
    2. Depends on the number of mono channels from A to C. If this is two or less then B can consist of a CC8530 (up to four channels) while A and C can be CC8520 (up to two channels). A CC8530 can at the maximum support up to 4 mono channels
    3. No - due to no way of changing from master to slave. Typology for CC85xx devices is always a star network with up to 4 slaves. Slave to Slave communication is not possible without going through a master.
    4. The number of devices that can be paired to a master is 4 no matter if these for sink/source audio or not.

    Regards,
    Kjetil 

  • 0 in reply to Kjetil
    Prodigy 155 points

    Ouch. It seems there may have been some limiting decisions in the design..

    I am confused.. so master to master communication is not possible, unless it is through IC I/O on the same PCB ??

    And a device must be configured at manufacture as either slave or master, and can not be changed, even when MCU is present.

    If stereo PCM16 or PCM24 is desired, then two CC8520 will be required on the 'range extender' (B above),  because 4 channel only works with compressed audio on the CC853x. You imply that for each channel it can only receive OR send, so does this mean they will both be masters, and therefore support up to 8 slaves on two distinct networks I guess it would support 4 slaves from the transmitting CC85xx plus one from the receiving one?

  • 0 in reply to M. B.
    TI__Mastermind 21125 points

    M.B

    Our solution do come with limitations, especially within applications that stray from the originally intended applications for this platform.
    Not my intention to confuse. I would recommend our 120+ page User Guide that goes into details on a lot of this:
    http://www.ti.com/lit/ug/swru250l/swru250l.pdf

    Section 1.2.2 lists target applications that we had in mind during design and FW implementations
    Section 2.2.1 goes into detail on Network Topology 

    Short answers:
    Master to Master communication is not possible as a PurePath Wireless network can only consist of one master
    Only way to change master/slave configuration of a device is to re-program the device. This can in theory be done from a host MCU, but the on-chip flash memory in the CC85xx device is limited to ~1000 erase cycles.
    4 channels do require the use of our SLAC compression algorithms and if this is not acceptable then you will need two CC85xx in the 'range extender'. Both would be masters. This is not something we have tried so how well this would work is highly uncertain. You will have to start considering isolation between the RF antennas and how to handle clock within the two networks etc. 

    Believe I can provide more targeted feedback for the application you have in mind of you can explain more in detail on what you want to achieve. If public E2E is not a good option, please go into my profile and add me as friend for a more private discussion.

    Kjetil