• State
  • Locked Locked
  • Replies 19 replies
  • Subscribers 54 subscribers
  • Views 3888 views
  • Users 0 members are here
Support feedback
Options
Options
Similar topics

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.

Johanson Tech: Sub-GHz Impedance Matched Balun - NEEDED?

nowell outlaw
Intellectual 405 points
Other Parts Discussed in Thread: CC1310, CC1190, TPS62063

Ok - we are working on (again) design for CC1310 chipset for using higher power (per designs) with CC1310 chips.  

Now we have a new ceramic filter for the CC1310 for RF frequency, etc.  


So - TI - is this a needed CHIP on a board?


 (   ) 

  • 0
    TI__Guru**** 317180 points
    Not sure what you mean by needed.

    You have two options to give the CC1310 the optimal load impedance and matching to 50 ohm,
    1) Use the discrete network as described in our reference designs (found under www.ti.com/.../technicaldocuments)
    2) Use the Johanson or Murata balun

    The first option given somewhat better performance, the second gives a far more compact layout.
  • 0 in reply to TER
    Intellectual 405 points

    Thanks.  Yes we are using the reference de design but having a problem with the RF.  This is an existing board that we added the amplifier to.   My engineer came to me this am saying we need to add this other board.   But I don't believe it, so trying to figure it out

  • 0 in reply to TER
    Intellectual 405 points

    so in.  


    top image is from the TI document.  Bottom drawing is from TI cc1310 dev kit...???

  • 0 in reply to nowell outlaw
    TI__Guru**** 317180 points
    Could you please post the schematics showing the full RF system including the amplifier (If you don't want it open on the forum, send it in a privat message to me). Could you elaborate a bit on RF problems?

    My previous post is still valid even when you want to use an external PA. The discrete version is possible to do simpler dependent on how much you have to filter before the PA.
  • 0 in reply to TER
    Intellectual 405 points
    ok. let me see if I can get a copy posted here. I think the issue is that we used the discrete solution for our old design and what he is telling me is that the values rely on manufacturing tolerances for the board. If we add the skyworks PA (or cc1190) then it's sort of guaranteed to work on the board.

    Not sure why TI didn't use the CC1190 on their long range module designs (?). That's why we went with skyworks design for this revision of our board. Due to it being suggested in the long range design PDF. It also limits you to using 915 mhz and not configurable down to 433 Mhz.
  • 0 in reply to nowell outlaw
    TI__Guru**** 317180 points
    Our discrete 14 dBm reference design is fairly robust with respect to component tolerances. Do you have some measurement results proving that the CC1310 + Skywork is dependent on component tolerances? That would be useful input for us.

    TI is a large organization and different groups do different things. I think the Skyworks design was done for a customer originally. A CC1190 design will also will be made available, I'm working on it.
  • 0 in reply to TER
    Intellectual 405 points

    Yes we used that for the first board design (14 dBm), but wanted to add the Skyworks PA to get better sending range. I'll see if I can get some data for you. Right now we are seeing if we can get these initial boards to work effectively - but my guess is that we will need to add the Baylun to make things work well in production mode. It's not clear yet.  At .55c USD each it makes it worth adding to the board as a safety measure.

  • 0 in reply to nowell outlaw
    Intellectual 260 points

    See the attached document for details about the design.

    Everything on the board works fine except RF transmitting. If i get a new board and disconnect the PA shunt and just put an antenna on the 

    (forgot the name but the small rf connector on the board) than the transmit / receive works fine.

    When I remove the 100pf shunt to the rf conn and put the shunt in for the PA than after 1 transmit the RF stage of the CC1310 is damaged.

    After that It won't initialize anymore (gets stuck at setfrequency in easylink init).

    TI TPS62063 is used as the power supply' big power and ground planes; Components selected with TI we-bench. 

    No disturbances in power delivery detected (checked with scope).

    ---

    The board probably needs a redesign. I don't mind swapping out the sky-works chip for the 1190. 

    I have designed multiple CC1310 board before using the discrete solution. And have not had any problems other than the basic impedance matching tuning that had to be done. All those other boards worked without issue.

    Stackup.pdf

  • 0 in reply to Thijs Sillen34
    TI__Guru**** 317180 points
    Sounds strange.

    - Have you checked that the Skywork part work if you disconnect the CC1310 part and apply a signal from a SG to the SKY input?
    - What is the max input power on this SKY part?
    - Do you get the same issue if you remove L10?
    - From the schematic/ layout snip everything looks correct but from the description it sounds like it's a short (?) between 100 pF serial cap and the SKY input?
  • 0 in reply to TER
    Intellectual 405 points

    So we are thinking that something is up with the skyworks PA that's not making it work right.  Not 100% sure.  Current plan is to pull the skyworks PA off the board and switch to cc1190 as the PA.  Not sure if we need to add Balun to the board design or not.  Has TI tested the CC1310 with CC1190?  Did they use a balun?  Why is the sample design using skyworks chip instead of TI 1190 chip?

    If you want since these aren't working at all I can send one to you if you wanted to check it out.  It's either that or they go in the garbage.


  • 0 in reply to nowell outlaw
    TI__Guru**** 317180 points
    I'm currently testing CC1310-CC1190. The version I'm testing looks good but need some more measurements to finalize. On this the discrete balun is used, the IPC was not available when I started.

    The Skyworks design is not made by the BU making the CC1310 and the app note contains some errors, the Rx sensitivity is not measured correctly.
  • 0 in reply to TER
    Intellectual 260 points

    Hi,

    See attachment for CC1310 + balun + CC1190 design.

    If you can spot any errors / concerns please post them here.

    Thanks,

    Thijsitems.zip

  • 0 in reply to Thijs Sillen34
    TI__Guru**** 317180 points

    CC1310_LaunchPad _ 2_CC1190RF.pdf The schematic + layout is not optimal. Try to do it more like the attached. The is for 869 MHz, 0.4 mm between the first signal layer and the ground layer

  • 0 in reply to Thijs Sillen34
    TI__Mastermind 30710 points

    You will also need a SAW filter after the impedance matched filter-balun if you are planning to use a external PA.

    //Richard

  • 0 in reply to RGW
    Intellectual 405 points

    so maybe that's what is missing from the TI / Skyworks design?  

      

  • 0 in reply to nowell outlaw
    Intellectual 405 points

    so this PDF which is 1190 -> 1101 board design from 2011 has a SAW on the board - no Balun that I can see.

    It sounds like for 1190 PA + 1310 , you also need SAW + Balun to make it work? Not 100% sure why skyworks design didn't have the SAW chip in it? Was that board ever actually made?

    Can we private send you our design to have it checked to see what you think?


    Here's the 1190/1101 file.   

  • 0 in reply to nowell outlaw
    TI__Guru**** 317180 points
    The SAW filter is needed in the CC1310-CC1190 design to be compliant with regulations. Typically to filter out phase noise. But you will get good performance also without SAW filter with regards to output power and sensitivity.
  • 0 in reply to TER
    Intellectual 260 points

    Added saw filter for 915MHz; changed / added a few component values; changed pcb layout.items2.zip

  • 0 in reply to Thijs Sillen34
    Intellectual 405 points
    Any Chance for a review? 915 mhz, not 868.