• State Resolved
  • Locked Locked
  • Replies 5 replies
  • Subscribers 51 subscribers
  • Views 2167 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.

CC2650: Range problem with own PCB

alex22
Intellectual 630 points
Part Number: CC2650
Other Parts Discussed in Thread: TIDC-BLE-TO-WIFI-IOT-GATEWAY, CC2640

Hello, we have made a hardware solution using the CC2650 and based on the BLE part of this design:

Time ago we have also have assemblied the texas design as it is, and we have it here for comparison, the problem is that our custom hardware BLE range is very poor, using both the integrated antenna (which is the same that the texas design uses), or a external antenna through the U.FL connector, we have a range of 1 meter, more or less, while the texas design range is far superior. We have noticed that, despite TI recomends the integrated balun Murata LFB182G45BG5D920, and is the one that figures on the schematic, in the BOM they use the Murata LFB182G45BG2D280, which was the reccomended for the CC254X, and not for CC26XX. I guess this is not the problem because both hardware are using the same balun (Murata LFB182G45BG2D280), despite of not being the recommended.

I'm trying to figure out the cause of this poor range, and until now I have not found a solution to our problem, below are some screenshoots of our PCB, compared to the TI PCB, it would be great news if anyone could help us figuring the problem that is causing this poor range

Thanks in advance

  • 0
    TI__Guru** 106705 points
    Hi Alex,

    For CC26xx you must use the LFB182G45BG5D920. Additionally you have some other weaknesses in your design that might affect the RF performance:
    - The balun should be placed closer to the CC2640 and the differential traces should have the same length
    - The trace(s) between the balun and the antenna must have 50 ohm impedance (it does not look like that is the case now)
    - The chip antenna must be placed on the board edge, it should not have copper on the lower side (as shown in your layout figure above. Refer to the antenna datasheet
    - You should place more ground vias around the balun

    I can also see that the decoupling cap for VDDR-RF is not placed optimally. It should be on the actual VDDR trace, not on a quite long separate trace. The way it is now, the effect of the decoupling cap is reduce by the inductance in the trace and the ground return path.

    See these pages on our BLE wiki for general layout advice:
    processors.wiki.ti.com/.../CC26xx_HW_training_Layout_Considerations.pdf
    processors.wiki.ti.com/.../CC26xx_HW_Checklist

    Cheers,
    Fredrik
  • 0 in reply to Fredrik K
    Intellectual 630 points

    Thank you very much for pointing me those errors, I'll try fixing them and I'll repeat the tests again.

  • 0 in reply to Fredrik K
    Intellectual 630 points

    Hi Fredrik, thanks again for the tips, I have redisigned my board, maybe could you give a quick look to it?

    The board thickness is going to be 0.4 mm and using 2 oz as copper weight. I have used this parameters in the TXLINE 2003, in the CPW Ground, and I have managed to get 50,7 ohm impedance in the whole RF trace, excluding the traces between the CC2650 and the balun (which now is the Johanson 2450BM14G0011). Those lines are 56.7ohm and I'm unable to decrease that impedance to 50ohm, could this be a problem?

    Thanks again in advance

  • 0 in reply to alex22
    TI__Guru** 106705 points
    Hi Alex,

    This looks better.

    Can you add some more ground vias on the right side on the RF trace? The VDDR decoupling should be connected to the actual VDDR trace (you can rotate it 90 deg. clock wise). Ideally the 32 k load capacitors should be placed with ground pads towards each other.

    ~57 ohm trace impedance will be fine.

    Cheers,
    Fredrik
  • 0 in reply to Fredrik K
    Intellectual 630 points
    Thank you very much!