• State Resolved
  • Locked Locked
  • Replies 9 replies
  • Subscribers 30 subscribers
  • Views 3739 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.

Copper thickness for manufacturing of CC85XX-CC2590EM

Torje Nikolai Thorsen
Intellectual 520 points
Other Parts Discussed in Thread: CC2590

Hi,

We have been using the CC85XX-CC2590EM reference design in our products for a quite a while now. Our RF boards are exact copies of the reference design.

We have been using two different subcontractors for the manufacturing of these boards, let us call them subcontractor A and B. It turns out that the RF performance of the inverted F antenna when using boards from A are way better than when using boards from B. The differences are quite extreme, even though both subcontractors have been given the exact same gerber files and specifications. When using master and slave boards from B, drop-outs are constantly occuring, and the system is useless. When using master and slave boards from A, no drop-outs occur at all, even when trying to provoke it by all possible means.

I am now trying to figure out how this is possible, and how to avoid it. My current suspicion is that the two manufacturers are not using the same dielectrics and/or not interpreting the manufacturing specification the same way.

Question 1:
Is the outer layer copper thickness crucial for optimum RF performance?

Question 2:
As stated in the CC85XX-CC2590EM readme file, the outer layer copper thickness should be 35 um. Is this the non-finished copper thickness or the end thickness after plating?

Question 3:
Outer layer plating is not specified. What is the optimal type and thickness of the plating?

Question 4:
Can an inadequate choice of dielectrics type and/or thickness cause the effects described above?

Best regards,
Torje

  • 0
    Genius 3055 points

    Hi Torje,

    the outer layer copper thickness or plating is not really crucial for the RF performance. I could imagine it having influence doing ultra fine performance optimization to get 101% from the chipset but is not the case you describing.

    The inadequate choice of the dielectric has an influence on impedance of your RF traces.

    The prepregs dielectric constant and thickness between top layer and the next ground plane and the distance between RF track and ground on the top side is crucial to keep the right track impedance and must be kept exactly as designed. You can easily recalculate and adjust the RF path geometry to any PCB stack your PCB supplier uses.

    You must keep the RF path having 50 Ohm impedance. The RF path used in PurePath Wireless application is called coplanar waveguide with ground and you can adjust the geometry playing with the online calculator from chemandy electronics http://chemandy.com/calculators/coplanar-waveguide-with-ground-calculator.htm - I used it a lot of times and it gives very good results.

    Cheers,
    Tomasz

  • 0 in reply to Tomasz Chumiecki
    Genius 3055 points

    The RF components on the RF path - either directly from CC85xx or from CC2590 to the antenna, having also influence on RF performance. Are you using the recommended chip inductors and capacitors from TI? Keep in mind the components should be HF RF components. TI design uses high quality wirewound ceramic inductors. I suggest also using high quality low ESR capacitors in RF path. Are you populating this same components on both boards?

  • 0
    Genius 3055 points

    Torje, Can I ask why you use the TI CC85XX-CC2590EM board directly in your product – it is so easy to customize the board to fit directly the requirements of your product. I designed plenty of different boards with CC85xx/CC2590 and never have any problem with the performance. I even think the PurePath Wireless is really robust and tolerant to changes. I also built modules based on CC85xx/CC2590 using different PCB stacks, materials and thickness - 1,6mm and 1mm PCB, every time with perfect results. The only problem with RF performance I had was with the ‘Molex 47948 SMD On Ground Antenna’ – we never solved the issue just change to tiny PCB antenna. This is why I would not recommend do use the antenna with CC85xx/CC2590. Below you can see some of my designs.

    If you decide once to customize your RF board, make your own one or integrate it into your existing product PCB and you need any help you can contact me anytime at tomasz ( at ) chumiecki . com - I can help you with the CC85xx/CC2590 RF design.

    Cheers,
    Tomasz

  • 0 in reply to Tomasz Chumiecki
    Intellectual 520 points

    Hi Tomasz,

    Thanks a lot for the tips and the very useful information concerning dielectrics and copper thickness.

    One of the manufacturers has indeed changed some of the components in order to optimize the RF range. They claim that this modification increases output power with about 3 dBm. I guess this indicates that their buildup does not correspond to the one used by TI for the reference design. I am waiting for some detailed information from them about their PCB buildup.

    By the way, what about the solder mask? Do you know whether it has a big impact on RF designs?

    We are using the reference design as it is because it already fits quite nicely the requirements of our products, so we have no need for adapting it.

    Cheers,
    Torje

  • 0 in reply to Torje Nikolai Thorsen
    Genius 3055 points

    Hi Torje,

    3dBm would be a lot just changing the passive components… let us know if your supplier gets the performance optimization running.

    I do not know if solder mask influences the RF performance. I would say not, or not significantly, but I do not have so deep RF knowledge. I’m not an RF designer... I was a ham radio freak for a long time therefore I have some RF background. I am Audio- and Power Electronics designer. Designing my boards I remove soldering mask from PCB antenna because it looks cool – especially if the boards is gold plated :) I cannot prove that it has an influence on the RF performance – I don’t think so. On the other side if you take a look on precision GHz RF designs like HF front end or HF mixers of spectrum analyzers they have gold plating and no solder masks on the boards…

    Cheers,
    Tomasz

  • 0 in reply to Tomasz Chumiecki
    Intellectual 520 points

    Hi Tomasz,

    Thanks again for your opinion, I find it rather useful :)

    Cheers,
    Torje

  • 0
    Prodigy 20 points
    Dear Torje, i have seen your board with CC85XX-CC2590EM...Really very nice job!
    What do you think for 4.1 system? Is it possible to implement another channel (with 200Hz max of bandwidth) in order to transmit also subwoofer channel? Have you tried? I am interested in this possibility, if you can help me i will be glad! Many thanks. Andrea
  • 0 in reply to Andrea Faletra
    Intellectual 520 points

    Dear Andrea,

    Yes the boards on the photos are indeed very nice! It is Tomasz J. Chumiecki who has designed them, so the honour for that nice piece of work goes to him.

    As far as I know it is not possible to implement any 4.1 system using this technology. I have done some experimentation with 2.1 systems (L + R + 1 BW limited subwoofer) and 2.2 systems (L + R + 2 BW limited subwoofers), and these configurations work rather well. But I have never tried a system with 5 channels or more, and I do not think it can be easily done.

    However, I know that the PPW protocol contains a neat feature called "dual masters" which lets you stream up to four channels in full quality by using two synchronized master modules (hence two antennas) instead of one. Maybe there could be a possibility of using the same principle for streaming 5 channels or more, with some restrictions? If you are curious about that you should ask someone at TI. I do not think such a possibility is described in the user's manual.

    Good luck!

    Best regards,
    Torje

  • 0 in reply to Torje Nikolai Thorsen
    Prodigy 20 points

    Dear Torje,

    thank you very much for your infos i will ask to TI!

    Tomasz, compliment for the wonderful pcb board! 

    Have a nice work!

    Best regards,

    Andrea