PAN1026 low output TX power - certification tests

Hi. We've tried to certificate our product's Bluetooth feature under Brazilian's telecom regulation (Anatel), though during the tests, the measured output TX power while running either on TX burst or non-modulated mode is about -20 dBm other than +4 dBm as specified by PAN1026 datasheet.
This is our current scenario:

  • We've removed on-module chip antenna and soldered a short SMA connector to the soldering pads.
  • Due to our current hardware design, PAN1026 D9 antenna pin wasn't fanned out, so we aren't able to use it with a UFL connector.
  • We've read that PAN1026 has TX power fixed at +4dBm and it isn't configurable, so we're ruling out firmware setting issues.

Questions:

  1. Does soldering an SMA connector on chip antenna pads cause an impedance mismatch, leading us to the low output power?
  2. How can we workaround that scenario considering we cannot have a connection to the D9 RF pin? Is there another place we can or should have soldered our connector to get the maximum output power?
  3. Is keeping the chip antenna on a possible solution? (Unfortunately we don't have an spectrum analyser at hand to verify it)

Thanks in advance.

21replies Oldest first
  • Oldest first
  • Newest first
  • Active threads
  • Popular
    1. Does soldering an SMA connector on chip antenna pads cause an impedance mismatch, leading us to the low output power? <RT> Most likely, yes.

    2. How can we workaround that scenario considering we cannot have a connection to the D9 RF pin? Is there another place we can or should have soldered our connector to get the maximum output power?   <RT> Normally circuit changes are required when the antenna is removed. I'll post these changes in this thread early next week.

    3. Is keeping the chip antenna on a possible solution? (Unfortunately we don't have an spectrum analyser at hand to verify it)   <RT> I'd recommend that you request Panasonic to perform the certifications.  With what Panasonic Sales Company and and which sales person are you working?

    Like
  • Hi. Thanks for the quick response. I'll be waiting for the circuit changes then.
    We've worked with Future Electronics and Arrow, which both have supplied the modules to our contract manufacturer.

    Like
  • Rich Trueman Another important detail is that I believe Panasonic cannot perform the certification for Brazil's regulatory agency since we're doing that for our product as a whole and not the standalone module.

    Like
  • Like
  • Soldering an UFL connector to the antenna pad will result in the very low output power as measured -20dBm is realistic. The reason is that the antenna pin is connected to the GND of the module. For the antenna this is normal but the UFL will output only very little power. This can be changed when cutting the trace to the antenna pad as shown in the picture I posted previously. Then the UFL should be soldered close to the pad of the C22.

    Like
  • While I'm not an Anatel, expert, I can say that having the module certified typically makes getting product certification less rigorous. 

     

    Alternately, if the certification body (Anatel) will allow you to use the module's existing test results -- that were produced for other countries -- this too may ease the certification process.  Test results are available to test laboratories after an non-disclosure agreement has been executed.   

    Like
  • Ingo Wagner Thanks a million for the clear explanation. We'll try it out asap! I noticed both antenna pads are connected to the ground too late and then I couldn't figure out how to workaround. We even disassembled one module to find a common point between the D9 antenna pin and the board top layer as a plan B. Anyhow, I'll let you guys know as soon as I test it.

    1. Would soldering an SMA pigtail instead of UFL have the same result?
    2. Should I keep the antenna on and solder UFL between the C22 pad and GND or simply replace it by the connector?
    3. The only thing that is not clear is that how cutting that trace will make a difference as the pad will remain connected to GND due to the plane. Do I cut it completely to let the pad floating?
    Like
  • Rich Trueman You're correct, in fact we can "inherit" the certification in case the module had already been tested for Anatel. However, we'd have to perform two certification processes anyway, which would be too costly and take too long, a time we unfortunately do not have due to our product release deadline. We'll have to certificate it in Argentina very soon. Thanks for the help!

    Like
  • Ingo Wagner Could you please review the modifications and questions on the picture below?

     

    1. Should I cut at both green and blue traces or green only?
    2. Does the SMA pigtail soldering seem fine to you?
    3. Should I keep the antenna mounted?
    Like
  • You should cut the trace and solder the SMA cable to the PINs of C23. I mentioned C22 by mistake. C23 is the correct part where the RF output is connected. The antenna you can leave mounted when the trace is cut.

    Like
  • Anderson Felippe  I've placed a request on our regulatory department to investigate Anatel certification.  BTW: Which sales person are you working with in Panasonic Brazil?

    Like
  • Rich Trueman Hi. We've worked with Future Electronics and Arrow.

    Like
  • Ingo Wagner It's crystal clear now. Thanks again!

    Like
  • Ingo Wagner Hi. The modifications you suggested have increased the signal output power up to -4 dBm. Unfortunately, we couldn't get better than that, which isn't enough for the certification tests. It's been also very difficult to keep anything (UFL connectors, SMA pigtails etc) properly soldered on that spot since traces and pads are too small. The mechanical stress is always breaking and pulling traces out. That  brings me another question: I can see that the output pin that's fanned out of the TC35561 is the same pin as the D9 below. If I remove the radio shield and desolder every component from the output filter, would it still work with max output power or is the filter a necessary circuitry to assure the +4 dBm output?

    (I'm asking it since it looks a more reliable spot to solder the connector as there are more pins to share the mechanical stress)

    Like
  • Rich Trueman Since we haven't been able to work this issue out, I think it might be an idea for ourselves to order a PAN1026 development kit, submit it to Anatel's certification an then later on get our product to inherit that certification.

    Like
  • Anderson Felippe  Adding a u.fl connector to our modules requires precision soldering  to avoid impedance mismatching and the subsequent low power output.  I recommend that you complete your design with the PAN1026A and use the EVAL_PAN1026A kit as well.

    Like
  • Anderson Felippe  Please share the name of the Future FAE/Salesperson with whom you are working.

    Like
  • Rich Trueman For now, we've soldered a flexible 50-ohm cable with a U.FL connector which might work. Next week we'll have a result and I can share it here. However, if it fails, then we'll likely go for the evaluation kit in the next step.

    The Future Electronics salesperson which has answered me and is aware of this issue is called Ricardo Bittencourt ( ricardo.bittencourt@futureelectronics.com ).

    Like
  • Anderson Felippe If the objective is to solder a 50 ohm cable to the module, I'd say the chances of success are small.  Is Ricardo familiar with your project on a commercial level?

    Like
  • Rich Trueman Agreed. We'll have a second try for the certification this Thursday, so let's see how it'll get on. And yes, Ricardo must be familiar with our project, plus I spoke to him a couple of times about the certification issue.

    Like
  • Rich Trueman Hi. I'd like to inform you we've managed to pass the tests at 0 dBm. It'll be just a matter of updating our documentation. This was only possible because we took a look at the FCC reports for the module and the maximum peak conducted output power results were also about 0 dBm. It seems +4 dBm it's not reachable with those test.

    Thank you very much for your support!

    Like
Like Follow
  • Status Answered
  • 2 yrs agoLast active
  • 21Replies
  • 405Views
  • 3 Following