CC2500 Inverted F Antenna - Design Review

Hi Chris,

Can you post your design so we can look at this and what is the PCB thickness?

If you have any mechanical surrounding, batteries or hand effects; this will bring down the resonance frequency so make sure that you are measuring the resonance of your antenna as the product would appear in final production (i.e. with batteries, display etc).

Regards,   Richard.

Hi Richard,

Please see the attached gerber files. It is designed to be fabricated using 0.030" FR4 using 1 oz. copper.

The initial testing I did was using this board, with a semi-rigid cable solder to the feed and ground connected to my NA. In open air, with the board parallel to the ground ~6" from my work surface the resonant frequency was ~2.67-2.7GHz. When I placed it in housing, depending on the distance to the plastic and orientation it shifted down towards ~2.4GHz.

My primary issue/concern is that I'll be having it tested for FCC modular approval, and when tested the results will be different then when installed into end products. This is likely an issue with any module as the resonant frequency will inherhenitly shift.

I did my best to follow the antenna layout as provided by TI (DN007), it just seemed odd that when compared using the NA the resonant frequency between mine and the design of the antenna evaluation kit were off by ~200-250MHz.

Thanks,

-CC

I downloaded the design files for the DN007 antenna, which can be found in the CC2430DB reference design.  I noticed two differences between your design and the reference design.  The DN007 antenna is intended to be used with no bottom layer ground plane, while you are using a 2-layer board.  And, your PCB is much shorter that the CC2430DB reference design.  From personal experience I have never really noticed a big impact from using different PCB thicknesses.  In fact, DN007 says that this should have minimal effect.  But, I have noticed that a smaller ground plane generally has a negative impact on the impedance matching.  These two things aside, what I notice the most about your board is the top layer ground plane.  I see in your top layer that part of the ground plane against the antenna has been removed (see 4382.IsoPro2.pdf , your top layer is blue and the DN007 top layer is red).  This missing bit of ground plane on your top layer copper causes the distance between the antenna and the ground plane to increase, thereby reducing the capacitance between the two.  And, In resonant systems, a decrease in C results in an increase of resonant frequency.  I would suggest that you place a piece of copper tape over that part of the ground plane, (if you want, you can scratch away the solder mask and solder the copper tape to that part of the board) and see if that doesn't bring your resonant frequency down to around 2.4 GHz.

As far as your concerns with end products, the antenna design and tuning is best done in the end product, but when that is not possible, you just do what you can.  Most everybody who has used laptops is aware that some laptops are better than others at picking up the wifi net.

sparkchaser

Thanks for the feedback - much appreciated.

I looked at the CC2430DB again, and as far as I can tell the bottom layer is a ground plane... perhaps I'm missing something obvious. I do agree that the physical size is quite a bit different then the reference design. I'll add that copper back to the top layer. That should was the result of the EMI shield pads being pulled back.

Further testing over the last few days, has shown that the reference design antenna from TI and our design produce very similar results. Once installed in the housing, the resonant frequency slides down closer to ~2.45GHz which is certainly acceptable.

Again, thanks for your input on this. Its greatly appreciated.

-CC

Chris,

The reason you'll see differnces in resonant fequency is due to small changes comming from the capacitance of the radiator to the ground plane.  The radiator on standard 1.6mm (62mill) board is 705 mils.   This divided into 2.45GHz or 3.45MHz per mil.   The capacitance for at 2.45GHz is in the pF.   The more important number is did your antenna efficiency change.  This is the measurement of power out divided by power in.     What effects this is large changes in the ground plane. 

For small changes in frequency you can adjust the short stub to re-center for 50 ohm impedance.   Using new antenna impedance value at 2.45GHz and rematch the antenna input to you balun/filter output.   At the most you will only have a 2-3 % change in efficiency.    This is less than 0.01 dB change in power.  If your antenna is not getting the same range as TI reference kit,  then you have more serious problem and most likely is related to the ground plane size. 

You can place the reference design ground plane on larger ground planes by using some layout techniques which force the grounds impedance to be different at 2.45GHz therefore keeping the omni-directional pattern and efficiency number.   Your first step is re-calculate your match and do some measurements.

Rgds,