Author Topic: Using Pulse W3006 Dualband Wifi Antenna with a CPW-G Trace  (Read 1725 times)

0 Members and 1 Guest are viewing this topic.

Offline jacksextonTopic starter

  • Newbie
  • !
  • Posts: 4
  • Country: ca
References:

Pulse W3006 Dual Band Chip Antenna: http://www.compotek.de/fileadmin/Datenblaetter/Antennen/9_SMD/Pulse_W3006_ds_2.2.pdf

Calculated Strip Line:




We (by we 3 I mean undergrad university students in Electrical/Computer-Systems Engineering) are designing a custom PCB which uses a RF-controller with a single RF-out capable of doing 2.4G/5G/Bluetooth as part of a larger custom PCB/micro-computer. This question has 3 parts:

  • If we are using an antenna that is around 50 ohms and a Broadcomm Chip (Broadcomm WM-BAN-BM-13) that also has RF outputs referenced to 50 ohms, we only need to design a trace that has an impedance of 50 ohms with little to no impedance matching network (such as dual pi-matching set-ups for resonating at both 5G/2.4G). Or should there still be a coupling capacitor/series inductance at the end even though everything is around 50 ohms?
  • With the calculations shown in the image for the PCW-G transmission line, what is the purpose of the Physical Length (L) and Electrical Length? From my understanding, the electrical length of the trace measured in wavelengths.
  • Since the chip needs to operate at 2 given frequencies but we can only have a single length for the trace, how would I compensate for 5 Ghz being almost double in terms of frequency which impacts the wavelength and length calculations?

Thanks in advance.
 

Offline whollender

  • Regular Contributor
  • *
  • Posts: 58
  • Country: us
Re: Using Pulse W3006 Dualband Wifi Antenna with a CPW-G Trace
« Reply #1 on: May 07, 2016, 11:07:05 pm »
Quote
If we are using an antenna that is around 50 ohms and a Broadcomm Chip (Broadcomm WM-BAN-BM-13) that also has RF outputs referenced to 50 ohms, we only need to design a trace that has an impedance of 50 ohms with little to no impedance matching network (such as dual pi-matching set-ups for resonating at both 5G/2.4G). Or should there still be a coupling capacitor/series inductance at the end even though everything is around 50 ohms?
Assuming both the antenna and broadcom chip are both matched (~50 Ohms) at the frequencies of interest, you shouldn't need any matching network(s).  The most important thing is to follow the manufacturer's recommendations for matching the components. 

For instance, the datasheet for the antenna shows a 50 ohm CPW feed, with an additional 1.5nH shunt inductor to help the match, and the ground plane(s) cleared from underneath the antenna itself.  It also mentions that the exact inductor value will change based on the specific application, which I take to mean different board materials, surrounding components, etc.  If you aren't too worried about the antenna's performance, then you may be able to get away without it, but I would include the pads anyway so that you can play around with different inductor values if you need better antenna performance.

I couldn't find any information on the broadcom chip, but you should try to follow the datasheet recommendations or try to copy an existing layout, assuming that you have a datasheet or existing layout available.

Keep in mind that the shunt inductor that the antenna datasheet recommends will pull the center conductor to ground, so you may need a DC blocking cap between the Broadcom chip output and the antenna if the broadcom chip puts a DC voltage on that pin.

Quote
With the calculations shown in the image for the PCW-G transmission line, what is the purpose of the Physical Length (L) and Electrical Length? From my understanding, the electrical length of the trace measured in wavelengths.

The physical length gives you the actual length (in mm or inches) of the CPW trace that corresponds to a given electrical length in degrees (or fractions of a wavelength).  This is important if you are designing a matching network that requires a shorted or open stub of a specific electrical length at a given frequency.  For connecting two matched components, it doesn't really matter.

Quote
Since the chip needs to operate at 2 given frequencies but we can only have a single length for the trace, how would I compensate for 5 Ghz being almost double in terms of frequency which impacts the wavelength and length calculations?

Because both of the components you are connecting are pretty well matched to the CPW line, there's nothing that you need to compensate for.  You will have slightly more loss around 5GHz than around 2.4GHz, but that shouldn't be a problem.  I would just recommend keeping the RF trace as short as the layout allows.
 
The following users thanked this post: jacksexton


Share me

Digg  Facebook  SlashDot  Delicious  Technorati  Twitter  Google  Yahoo
Smf