CC2500 Layout

Two different ways to solve the same problem.  Both of them are good.  The goal is to get a good RF and thermal ground under the device pad while avoiding the problem of solder being drawn down open vias and away from the device. I like the approach shown on page 51 of the data sheet when using a "quick turn" pcb fab such as pcb123 where the vias through the copper pour are automatically tented and the layout tool does not permit multiple vias in a pad.

When using Altium and a full feature pcb house I use a single large pad with several vias with solder mask (tented) through it as shown in the reference design document because I believe both the electrical and thermal ground is a little better with the via directly in the pad.

I have 5 or 6 prototype designs with open vias in 4 to 6 pads under the device for a quick turn initial design and have only had one board that required additional solder down the via and reflow.. I would not use this approach for a production design.

I suggesting picking a approach that matches your pcb fab house and your layout capability.

Hi Stewart,

I want to know if you can send me a link where i can read about RF design, i downloaded from this forum a PPT about this, but i want to know about the influences of the thick of the pcb, grounds planes, layout . This is because i want to make my own board using the cc2500 but i have some problems using this chip.

Thanks.

Gastón

Hello Gaston,

There really isn't a "short course" on high frequency design and for several reasons the design of these circuits is part analytical and part empirical. This is because the parasitic values of the caps and inductors are becoming significant compared to their stated values at these high frequencies. Also the CCxxxx devices are not simply matched into 50 ohms. They require loads with complex phase values out through several harmonics to function and meet certification requirements, and these values are not provide in the data sheets. With experience and expensive CAD and lab tools it is possible and Ti has done a good job of capturing designs in reference designs that work very well.

Ti offers many different reference designs on various PCB stack ups in both 2 and 4 layers. It is generally best to start with one of these for the RF portion of the design and keep the modifications to a minimum. Most of the unique design (e.g. adding sensors, types of data, interfaces) is centered on the digital portion of the devices which is more easily modified. For the RF portion look at reference designs for devices with a similar radio to the CC2500 such as the CC2510.

If I have missed the mark on what you are trying to do or you want suggestions on how to proceed let me know what you are attempting and I can help.

BTW here is a web site that will give you a feel for small changes in PCB parameters:
http://chemandy.com/calculators/microstrip_transmission_line_calculator.htm

H

Thanks Stewart for reply,

Like you say  in this matter isn't a short course. I tried to make my own board using the Disign note DN004 also I followed  the cc2500 datasheet " pcb layout recommendations" (Pag 51 datasheet cc2500), but my board was not working. I check the signal on the pin 6 of cc2500 ( was there), i check the communication spi pins  and the communication was perfect. But there was not link between my own board and one msp430rf2500 in Access Point mode. So i believed that the problem could be the antenna, but i can't check if this is true. So my guess is that i need more bias between de top and bottom layer or could be the thick of my pcb.

Thanks Stewart

Regards

Gastón

If you can post or send me (via friends) the top side layout I should be able to help. Include the thickness from the top side to the first ground layer.

 BTW screenprint32 is a great free tool for capturing a part of the screen to a jpg / bmp file. See screenprint32.com

Chai,

Copy the reference board this is probably incorrectly stated in the applicaiton discussion.  The important point is you need enough vias to connect top layer grounds to middle or bottom layers to reduce any high frequency oscillations at 2.45GHz or higher.   The inductance of these vias need to be as low as possible.   If you experience high radio sensitivity numbers above what is quoted in the data sheet you probably did not use enough vias.

Low Power Wireless Group

Texas Instruments