50 Ω trace - difference between TXline calculations and reference design (CC2500)

Thanks for the quick response. I would be great if I could get clarification on the very close similarity between the two reference designs which currently confuses me a bit.

I checked again and there is a 1 mil difference in spacing between the ground plane and the feed line itself between the different designs. Also, another difference is that the 0.8 mm board has this spacing also between the ground plane and the line where it goes from the balun to the matching network. The 1.6 mm design just has 10 mil spacing here so maybe these differences are enough.

Emil,

Since layout are compac and wavelengths are far shorter than 1/4 wavelength traces between components are not 50 ohms.   Instead they are used to fine tune the circuitry.   TI only calls out lines which are 50 and you should reconginze the height to width ratio means a 2 : 1. for given heigth above ground and width of trace.   So in order to utilize board space many of the traces are not 50 ohms but inductor values which tune the final design.  Keep in mind it is very difficult to get a discrete component with the resolutioin to tune these circuits.    So by using the traces you can fine tune the balun and matching filter network.  You only need 50 ohm traces to SMA or antenna inputs.   It is difficult to get those lines 50 ohms for all frequencies and generally tune to only one frequency as in your case 2.45Ghz.   To get a perfect match you must adjust values to within 100nH and 100ff. 

Regards,

I have a question,

 i'm redesigning a circuit  when using the CC2500 and the SMA connector for rx and tx trasmission a 2.45GHZ,  the circuit works well but with txline calculator I calculated the impedance between the balun and SMA connector with the method CPWG getting a impedance of 80 ohms;

data analysis Tx calculator

-Phisycal Lenght: 178mil

-width trace: 10 mil

-gap betwen trace and Ground: 10 mil

-height:  PCB 75.9843 mil

-thickness: 1.4 mil

-dielectric costant: 4.7

-conductor: copper

-FREQUENCY 2.45ghz

if I change the impedance to 50 ohm with  trace width of 50 mils will see significant improvements in rx and tx of data?

Reguards.

- T line calculators uses ideal equations and are only fairly accurate for long lines

- Normally a line has to be longer than about 1/10 wavelength to be treated as a transmission line. If I don't misread your data your line is 178 mils or 4.5mm long. The wavelength for 2.4GHz is 125mm and hence your line is too short to be counted as a transmission line.

Using calculators will give you an idea for the required dimensions but for short lines and/or for more accurate estimate a EM simulator should be used.

But in the reference design cc2500 the pcb is 031mils and 062 mils; the trace between baun and sma connector is 230 mils why the trace is not 125mm?

Hi,

With a trace thickness of 1.4 mm; then the characteristic impedance for a 0.8 mm (32 mil) thick PCB is approx. 51-52 ohm without taking into account the solder mask. When applying the solder mask, then the impedance will be lowered slightly.

For a 1.6 mm thick pcb, the trace width should be theoretically increased to 3 mm but this is impractical when the SMA pad diameter is 1.8 mm and the trace length is only 5 mm long. So for practical reasons, the width of the trace was kept the same as the 0.8 mm thick pcb design.

Regards,

   Richard.

Thanks for your response and your solution, my scenario is show in  picture.

 and applying the solder mask, I'll have a 50 ohm impetenza as suggested by Richards Wallace. I think that the setting is correct AWR?

Looks fine.

This is great freeware from AWR.

Major parameters that effect the characteristics impedance is the height (H), frequency, width (W) and then the dielectric constant. The gap (G) becomes more crucial when G approaches H or is smaller in size. The thickness is not too sensitive but this is normally in the region 17um to 35 um for "standard pcb".

Best regards,

   Richard.

Hi,

There is a couple of ref designs based on the CC2500 with contains the schematic, pcb stack-up, layout and BOM.

CC2500 EM Discrete based Ref Design: http://www.ti.com/tool/cc2500em_refdes

CC2500 EM Folded Dipole Ref Design:http://www.ti.com/tool/cc25xxem_ref_des

Regards,

   Richard.

Hi Richard

I'm reviewing a new wireless device , that device is constituted by the CC2500 radio chip and antenna dongle dongle . the antenna nano dongle (Design / Application NoteAN043 * 2). The antenna if the wireless device stand alone and wireless transmitter are tuned to the same frequency as I carried out the measurement by means of the spectroscope ,

my problem on my stand alone device is that it has trouble receiving the messages you send , it has a strange behavior :

sometimes receives messages and other times I discard messages such as

if I place the wireless device stand alone 3 meters from the transmitter and sending 3 sequences of 50 mmessaggi happens to me

on which :

first 50 messages 48 m discards them

50 messages he receives 15 seconds then discard the subsequent messages 20 and 15 final messages they receive

third 50 posts discards 46 and the last 4 receives them

This situation also happens to 5 meters and 7 meters reception becomes almost nothing in practice messages? 2 to 50 with an RSSI of - 84dbi.

I have tested the radio circuit board with radiotest texsas, using a test module with CC2500 on the first board and the second board I connected my RF module stand alone. With my rf module stand alone I can send packets to 5000 packets I lose only 400 with an RSSI of -74.3 dBi. While receiving messages if I perform with my rf module stand alone in 5000 * message I receive only 45.3%.

I do not know what could be the problem the balun work perfectly is the drawing of texas and also the components are those recommended by texas. The problem could be due to some component that is not correctly positioned seconto specifications texas or disorders may be due to the noise of the gnd and / or the positive power supply?