CC1120 Impedance Tx/Rx

Hi,

please see the following post for information: http://e2e.ti.com/support/low_power_rf/f/155/t/127993.aspx#482092

Hello

The Recommended post you give is not a useful for user who do matching impedance of PA and LNA Port.

If user who is RF engineer is trying to design RF port, they are looking for RF parameter like S-Parameter.

Otherwise they are dubug through Trial and Error so many times to get RF performance.

Thanks 

Please note that the optimal load to a PA or an LNA in many cases is not the complex conjugate of the input impedance in the port in question. Using a complex conjugate match will give you a good result if the PA is linear but in this case the PA is operating as a class AB and a different method to find the optimal match is needed. What we do is using load pull and source pull measurements to find the load impedance seen by the PA that will give maximum output power with minimum power consumption and minimum generated harmonics. In addition the network needs to suppress harmonics to comply with the regulations in the region the product is going to be used.

We provide reference designs for most frequency bands to enable the customer to do fast development of the RF part without having to do trial and error.

- The optimum load impedance for the LNA is given as 30+j*30. This is a simulated value and are not verified by measurements

- The optimal load seen by the PA is around 30+j*30 for our reference designs

Dear TI,

I am engineer at NURI Telecom and  developing the module using the CC1120 with CC1190 together.

As the above the question, i can not find the impedance the PA impedance and the differential LNA impedance.

As you can know, when the manufacturers release the RF chip, they must provide the impedance for PA and LNA.

So, i think that engineers who use the CC1120 and CC1190  can develop the RF circuit and fix the impedance for optimizing the RF output and sensitivity. 

If engineers don't have the basic imformation provided manufacturer for impedance, they try on and on many times.

And i also think engineers can not debug the RF circuit. 

Can  you debug the RF circuit without the impedance whether RF chip(CC1120 or CC1190) has malfunction itself or malfuntion of RF componet(for example, mismatching the impedance)???

I think impedance may differ each freqency  170 / 433 / 868 / 915 / 950 MHz  band. 

Therefore, TI must provide the impedance for PA and differential LNA over 170 / 433 / 868 / 915 / 950 MHz ISM/SRD band.

I and user using the CC1120 and CC1190 do not want to get  the simulation value for impedance(30+j30)  but get the verified value by measurement.

Please guide the CC1120 and CC1190 for impedance a.s.a.p.

-P.S

I also have another question: does TI have the plan providing the measured data(impedance) to user using CC1120 and CC1190?

if TI don't have the plan and can not provide the impedance, i think engineer won't apply the them on their circuit.

Please note that we have developed two reference designs for CC1120 + CC1190:

For 915MHz: http://www.ti.com/litv/pdf/swra387a

For 868MHz: http://www.ti.com/litv/pdf/swra393

Using this will reduce the design time significantly and reduce risk.

Dear TER and TI,

Thank you for your response, but i think that it's not proper reply.

Your datasheet is comprised of  10 chapters.

Chapter 1 ~  4.2  are gerneral infomation and electrical characterics. / Chapter 4.3 ~ 4.4 are receive parameter and transmit parameter respectively.

Chapter 5 ~ 6 are the tools and the radio pins usage / Chapter 7 is the reference design / Chapter 8 ~ 10 are the general information

There are only three chapter about RF information.

* Unfortunely these are not the conditions for optimizing RF performance and for making the RF circuit but the measurement result on the output port(50ohm).

* If there is without the conditions(for example, PA, LNA impedance, PCB routing pattern length, width, die ground conditions), 

can you accomplish the proper result as your datasheet?

Your reference use the discrete components(L, C) for LNA circuit, but RF flatness is so bad because there is the difference of  the component tolerance, temperature effects.

So, most of RF engineer use the Balune chip fixed the insertion loss and applitude difference. (for example, JTI (www.johansontechnology.com / 0896BM15A0001)

If you don't know the LNA impedance value, can you make RF circuit and apply these kind of Balun chip to the circuits?

As you can know, impedance value is differ each frequency band. Nevertheless datasheet doesn't provide the impedance value over 160MHz ~ 950MHz.

It only provides TBD. TBD means "To Be Determined". 

I and other engieers who using TI chip think that these impedance must be determined not from user but manufacturer.

For example, if you don't know the register value, control bit(controlling GPIO port) of TI CPU, can you turn LED on or off?

Please let me know TI have a plan to provide the PA and LNA impedance value over 160MHz ~ 950MHz bands or not.

and send me impedance value verifed by TI.

I an so interested to find out really why TI doesn't provide the PA and LNA impedance value to user

We have no plans providing the LNA and PA impedance values.  Note that we have never measured these paramenters because our RF reference designs are based on optimal load found by load pull measurements.

Note that the balun chip you are referring to is made for CC1101 and other TI chips that uses the same PA/ LNA. Balun chips designed to be used with CC1120 will be available on the market shortly.  

TER said:

We have no plans providing the LNA and PA impedance values.  Note that we have never measured these paramenters because our RF reference designs are based on optimal load found by load pull measurements.

Note that the balun chip you are referring to is made for CC1101 and other TI chips that uses the same PA/ LNA. Balun chips designed to be used with CC1120 will be available on the market shortly. 

I have been involved in another thread here: http://e2e.ti.com/support/low_power_rf/f/155/t/145325.aspx 

Now I have my answer.  This is unfortunate, but fortunately I'm not scared of a little tweaking.  At least we have a reference to go from.  Should be able to figure out tuning and matching from there.

Not the best answer, but if this is what it is, it's what it is.

I have an application at 300MHz in which I would like to separate the TX and RX ports.

Has anyone at TI made progress in measuring the optimal load impedance (using load pull) across frequency for the PA?

And likewise, is there data available across frequency for the differential input impedance for the RX LNA port?

We only have data for the frequencies where we have EMs:

For LNA:

900MHz: Zopt=60+j60

450MHz: Zopt=100+j60

175MHz: Zopt=140+j40

PA (simulated values on the reference design):

170 MHz: Zopt (57 - j*4) ohm

433 MHz: Zopt (34 + j*34) ohm

868 MHz: Zopt (27 + j*34) ohm

915 MHz: Zopt (28 + j*37) ohm

955 MHz: Zopt (40 + j*42) ohm

OK, great. This is a big improvement over some of the info posted earlier.

A few more questions ...

- There is a reference design available for the 274-320 MHz range.

CC112x/CC1175 Operation in 274-320 MHz Frequency Band

http://www.ti.com/lit/an/swra398/swra398.pdf

Is the data available for this frequency range as well?

- What does this comment mean in th eabove posts?

"The optimal load seen by the PA is around 30+j*30 for our reference designs"

Is this referring to the expected antenna impedance?

It doesn't match any of the data just posted and has no associated frequency.

For the 274.320 MHz reference design I got the impression that the method was to start with the BOM for 433 MHz and scale it to lower frequency. The optimal load impedance was not checked for this design.

I wrote around 30+j30 for Zopt, 868MHz. This is basically a rounded version of the numbers I gave in the previous post, note that if you look in the smitch chart it is basically the same value.

Thanks for the clarification.

I appreciate the timely help. I am, however, going to add an editorial comment.

It is really unacceptable that a company as big as TI doesn't have basic information about port impedances and optimal loads across frequency in the data sheets for these 112x parts. These are active devices and anyone (like me) who wants to use them at a frequency and in configuration other than those of the reference designs needs this information to perform reliable RF modeling and design.

It shouldn't be guess work. There are a number of measurement companies (Modelithics for instance) who could do the load pull and RF port characterization across frequency.

Who is the right person at TI to address this issue?