This thread has been locked.

If you have a related question, please click the "Ask a related question" button in the top right corner. The newly created question will be automatically linked to this question.

CC1120EM-169-RD: Problems simulating example impedance matching circuitry for CC1120 169MHz band

Part Number: CC1120EM-169-RD
Other Parts Discussed in Thread: CC1120

Tool/software:

Hey everyone,
As I am designing a circuit for the 139MHz - 160MHz band with the CC1120, I am looking at the impedance profile. However, I feel like there is important information missing on how I should calculate the impedance. As of now, not even the 169 MHz band that I tried to model as a sanity check, doesn't match up.
First of all, how does the TXRX switch work, is TXRX GND when the transceiver is in TX mode, and floating when it is in RX mode?
Talking about the 169MHz band, I tried to model the impedance at 169MHz in TX mode in the circuit simulator LTspice XVII. However, the impedance I received at 169 MHz didn't match up with the expected 80 ohms I should receive from copying the example schematics from TI.
Here is the circuit I build:

Explanation to some details:
- T
  • The Current Source on the left represents the TX output of the device.
  • I put 80 ohms as a termination resistor at the right end of the circuit in order to model an 80 ohms load impedance
  • The components used are ideal components, as this is a low frequency band, capacitor mostly have capacitive impedance, inductors have inductive impedance, and neither of their value changes at this frequency.
  • The GND which is connected to C8 represents the TXRX switch line, which is GND when the transceiver is in TX mode. 
  • I used two 72 ohm termination resistors on the top to model the impedance of the RX signals (I got that from the data sheet)
  • I did an AC sweep from 45 MHz to 245MHz as a simulation command
  • The resulting impedance at the input (TX signal, current source in the schematic), should be around 80 ohms. It is calculated by dividing the input voltage by the input current
The graphed result I receive, the x-axis represents frequency, the right y axis represents Reactance, and the left y axis represents real resistance.


Here the impedance specifically around 169MHz:


As the impedance is going to be Z = sqrt(Real^2 + Imag^2), this is a bit too far off from 80 ohms total impedance.
Therefore, my question. Did I not implement something correctly from your circuit, in which case I would hope for you to correct and guide me upon this, or is it designed to be that way?
Also, what would be the considerations for modeling the circuit in RX mode then?
Thank you everyone for helping!
  • Hi,

    We recommend following the reference design: CC1125EM-169 (SWRR100): https://www.ti.com/lit/zip/swrr100

    This reference design accounts for the factors discussed in Section 9 of SWRA640 (CC13xx/CC26xx Hardware Configuration and PCB Design Considerations): https://www.ti.com/lit/swra640 - this also also relevant to why we recommend the reference designs for all of our Wireless Connectivity devices.

    The following thread is relevant here: https://e2e.ti.com/support/wireless-connectivity/sub-1-ghz-group/sub-1-ghz/f/sub-1-ghz-forum/740905/cc1120em-169-rd-cc1120-matching-circuit-for-145-mhz

    You likely do not need to go to this extent to get the desired performance, especially as the PCB parasitics have not been simulated. Testing the initial design at the frequency band of interest and performing some bench tuning (if required) is likely more efficient for this application.

    Regards,
    Zack

  • Hey Zack,

    Thank you for your response. I looked at the reference design you sent me, and it seems to be the identical circuitry I found/included for the CC1120 circuitry, therefore this wouldn't change the simulation circuitry. Moreover, I am aware that simulation might not be the most optimum form of determining my circuitry, however, I want to remind you that my team wants to use a 2m band (140MHz - 146MHz) therefore I need to simulate whether I might need to do some fine tuning for some components in order to reduce insertion loss, as there is only an example schematic for the 169 MHz band. Additionally, we will be using a 50 ohms impedance antenna and traces (not as the recommended 80 ohms), therefore, there could be more values to change. But as of now, I am just trying to verify the 169MHz band, so that from that point we can continue tuning the impedance matching circuit. And again, as you mentioned I haven't modeled other PCB parasitics, but as of now, the simulated 169MHz band seems to have an impedance of 64 ohm, hence, will the PCB parasitics add the missing impedance of almost 20 ohms?

    Best,

    Mohammed

  • Hi Mohammed,

    As a starting point to see how the parasitics can affect the impedance matching you can use component vendor models for the passive components (Murata or another vendor) - S-parameter measurements should be available for the relevant component series at least.

    We do not provide load-pull data for this device (and not at 140-146 MHz), so manual tuning is likely necessary regardless.

    To simulate (more) accurately, you would need to simulate the PCB layout and include vendor component models. Regarding the TXRX pin: https://e2e.ti.com/support/wireless-connectivity/sub-1-ghz-group/sub-1-ghz/f/sub-1-ghz-forum/607375/cc1120-lna-information-tx-rx-switch 

    Regards,

    Zack

  • Hey Zack,

    Again, thank you for your response. I understand that in the end I will probably need to hand tune to some extent, but again, even if I simulate the PCB parasitics and the put in actual components I am unsure whether I would achieve these 80 ohms. But that doesn't matter, I guess I will just have to deal with it. However, before going ahead with this procedures, I would like to ask you if you could take a look at the circuitry and if you notice something odd, or something which should be different, so that I can at least make sure that I am not getting wrong results because of an implementation error, e.g. whether using 80 ohms termination resistor to model an antenna is fine or not. And also, did TI determine this circuitry by testing an actual physical product, by simulation, or by calculation?

    Best, 

    Mohammed

  • Hi Mohammed,

    TI did all of the above to determine the final circuit values and the results were included in the datasheet because it underwent characterisation at these frequency bands. However, I cannot provide the simulation or load-pull data for this.

    You will want to match to 50 Ohms for your setup, not 80 Ohms. You will also want to match to the impedance of the antenna at a later stage once the conducted performance has been verified, transforming from 50 Ohms to the antenna impedance (probably by including pads for a Pi-network), so at this stage of development you would match to 50 Ohms at the output (R3 in your setup).

    C1 in your setup (C173 in the TI reference design) is listed as DNM.

    The effects of the layout could be significant and without accounting for this it is difficult to say what the results should be here. The final values were fine-tuned on the bench.

    The RX path would not "see" 72 Ohms when in TX operation; the (70 + j20) Ohms target source impedance from the datasheet is only for when the device is in RX operation. The RTX pin is pulled low (to GND) in TX operation.

    I am not very familiar with LTSpice but we would normally perform an S-parameter simulation using Keysight ADS or a similar EDA tool (for context) - for LTSpice you could try using a voltage source with the target load impedance?

    I would strongly recommend testing with the 169 MHz circuit on the bench to see what the performance actually is at 140-146 MHz (the boards looks to be in stock on TI.com) and in the meantime check what frequencies the notch filters are targeting (in your circuit that would be L2/C5 and C7/L4):

    • It looks like L2/C5 target the 2nd harmonic, which at 169 MHz would be 338 MHz. Therefore, this would need to be adjusted to resonate at the  2nd harmonic of the centre frequency of your band of interest (i.e. 2f0 = 286 MHz). I would suggest starting with a C5 value of 18 pF (if using vendor models this targets the 140-146 MHz band).

    The layout could affect the notch filters in particular, so this might need to be tuned on the bench if the harmonics violate the spectral mask with the original BOM - this is especially true if you do not simulate the PCB parasitics (as your simulations will be less accurate). I don't think that retuning the notch filters using only  C5 and C7 will affect the target load impedance much, however, only the harmonic attenuation.

    Regards,

    Zack