3
\$\begingroup\$

sx1276 lora schematic

I want to use the sx1276 LORA ic as a node that can transmit and receive at 868MHz. In order to use that band of frequency one has to use Higher frequency rf pins of the chip (rfo_hf and rfi_hf). The image i have attached is from the manufactures design reference. For my purpose I don't need the lower frequency thus I will not be connecting the rfo_lf and rfi_lf. In the image i have attached I have marked on schematic according to my knowledge (I maybe wrong please correct me) what the functions are of each section of the passive components. In order to simplify the design I have choose to use an off the shelf SAW filter rather than building one out of passive components. The off the shelf SAW filter i have choosen has an input and output impedance of 50 ohm. My question is regarding the balun and matching network aspect of the schematic. Are components I have highlighted for the balun and matching network all I need in the design before the SAW filter or will L7 fall away as I will not be using the lower frequency? Do i fit L3? I would like to have the PA_boost connected so I can have a higher transmit power (20dBm).

Any help will be appreciated.

\$\endgroup\$
  • \$\begingroup\$ Just saw this question in the feed: if you are using only the higher band take a look at SX1272 \$\endgroup\$ – Jan Dorniak Nov 29 '17 at 8:10
  • \$\begingroup\$ Just a thought 14 dBm ERP is the highest output power you can legally use in that band in Europe. See ETSI EN 303 204. \$\endgroup\$ – user94729 Sep 13 '18 at 10:45
1
\$\begingroup\$

VR_PA provides DC biasing for the amplifiers. L1 for example is letting this DC pass but is blocking the RF from coming back to this power supply. L1 and L2 are 33nH (quite a lot). C28 and C9 block this DC from going any further (it's only for the amps). Moving to the right are what I believe all bandpass filters. Each component has an effect of the impedance. Going through the whole filter, the impedance may land back in the same place, then it's strictly for filtering, otherwise it may be matching from a 50-Ohm transmission line to whatever the rf inputs and outputs of the transceiver want to see in addition to filtering. Then you have the rx/tx switches and at the end that is most certainly for matching to 50 ohm for the antenna port (although here too, it may serve for filtering as well). That's about all I know, and definitely let me know if I'm wrong with anything.

\$\endgroup\$
0
\$\begingroup\$

I am also currently trying to understand the TX path (on the LF side though) and I can tell that most of the time, the recommended schematics are full of unnecessary components (depending on what your application is), especially in RF.

About your analysis of the RX path, you got something wrong. C14, L5, C15 and C16 could be for filtering AS WELL AS matching the SX1276 RFI port (LNA). You can use a VNA to measure the impedance of that port (make sure to put your calibration plane @ C15) and then using these footprints (double L network, which is more than enough) to match the port to whatever reference impedance you need (usually 50ohms). Have a look at this AN:

www.semtech.com/images/datasheet/AN1200-16_SX1232_API_RF_ST.pdf

It explains the optimal way (different of mine) of matching the LNA (RF inputs) of the SX1200 family.

When designing such circuits, every single components have an effect on impedance matching. When some are chosen only to filter or DC bias/decouple the signal, others strictly have impedance matching purposes (I think C20 and L3 are some of those). C9, for example is used to block DC, but its value was also chosen accordingly to play a certain role in the TX matching network.

U4 (meant to be a SAW filter) and C17 are just shorts. There is no DC to be blocked here. C17 will be around 1nF which is a good RF short @ the operation frequencies of the chip. I would only put that capacitor since the 0 ohm resistor can have unwanted parasitic inductance (use thick film if you still want to put that 0 ohm for whatever reasons). Thats for RX.

TX is somehow a bit more complicated because it includes the Power Amplifiers Boost (PA_BOOST) options circuits. You can actually choose if you want to use that option or not. For this particular schematic, it seems the designer chose to use the PA_BOOST instead of the RFO_LF. And he used the standard PA for his HF band (RFO_HF).

Have a look at this AN:

http://www.semtech.com/images/datasheet/rf_design_guidelines_semtech.pdf

Page 12 reveals the purpose of most of the components. Then, the way they describe to match the port can be used if you have the right equipment. What you call the "Balun and Matching Network" circuit is not that at all. L1 and L2 are bias components that bring DC to the PA (PA_BOOST or RFO_XF). They are 100nH, so they basically block RF and let DC go through. C7 and C8 are large values that act as a RF RF ground. So, if some RF manage to get through L1 or L2, most of it goes down C7 and C8. They could also act as quick energy sources if the PA needs instant power. That said, you can see the PA_BOOST and the RFO_HF branches as two separate circuits. Nonetheless, you have to consider L1 and L2 when matching the RX/TX ports.

Here you'll find everything you need to know about the SX1276 and many reference design with Altium projects:

http://www.semtech.com/wireless-rf/rf-transceivers/sx1276/

Another note: The Pi-Networks next to the antenna connectors are only there to match the antennas to the RF switches which are usually 50 ohms. The commercially available antennas are rarely 50 ohms stock.

A+

\$\endgroup\$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.