Trying to figure an antenna for the Si4012 RF chip. Page 26 of https://www.silabs.com/documents/public/application-notes/AN639.pdf talks about feeding the two sides of the loop antenna through two pull-up inductors (LP and LM) which it fails to provide any values for.

There is also a Word document named Panic Button Differential Loop(Printed) Along the PCBCircumference in the 434 MHz Differential Antenna Matrix (WES0042-01) whose link I can not find anymore that presents a diagram figuring the same two inductors with "tbd" values. And they never got discussed.

Lastly, I have found this document: https://www.silabs.com/documents/public/application-notes/AN436.pdf that on page 3 (as well as throughout the file) gives a diagram and some values for a pull-up inductor Lchoke that appears to have the same purpose as the one I am asking about.

Thing is this one is for a single ended antenna while my design comprises a differential antenna. Would it be safe to take out the value from this pdf and use it on my design ? And another detail - I suppose THT inductors (as well as caps / resistors) are out of question on this side of the PCB and 434 MHz, correct ?


I still don't get how to size the RF choke. In the https://www.silabs.com/documents/public/application-notes/AN436.pdf they ate giving the example of 315 MHz. |Z_load| here is around 56 ohm and the recommended inductance for the choke is 270 - 390 nH, thus making for an impedance |Z_choke| of 534 - 771 ohm, roughly tenfold compared to Z_load.

In my design, however, |Z_load| is somewhere between 500 and 600 ohm (434 MHz differential antenna). Wouldn't it be required to pick an inductor having its impedance in the 5k - 6k ohm interval in order to actually be effective at blocking the fundamental + first few armonics ?


The value is specified as "large". The idea is to use the inductors to provide a DC bias without affecting the RF circuit. This works because an inductor has 0 impedance at DC, but increases with frequency. With a sufficiently "large" inductance, by the time you get to 434 MHz the impedance will be so large that it's effectively an open circuit.

I wouldn't say through-hole components are "out of the question". The important thing is to keep the layout as compact as possible, which is easier with smaller components. Leads may add additional inductance compared to surface mount components, but frequently that's either insignificant, or can be compensated.

  • $\begingroup$ I agree (and upvote), though I must say that at 400 MHz, the SMD inductor is simply pretty save to directly put onto the matched trace and be done with 0603 and 0402 SMD components which would add so little parasitic capacitance that you'd actually get around re-matching / verification. $\endgroup$ Apr 20 '17 at 21:14
  • $\begingroup$ @MarcusMüller Huh ? I don't get it. Those two pull-up inductors are supposed to pull each of the PA differential output pins up to Vcc. Will they still do that if laid directly into the traces ? And why would I want to it this way instead of following the design from their sample PCB ? $\endgroup$
    – kellogs
    Apr 20 '17 at 21:28
  • 1
    $\begingroup$ close to that at least makes the most sense. $\endgroup$ Apr 20 '17 at 22:01
  • 1
    $\begingroup$ From the paragraph just above the schematic: "The PA output devices within the Si4010 chip must obtain VDD supply voltage through the TXP and TXM pins themselves; that is, there must be a DC path from the VDD supply to the TXP/TXM pins." $\endgroup$ Apr 21 '17 at 11:57
  • 1
    $\begingroup$ @kellogs Making the impedance of the inductors 10x or more of the load impedance is probably a good rule of thumb: the goal isn't so much blocking harmorics as it is just keeping most of the current in the antenna (due to the relatively low impedance). If you want more detail, I suggest asking a new question. $\endgroup$ Apr 23 '17 at 1:40

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.