Reply
Posts: 8
Registered: ‎01-30-2016

Inexplicable parameters in AN732

       In the Figure 2 of AN732 - Si4438 TX MATCHING , C0= 6.8p, L0=24nH, I calculate the resonate frenquency F0=394MHz, but it's working frequency band  noted following the Figure is 470~510MHz,It is inexplicable.Someone can tell me the reason? and the components parameters of the matching cicuit (from CM to CM3) are also inexplicable. It looks as cannot match the impedence from 24.87 + j28.66 to 50 ohm.why?

Posts: 127
Registered: ‎01-06-2015

Re: Inexplicable parameters in AN732

Hi,

 

The final L0 and C0 values in the matching should not create a resonance at the fundamental frequency, but at a lower one. If you take a look at Figure 33 in AN627 (page 73):

https://www.silabs.com/Support%20Documents/TechnicalDocs/AN627.pdf,

you can see that the original L0-C0 is followed by an LX-CX element, which makes the impedance match to 50ohm. The orignal L0-C0 (which is in Figure 33) should resonate at the fundamental frequency, but the final L0 value that you can see in the schematic is a sum of the original L0 and LX, this way the calculated resonance frequency is lower than the fundamental. CX and CM are also added together, resulting a final CM value which is generally higher than the final parallel capacitor value in the filtering network (CM3).

 

Daniel

Posts: 8
Registered: ‎01-30-2016

Re: Inexplicable parameters in AN732

Dear dasimon :

      Thank you very much for your  reply ! But  I  have different opinions.The circuit structure of  Figute 2 in AN732 is different from the circuit structure of Figure 33 in AN627. the  main different item is the sequence of C0 and L0, In Figure 2 (AN732) , C0 first ,  L0  followed , but in Figure 33 ( AN627), L0 first , C0 followed. so they are different design ideas.so I think that  the schematic in Figure 33 ( AN627)  can't used to interpret the schematic in Figure 33 (AN732)。

Posts: 127
Registered: ‎01-06-2015

Re: Inexplicable parameters in AN732

[ Edited ]

Hi,

 

Generally both sequence of L0 and C0 generally means similar performance, the difference is within 0.1 dB in output power. In our application notes the first element after TX pin is C0 so that it is easier to explain how L0 and LX can be combined together.

 

Daniel

Posts: 8
Registered: ‎01-30-2016

Re: Inexplicable parameters in AN732

Thank you for your reply !  I will recalculate the parameters according to your idea.

 

boom

Posts: 127
Registered: ‎01-06-2015

Re: Inexplicable parameters in AN732

Hi,

 

In case your application will operate at a standard frequency, which we make reference designs for, you do not need to calculate parameters. The reference designs are fully tested and validated, their aim is to help you starting your own design. This way, if you copy the schematic and layout of the RF section, apply the general layout rules, and your PCB substrate is similar to the reference design in terms of number of layers, then you should have the same performance as the reference design has.

 

Daniel

Posts: 8
Registered: ‎01-30-2016

Re: Inexplicable parameters in AN732

Unluckily,Silabs does not provide parameters at the frequency of 433Mhz.Many parameters should be recalculated and simulated by myself.

 

boom

Posts: 8
Registered: ‎01-30-2016

Re: Inexplicable parameters in AN732

As shown in Figure 2 of AN732, C0= 6.8p, L0=24nH , but on page 14,  C0 = 6.81 pF  L0+LX = 32.15 nH . How to interpret the disaccord of L?

 

boom

Posts: 127
Registered: ‎01-06-2015

Re: Inexplicable parameters in AN732

Hi,

 

On page 14, the values are calculated ones. The final values that are included in our reference designs can differ from the calculated ones due to the fact that the calculation does not take into account the parasitics.

 

For 20dBm 433 MHz, you can use the following reference design:

http://www.silabs.com/Support%20Documents/Software/4463-PCE20C460.zip

 

This matching is developed for 460 MHz, but 433 MHz and 460 MHz are quite close to each other, it should work properly at 433 MHz as well. Although the design uses Si4463, the same matching can be used for Si4438.

 

Daniel

 

Posts: 8
Registered: ‎01-30-2016

Re: Inexplicable parameters in AN732

Hi, 

    I think the example (http://www.silabs.com/Support%20Documents/Software/4463-PCE20C460.zip) can not be used directly by my design. Because it works in Class-E Split TX/RX mode while my design works in Direct Tie Operation mode. How to migrate the design of Split Tx/Rx mode to 

a design of Direct Tie mode? 

 

Boom

Posts: 127
Registered: ‎01-06-2015

Re: Inexplicable parameters in AN732

Hi,

 

Have you already designed the Direct-tie layout, and the PCBs are ready, so you are only waiting for component values?

 

We have a preliminary Direct-tie matching for 20 dBm 434 MHz, but it is not full validated yet, this way we cannot guarantee PA long time reliability. Still, if you need that one, I can send you the schematics, but it is recommended to use the 4463-PCE20C460 switch design instead, if it is not a problem for you to make a new layout design.

 

Daniel

Posts: 8
Registered: ‎01-30-2016

Re: Inexplicable parameters in AN732

Yes,  I have designed the Direct-tie layout, and the PCBs are ready. only waiting for component values.

 

Boom

Posts: 127
Registered: ‎01-06-2015

Re: Inexplicable parameters in AN732

Hi,

 

Please find attached the preliminary schematic for 20dBm Direct-tie matching at 434 MHz. Please note that this matching is not validated yet, this way we cannot guarantee long term reliability. Although the radio in the schematic is Si4463, you can use the matching for Si4438 as well.

 

Daniel

Highlighted
Posts: 8
Registered: ‎01-30-2016

Re: Inexplicable parameters in AN732

Thank you very much! I will  use the component values to test my design.