Maker Pro
Maker Pro

RF balancing (50 ohm traces, baluns etc...)

24Volts

Mar 21, 2010
164
Joined
Mar 21, 2010
Messages
164
Hi,

I am using an RF MCU chip which comprises
an RF transceiver and an MCU.

The company that sells this RF chip provides a recommended RF circuit schematic which shows how the device should be wired. The schematic shows the RF MCU along with some external coils baluns and capacitors.

I also purchased an RF kit which has two of these RF MCU’s. The distance of the kit’s communication was about 1000 feet.

I had the RF circuit replicated and when I tried it it seems that the distance is shorter than their kit’s distance. I seem to only communicate about 200 feet.

The company that sold me the RF kit (which also provided the MCU-RF schematics) said to me that every time an RF circuit is replicated there needs to be some small adjustments made to the RF traces portion of the circuit.

My question is, what are these adjustments exactly that need to be done? Thanks
 

kellys_eye

Jun 25, 2010
5,092
Joined
Jun 25, 2010
Messages
5,092
Schematic? Layout?

What frequency are you talking about? I'm presuming 800+MHz where circuit board traces become important and tracks can act as inductance/capacitance etc therefore need to be taken into account for correct power transfer....
 

24Volts

Mar 21, 2010
164
Joined
Mar 21, 2010
Messages
164
Schematic? Layout?

What frequency are you talking about? I'm presuming 800+MHz where circuit board traces become important and tracks can act as inductance/capacitance etc therefore need to be taken into account for correct power transfer....

Hi Kellys_eye,

Thanks for replying,

It’s 900 MHZ...

I could post the schematics if you want, but what I don’t understand is what is it that they would do when adjusting.... ?

Do they just adjust the thickness of the traces or do they also alter the components ???

Also, how do they do this.... ?
Do they mesure the impedance and calculate it?

And if so how do they calculate it ? And is there any special tools to do this? I have heard of the smith charts... how do we use the Smith chart?

And is it complicated if I wanted to do this myself? Is it even feasible for me to learn how to do this or is this something that is better handled by the layout/board firm company?

So many questions....Confused!!!
 

davenn

Moderator
Sep 5, 2009
14,075
Joined
Sep 5, 2009
Messages
14,075
I could post the schematics if you want, but what I don’t understand is what is it that they would do when adjusting.... ?

Do they just adjust the thickness of the traces or do they also alter the components ???

Also, how do they do this.... ?
Do they mesure the impedance and calculate it?

And if so how do they calculate it ? And is there any special tools to do this? I have heard of the smith charts... how do we use the Smith chart?

And is it complicated if I wanted to do this myself? Is it even feasible for me to learn how to do this or is this something that is better handled by the layout/board firm company?

So many questions....Confused!!!


well since you still didn't provide the details that Kellys_eye asked for it's almost impossible to advise you

schematic,
board photos,
link to web site(s) where the gear came from
 

kellys_eye

Jun 25, 2010
5,092
Joined
Jun 25, 2010
Messages
5,092
As with any tuned circuit the output stage (filter) has to be tuned correctly and component tolerances can make a difference.

Ideally you would monitor the input and output signals and tune the circuit for maximum pass (resonance) and, better still, use a sweep generator to produce a 'live' display of the passband and adjust components for the best result.

When you get towards μwave signal frequencies the actual copper traces have appreciable inductance and capacitance - the values of which vary according to width, length and substrate material parameters. This is very esoteric work and can't be guessed at but requires sophisticated test/measurement equipment to perfect.

However YOUR device sounds very much like LoRa-based systems and tuning should be possible just using the correct value discrete components and/or adjusting the transmit channel frequency to the centre of the passband of your output filter.

If your device CAN be tuned for frequency, try going from one end of the band to the other and see where you get your peak output.

If you change one (more likely a PAIR) of filter capacitors (or inductors) by one level of value (i.e. next in range) you could retune the band and see if there is any improvement. If not then change the component(s) one level of value in the opposite direction, rinse and repeat until you're happy.

It's rather brutal but without the pro test equipment it's about the best you'll manage.
 

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
Sometimes you have to make sure you match the trace impedance to the load. This is normally done by the PCB manufacturer. But you have to tell them to make a particular trace say 50 Ohms. This will ensure maximum power transfer for that particular output. We need to see the PCB design files if you have them.
Adam
 

24Volts

Mar 21, 2010
164
Joined
Mar 21, 2010
Messages
164
Hi folks,

Sorry for the delay!

I am using the erf32fg1p133f256gm48 RF MCU from Silabs and here is the kit I bought:

https://www.silabs.com/products/dev...gecko-915-mhz-2-4-ghz-and-sub-ghz-starter-kit

I have been reading up on RF tuning and balancing. I have attached the actual schematic and the RF side of the PCB layer. In reference to the attached files and the sub GHZ matching part of the RF circuit, here are some questions:

1) Is the balun there to balance the RF signal wave? I assume it means that the AC wave is balanced when the output of the balun SE (1) is opposite from the output GND (5). Is my assumption correct?

2) Is L6, L7, C6, C7, C8 are used to fine tune the 50 ohm impedance signal?

3) What is C3, C4,C5, L3, L4, L5 used for?

So basically, I had a firm draw out the layout and had another firm build the PCB board with all the components. The RF data transfers worked meaning I was able to send data from one RF module to another. However, the kit was able to go at least 1000 feet. The one I had replicated could only go about 250 feet.

One of the many things I don't understand, is that the sub-GHZ part was 100% replicated according to the manufacturer's gerbers/layouts/schematics and sepcs. The only thing I added was one connector and one micro switch on the opposite side of the RF PCB layer. So how could there require some fine tuning/Matching of the RF components given that everything copied perfectly??

So what typically needs to be done to to resolve such a distance degradation?
In other words, the person correcting the layout... what is he/she going to do to figure out what changes have to be done to this board in order to tune it so I can communicate 1000 feet? And how will they do that?

I have heard that if the traces on the PCB of the RF circuit aren't wide enough, the manufacturer will need to adjust them. To do this, what does he need to do? Does he need to measure their ohmic values, if so, how?? Is there special equipment to do this?

Since I am not creating the board and if there is some trace(s) that needs to be widened (or stuff like that), this i out of my control, so I don't know if its even worth it for me to dig into this stuff if tuning the RF portion of the components is something that the board layout/manufacturers typically do?

In my case, should All the RF fine tuning and impedance matching be left to the firm building my board or is this something the RF chip manufacturer needs to do(silabs), or is it something I should do?


Pretty confused huh :-(
Thanks for your help and replies
 

Attachments

  • RF_Tuning_Schematic.jpg
    RF_Tuning_Schematic.jpg
    62.8 KB · Views: 104
  • RF_Tunning.jpg
    RF_Tunning.jpg
    138.3 KB · Views: 71

Arouse1973

Adam
Dec 18, 2013
5,178
Joined
Dec 18, 2013
Messages
5,178
The first matching circuit is used to match the output of the device to the Balun. The Balun then converts the double ended balanced signal to an unbalanced signal both inputs and outputs need 50 Ohms matching. The other inductors form a PI filter and you would use them to match also.

It's important to have the correct PCB stack up which means the thickness of the PCB and the order of the layers especially the 0 V plane be the same distance from the feed line feeding the antenna as the original PCB design.

This can be done as shown using the top plane running both sides of the feed line or in some case a reference plane underneath the feed line on the other side of the PCB.

Thanks
Adam
 
Top