How do I design a 5 GHz circuit with RF components?

Click For Summary

Discussion Overview

The discussion revolves around designing a PCB for a circuit operating at approximately 5 GHz, specifically involving two SMA connectors and one resistor. Participants explore various design considerations, simulation tools, and transmission line technologies suitable for high-frequency applications.

Discussion Character

  • Exploratory
  • Technical explanation
  • Debate/contested
  • Mathematical reasoning

Main Points Raised

  • One participant seeks guidance on designing a PCB for a 5 GHz circuit, mentioning the use of Microwave Office for simulation.
  • Another participant questions the specific effects the original poster aims to simulate, suggesting that a realistic model of the resistor is crucial.
  • It is proposed that a coplanar transmission line (CPW) could be advantageous, as it avoids the need for vias.
  • Discussion includes considerations about the choice of substrate material, with FR4 noted as not ideal for 5 GHz applications.
  • Participants discuss the implications of using a 1 MΩ resistor at high frequencies, with concerns raised about its effective impedance due to parasitics.
  • There is a suggestion that leaving the line open might be a viable alternative to using the resistor.
  • Clarifications are sought regarding the dimensions and layout of the PCB, particularly the relationship between the resistor, connectors, and ground plane.
  • Participants express uncertainty about the behavior of high-value resistors at microwave frequencies and the importance of consulting datasheets for specific components.

Areas of Agreement / Disagreement

Participants do not reach a consensus on the best approach to the design, with multiple competing views on the use of resistors versus open lines, the choice of transmission line technology, and the implications of using FR4 as a substrate material.

Contextual Notes

Limitations include the lack of consensus on the appropriate resistor value for high-frequency applications and the potential impact of parasitics on circuit performance. There are also unresolved questions about the exact dimensions and layout required for the PCB design.

Who May Find This Useful

This discussion may be useful for electronics engineers, PCB designers, and students interested in RF circuit design, particularly those working with high-frequency applications and transmission line theory.

vaka85
Messages
8
Reaction score
0
Hi all,
I don't know if it's the right section in the forum, but here is my problem.
I have to design a small pcb working at about 5 GHz, and I don't know how to start exactly...
It's very simple, 2 sma connectors and 1 resistor. Here is the schematic:

http://img690.imageshack.us/img690/9175/immagineja.jpg

And here is what I've done sofar in Microwave Office... but in this step there are some problems for sure...

http://img80.imageshack.us/img80/8898/immaginelo.jpg

Anybody knows how to proceed in microwave office (or similar software)? I need help because this topic isn't covered so much on the web...

thanks
 
Last edited by a moderator:
Engineering news on Phys.org
What exactly are you trying to simulate? I.e. which effects are you after?

The main question is you need a realistic simulation of the resisitor (with all stray effects)?

You don't really need Microwave office for the rest, just decide on what type of connector and PCB (e.g. Rogers XXXX) you need as well as the type of transmission line (if you want to terminate one end e.g a coplanar waveguide would work, unless you really want a via for some reason, but beware of parasitics). Then you just draw the circuit, as long as you avoid sharp bends this will be accurate enough.

You can calculate the parameters of the transmission line using for example TXLine (free software, made by the same company that makes Microwave Office).

However, if you want to simulate the "full" resistor things get more complicated. I guess you are planning to use a small SMD resistor, right? Do you have a circuit model for it?
 
sorry, you're right, I've forgotten to say some things...

1. I don't have to do a simulation.

2. I have to design the pcb from the beginning. I have FR4 as substrate, 2 SMA connectors, and one SMD resistor. That's all.

The project specifics are:
- About 4.5 GHz ( Not so strictly specific, also 4 or 5 GHz would be fine )
- 50 ohm impedance for the entire circuit.

I don't know if use microstrip technology or something else... any advice? Coplanar could be also ok( maybe in this case I can avoid to insert vias... )?
And if I estimate the width and the length of the tracks with TXline, then I can make the pcb with a normal pcb software? (obviously setting the correct values of W and L)

(sorry for my bad english)

thank you very much, these informations are very useful for me...
 
OK, my suggestion would be to use a coplanar transmission line (CPW) since you then won't have to use vias.

And yes, once you have the dimensions of the line you can just use normal PCB software. That is what I always do with simple things like this.

One "trick" you can use when drawing CPW using normal PCB software is to set the width of the interconnects in the software equal to the widht of the centre line and then set the parameters for "copper pour" so that the distance between copper and interconnects is equal to the centre-line <->ground plane distance of your CPW. This makes it easy to draw bends etc.

BTW, FR4 isn't exactly a great material at 5GHz.

Also, what is the value of your resistor? I just noticed that you have R=1M in your simulation which doesn't make much sense, I am guessing it should be a 50 ohm resistor, right?
 
first of all: so many thanks! :)

I've don't understood exactly your trick, however tomorrow I'll study it better (now in Italy it's 9 pm ---> tired ..)
I'll make a test pcb and I'll show to you...

FR4 isn't great but, unfortunately, at the moment is the only material available...

The resistor is 1Mohm.
Why? I try to explain: I have a switch, on another pcb. It has 1 input (from an antenna) and 2 outs, the 2 outs go to this pcb I'm designing.
One sma goes to ground, to shortcircuit the antenna.
The other sma goes to an high resistance, to simulate the "open-circuit antenna condition".
I hope I made myself clear...

However, any advice to make a better design/schematic would be very useful.

The main goal is to switch the antenna from 2 positions: open and short-circuited.

thanks again
 
I've tried some variations on txline. here is the result:

http://img230.imageshack.us/img230/8035/immaginevf.th.jpg

What's your opinion?
(substrate fr4 wasn't in the list, so I've chosen another one, but with correct dielectric constant).

One more thing (maybe very stupid but very important for me):
The length (10mm in this case) is the length from connector to sma resistor or the length from connector to ground (resistor included)?

i.e.:
http://img42.imageshack.us/img42/5711/immagine2i.th.jpg

maybe it's a very stupid question because the track length from the resistor to ground should be zero, considering that this is a coplanar pcb...

:blushing:

thanks again
 
Last edited by a moderator:
That looks reasonable. Note that you can enter the material parameters yourself in txline

My guess is that the length isn't very relevant for you unless you are worried about the phase shift for some reason (the electrical length); but if you know exactly which frequency you will be working at it might be a good idea to think about where you want to place the resistor with respect to the nodes on the line.

Speaking of the resistor, I doubt there is such a thing as a 1MOhm resistor at 5 GHz, the parasitics in a real resistor will effectively shunt it meaning the impedance will be much lower than 1 MOhm, probably only a few hundred ohms.
Is there any reason why you can't just leave the line open?
 
f95toli said:
Speaking of the resistor, I doubt there is such a thing as a 1MOhm resistor at 5 GHz, the parasitics in a real resistor will effectively shunt it meaning the impedance will be much lower than 1 MOhm, probably only a few hundred ohms.
Is there any reason why you can't just leave the line open?


really a few hundred?!? :bugeye:
there isn't a reason why I use a resistor instead an open line.. however if you say that the resistor behavior is that little value, so I can left the line open.

In this case, how far must be the open line from the ground plane? the value G (gap from line to ground) right? However, maybe the open line is only the connector...

Last thing: if I don't use the resistor, W and L are useless... I connect directly the sma connector to the ground plane... Is this correct?

If I understand these things then I can make the pcb...

Thanks
 
vaka85 said:
really a few hundred?!? :bugeye:
there isn't a reason why I use a resistor instead an open line.. however if you say that the resistor behavior is that little value, so I can left the line open.

Don't take my word for it, I've never even tried using a resistor that large at a few GHz; but I would be surprised if you get anywhere near 1Mohm. You basically have to find the datasheet for the resistor you are planning to use. Also, how you solder it etc will have a large effect on the result. You can download some applications notes about using SMD resistors at MW frequencies from e.g Vishay, although they will mainly deal with lower values.

In this case, how far must be the open line from the ground plane? the value G (gap from line to ground) right? However, maybe the open line is only the connector...

You need a computer simulation to say for sure, but I'd say a gap of W should be OK.

Last thing: if I don't use the resistor, W and L are useless... I connect directly the sma connector to the ground plane... Is this correct?

If I understand these things then I can make the pcb...

Thanks

No, remember that at these frequencies the line itself will be a "component" (a common component in MW applications is the "stub", which is quite literally just a line) , both the open and the shorted line will for example give you resonances (just calculate the wavelength and you'll see why) of etc so it is definitely not the same thing as just a connector.
 
  • #10
Ok, I've understand.

Next days I'll try to make a pcb, then I'll show the result.

Thank you very very much!
 
  • #12
last step please :)
 
  • #13
you have an open and a shorted transmission line, is it time to test it?
 

Similar threads

Circuit Board Design: Basics of Designing a PCB
  • · Replies 11 ·
Replies
11
Views
4K
What is the Design Process for Building an RF Mixer Circuit?
  • · Replies 25 ·
Replies
25
Views
6K
Usefulness of an RF Circuit Design Class
  • · Replies 1 ·
Replies
1
Views
2K
How to Measure PCB Impedance and Component Coupling?
  • · Replies 12 ·
Replies
12
Views
2K
Designing Circuits: RF & Sensor Basics for Self-Study
  • · Replies 7 ·
Replies
7
Views
3K
Designing an AM Generator Circuit with BF494 & IFT
  • · Replies 14 ·
Replies
14
Views
7K
How Do You Design an RF PIN Diode Switch for High Isolation?
  • · Replies 3 ·
Replies
3
Views
11K
Design of interrupt handling hardware circuit for a real time scenario
  • · Replies 19 ·
Replies
19
Views
3K
How Do I Design a Regulated Power Supply Circuit in Proteus?
  • · Replies 1 ·
Replies
1
Views
3K
Op-amp compensation of cascade fudges audio signal oddly
  • · Replies 3 ·
Replies
3
Views
2K