How much RF Current can a PCB Trace Handle? - RF Cafe Forums

 Post subject: How much RF Current can a PCB Trace Handle?

Posted: Wed Nov 30, 2011 11:57 am 

Good day.

I am working on a dual band power amplifier design (PA die, matching network, lowpass filter, and antenna switch) that will be placed inside of a package.

My question => What is the maximum RF current the traces on the pcb will be able to handle?

PCB Specifications
Thickness = 100 um
Dielectric Constant = 4.4
Loss Tangent = 0.006
Trace Thickness = 25 um
Minimum Trace Width = 75 um
Minimum Trace Spacing = 75 um
Overall PCB size = 5mm by 5 mm

Power Amplifier Die Specifications
Frequency = 900 MHz and 1800 MHz
Output Power = 35 dBm and 33 dBm respectfully
Modulation = GSM and CDMA
Zout = 1.6 + j1 ohms and 2 + j1.5 ohms respectfully.

Matching Network
Insertion Loss (passband, according to ADS) = 1 dB

Lowpass Filters
Insertion Loss (passband, according to ADS) = 0.3 dB
Insertion Loss (stopband, according to ADS) = greater then -30 dB

Antenna Switch
Insertion Loss (passband) = 0.0 dB

While the system will be using GSM and CDMA, I am approaching my analysis from a CW point of view. This will give me a worst case situation.

If you have any information (magazine articles, books, seminars, etc) about how much RF current a trace can handle, I would appreciate knowing about them.

My company has the IPC-2152 Standard and I feel that it is of no value in this situation.

Thank you

 Post subject: Re: How much RF Current can a PCB Trace Handle?

Posted: Tue Dec 06, 2011 4:27 pm 

At the end of the day it all comes to the basic formula:

P=I²*R.

So if your impedance is 50 ohm (which you need to deliver the maximal power to the load), then you can find the current.

For CW (which is indeed the worst-case), the calculations is the same as for DC.

There are tables which define the width of the trace as a function of the current. Pay attention that at the same time, to keep Zo, you have to take substrate characteristics into consideration (Dielectric constant, TanDelta etc, which you noted)

_________________
Best regards,

- IR

 Post subject: Re: How much RF Current can a PCB Trace Handle?

Posted: Mon Dec 12, 2011 10:17 am 

Good day.

Thank you IR for your response.

The problem I am encountering is that most (if not all) of the tables/charts are of no value for this design.

For example the IPC specification (IPC-2152) only gives information about a single trace on FR4 or polyimide. Also the size of the board is large compared to what I am looking at. There is no information about how to adjust the charts for different board materials, board thickness, trace width, or trace thickness.

Michael.

P.S. There has never been a problem with determining how much current is flowing in a trace.

 Post subject: Re: How much RF Current can a PCB Trace Handle?

Posted: Mon Dec 12, 2011 11:32 am 

Good day IR.

I disagree with your comment about CW calculations being the same as DC calculations.

The skin effect will reduce the cross-sectional area that the current will flow through. At DC the entire conductor thickness is used, where at 900 MHz the skin depth is around 2 to 3 um. So the effective cross-sectional area is now 225 um^2 instead of 1875 um^2

I forgot to include this comment in my earlier reply.

Michael

 Post subject: Re: How much RF Current can a PCB Trace Handle?

Posted: Mon Dec 12, 2011 5:19 pm 

Michael,

The distribution of current in the skin depth is 63% of the total current flowing in the conductor. The remaining 37% of the current flows between the skin depth and the center of the conductor.

Knowing the skin depth is of no relevance for knowing the current (It can of course be helpful to calculate it for other reasons).

Hope this helps.

_________________
Best regards,

- IR

 Post subject: Re: How much RF Current can a PCB Trace Handle?

Posted: Tue Dec 13, 2011 12:22 pm 

Good day IR.

Again I want to thank you for your response.

Let me re-ask my question. I feel that we are talking about two different issues.

I have a trace 75um wide and 25 um thick. The trace is 250 um in length. There will be other traces (same width and thickness) next to this trace. The spacing between the traces will be 75 um (minimum).

How much RF current (at 1 GHz) can flow through this trace such that there is only a 10 degree rise in temperature (above the ambient temperature)?

To determine how much current is flowing in the trace is not my problem (I have been using the method you described in your first response).

Michael

 Post subject: Re: How much RF Current can a PCB Trace Handle?

Posted: Tue Dec 13, 2011 2:14 pm 

This is what I just found after a quick search in Google (and there are many others available):

http://www.piclist.com/techref/pcb_traces.htm

Good luck!

_________________
Best regards,

- IR

 Post subject: Re: How much RF Current can a PCB Trace Handle?

Posted: Mon Dec 19, 2011 10:57 am 

Good day.

IR - Thank you for your post. However, the website that you specifically mentioned (as well as just about every other one I have looked at) are based on the old and obsolete specifications (IPC-D-275 and IPC-2221). IPC-D-275 was replaced by IPC-2221 and now IPC-2221 has been replaced by IPC-2152.

We do have a copy of IPC-2152 and it does not work for the situation that I am working on.

The basic question - what is the maximum current (at 900 MHz) that can flow in the traces I have described.

Thank you
Michael

 Post subject: Re: How much RF Current can a PCB Trace Handle?

Posted: Thu Dec 22, 2011 11:54 am 

Gents,

This document provides a good summary of the current (excuse the pun) situation:-

http://www.ami.ac.uk/courses/ami4817_dt ... ah0221.pdf

I have use the old IPC D-275 curves in the past, I lashed up a curve fitted mathcad routine to do the calcs for me. Of course, it now looks like this method isn't very accurate!

The comment about simply using P = I^2 x R...... be careful. In practice, most real world RF circuits may often end up mismatched for one reason or another - classic case being antenna de-tunes/faults. In this scenario the assumption of 50 ohms for your power calcs may get you into hot water (actually hot solder!). I have first-hand experience of a mismatched power FET de-soldering itself! of course, that's not down tot he tracks themselves overheating but a mismatched track may end up carrying a much higher current than you calculated by assuming nice 50 ohms everywhere.

When all is said and done you could spend hours faffing around with simulators trying to model the whole shooting match and if your really careful you might get a reasonable result. However, I would opt for a simpler approach and just get a single track (or a few tracks) cut on a test board, run your amp into the track at maximum 'smoke' with the traces matched, then do the same with mismatch. I would get a bigger amp and see how far you can go. You can use a laser thermometer to measure the temperature rise.

Enjoy

 Post subject: Re: How much RF Current can a PCB Trace Handle?

Posted: Tue Mar 20, 2012 11:55 pm 

Here is another article from National Semiconductor that may help.
http://www.ti.com/lit/wp/snaa113/snaa113.pdf
It also covers the important grounding of mixed signal circuits.

_________________
Sunshine Design Engineering Services