[Amps] PCB Trace width for 1500 watts RF?

[Steven J Fraasch]

Dan

I have my Agilent ADS RF software at work, I will run some line widths
for you there next week.

I am not sure if you plan on microstrip (line over ground plane) , or
coplanar, where the line and ground plane are on the same layer.

The power and SWR determines the material selection in terms of max
voltage and current.  For high power HF, I would consider the thickest
board possible, 1/8" or greater.  A good exercise is to determine the
worst case VSWR and high impedance location, and determine the peak
voltage excursion there.  Likewise, determine the worst case ISWR current
maxima to determine safe current density.  For your case of 2.5:1 VSWR,
that is same as a inductive reactive quadrant load of 50 + 50j load, or
an impedance magnitude of 71 ohms.  The resulting peak voltage assuming
5.5 amps (real) into the load is 550V.   For the ISWR case, I determine
12.2 peak amps into a real 20 ohm load.  Therefore, I'd shoot for at
least 750V breakdown, and a line width that will support 15 amps, if not
more.  Remember, disconnected antennas or unterminated transmission lines
can present load cases far more severe than a slightly non-resonant load.
 I don't know if your amp transforms impedance along the way, or if it is
non-transforming, such as an antenna switch.  But, you will need to
re-calc max voltages and currents depending on the use of transforming
networks in the design.

Thicker boards results in wider traces for the same impedance.  For high
RF power, it is more important to keep the current density down in the
line, and voltages within ratings than to worry about a specific
impedance line traces.  Normally, you can accomodate characteristic
impedance lines between 25 and 70 ohms; less or more and the lines become
impractically fat or thin.  Your lumped components (Cs and Ls) in your
design can handle any rotation due to a non-characteristic impedance
line.  The problem is at microwave freqs where a guide wave can be on the
order of inches, and a good simulation program is required to model
lines, stubs, gaps, etc.  But, at HF, you should just need to pay
attention to the max current and voltage.

I'll send some charateristic line impedance widths vs. substrate
thickness next week, but ARRL handbook has this info, as does the
enclosed website.  Another suggestion is to use the LTspice from
www.linear.com or buy the League's RF cad software.  Either would be
high-value added in RF PCB design.

http://www.ece.uvic.ca/499/2002b/group11/microstrip.html

73, Steve Fraasch, K0SF

On Fri, 28 Jan 2005 14:05:24 -0800 "Dan Levin" <djl at andlev.com> writes:
> Anyone care to opine on the minimum pcb trace width appropriate (i.e. 
> safe)
> to carry 1500 watts between 1.8 and 30 Mhz into an impedance 
> mismatch of
> less than 2.5:1?  Assume 2 oz copper thickness.
> 
> Almost as useful would be if someone who knows something about RF 
> (since I
> know practically nothing, which is a separate subject) can tell me 
> the
> maximum current in a wire carrying 1500 watts into a 2.5:1 
> mismatch.
> 
> Thanks!
> 
>                         ***dan
> 
> _______________________________________________
> Amps mailing list
> Amps at contesting.com
> http://lists.contesting.com/mailman/listinfo/amps
> 
>