[TowerTalk] Circuit Board Question

[Jim Lux]

On 7/31/14, 11:19 PM, Matt wrote:
> Could someone help me out with some advice.
>
>
>
> How much separation between circuit board traces in an antenna switch box
> would be recommended to reliably handle the voltage associated with legal
> limit at 2:1 SWR?

You can start by looking at the underlying physics:
http://home.earthlink.net/~jimlux/hv/paschen.htm
http://home.earthlink.net/~jimlux/hv/sphgap.htm


A good rule of thumb is that the breakdown voltage across an insulator's 
surface is 1/3 that of the breakdown in air. for uniform fields: 
70kV/inch in air. This is "sharp edge to sharp edge" and another rule of 
thumb says that the needle gap distance is 3 times the "almost uniform" 
gap of big spheres.  So overall, I'd say a factor of 10.
In air, uniform field, 0.014 inches/kV
across the board, without coatings, maybe 0.140 inches/kV?


Conformal coating (with no voids) greatly increases the breakdown 
voltage, because you're not sparking across a surface. Now, you're 
looking at the breakdown strength

I'm sure there's some "design rules for HV PWBs" out there..
http://www.smps.us/pcbtracespacing.html
gives a UL test of 40V/mil (this is about half the air gap breakdown for 
uniform fields), and discusses various industrial standards..

http://www.smpspowersupply.com/ipc2221pcbclearance.html is a handy table


1.5 kW into 50 ohms is 273Vrms, or 383V peak.
5.4 Arms, or 7.67Apk (for current handling, though, rms is more appropriate)
2:1 VSWR is double that, or around 800Vpk (for breakdown), and 11 Arms 
(for thermal limits)

So, by my "rule of thumb, physics based", I'd say about 0.112 inch spacing.

The above site gives spacings ranging from 0.1811" down to 0.0677" (for 
800V) depending on which rule system you use.
For 500V, the graph shows 2.5mm (very close to 0.100") for IPC2221A 
which is for uncoated PC boards.

http://www.smps.us/pcb-calculator.html  has a handy trace width calculator
it gives widths ranging from 50 to 85 mils for 11 Amps..
I'd be real careful, though.. at RF, skin effect is something worth 
considering. At 30 MHz it's about 0.5 mil.  1 oz copper is 1.4 mils 
thick.  3 oz would be three times that, or 4.2 mil.  For 1 oz copper, 
then, the RF current will fill the conductor, but for the thicker 3oz 
copper, at least at the top of the HF bands, the current will be flowing 
mostly in the surface of the conductor.  That said, the thicker copper 
will give a better thermal path.



>
> Just tinned or conformal coating?
>
> How wide would you recommend the trace to handle the same signal using 3oz
> board stock - longest trace is about 2", most less than 1"?



>
>
>
> Thanks in advance
>
> Matt
>
> KM5VI
>
> _______________________________________________
>
>
>
> _______________________________________________
> TowerTalk mailing list
> TowerTalk at contesting.com
> http://lists.contesting.com/mailman/listinfo/towertalk
>