[RFI] DSC Alarm Panel 75 Meters/False Alarms

[Jim Brown]

On Wed, 17 Nov 2004 07:33:04 -0500, Brad Roope wrote:

>installed a common split core around the leads at the input to the panel.  

Common split cores are optimized for 50 MHz and us, and would not be expected to 
be effective at 4 MHz. 

>In addition he grounded the new panel to the common telephone/electrical ground.
>
>If DSC does not respond to their RFI Issue, I may ask my neighbor if I can install 
>#77 beads at the input to the Alarm Panel. I may get only one opportunity so I need 
>to know if multiple beads on each lead would be better that just one?

Your tone sounds defensive -- it should not be. DSC is the guy with the problem, and 
THEY should be working much harder to resolve this.  My guess is that there is a "pin 1 
problem" in their equipment, excited by common mode current on the interconnect 
wiring (receiving antenna). Choking that common mode current could fix it. 

Some basic principles of beads. The series impedance is proportional to the length of 
the bead along the wire. So yes, more is better, but the series impedance is not likely 
to be large enough to make a dent, even with multiple beads.  

The series resistance is also proportional to the square of the number of turns, and it is 
far more likely that a multi-turn choke of #78, #31, or #43 material would solve your 
problem.  

Fair-Rite makes 2.4" toroids of #78 and #43 material.  I would start with 8 turns around 
either of those. Fair-Rite also makes some big clamp-ons of #31 that can be used to 
form multi-turn chokes around wiring that either can't easily be disconnected or has a 
connector on it too large to fit through a solid core or toroid. 

>I may also install by-pass capacitors on each lead. 

Bypass capacitors are probably inappropriate if there is data on the line (as opposed 
to a dc signal generated by the sensor). And it is far more likely to be common mode. 

Another point. Look at the interconnect cable. Is it twisted pair?  It should be. Is it tightly 
twisted? It should be. Twisted pairs have solved LOTS of RFI problems when RF was 
getting into the output terminals of power amplifiers via zip cord used as speaker wire. 
My first choice would be a decent CAT5 cable, and it should NOT be shielded. Most 
cable shields can CAUSE common mode current to be coupled onto the signal pair as 
a differential mode signal. 

I wouldn't rule out coupling through the power supply wiring, but I would attack the signal 
wiring first. 

All of this, of course, is the responsibility of the installation company and the equipment 
manufacturer. Don't let them make THEIR problem YOUR problem. Let them know that 
this is all well established science, and that you are willing to cooperate by firing up 
your transmitter for testing on a schedule that is reasonably convenient to you (but not 
when you must take time off work). But make it clear that you know that fixing it is 
THEIR responsibility, not yours. 

BTW -- you are welcome to print this email and show it to your neighbor. And you can 
tell them that I am vice-chair of the Standards Committee of the Audio Engineering 
Society's Working Group on EMC, and that I'm a consulting designing sound systems 
for both large and small installations. 


Jim Brown
Audio Systems Group, Inc.
Chicago
http://audiosystemsgroup.com