Shield Drains and Ground Loops

Nathan Miller

Pre-takeoff checklist
Joined
Oct 5, 2019
Messages
377
Location
Washington, D.C.
Display Name

Display name:
Nathan
Things are progressing on my panel (re)stack. I've carefully studied Garmin's install manuals for all the equipment I'm putting in. One thing that surprises me is how they depict the grounding of the shield wires on cables that carry RS-232 - they connect one end to the back shell connector and the other end goes to the airframe ground. This flies in the face of every admonishment to avoid ground loops in their manuals. Is avoiding ground loops only important for audio signals (music, headset, and mic connections)? It seems data would be susceptible as well. I'm inclined to follow the manufacturer's advice, of course. Just wanted to ask ..wtf? :p

-Example of grounding one side of the shield wire:ShieldGround2.jpg


-Example of grounding both ends of the shield: (Left side, in this case, goes to the back shell and the right side goes to the airframe ground).ShieldGround.jpg
 
I’ve always understood it as a matter of keeping noise out (audio, grounded one side), vs keeping noise in (grounded a both sides). You’ll see all the 429/232/422 etc interconnect wiring grounded at both ends.
 
Do the next guy a favor, and install everything IAW the install manuals. I makes troubleshooting so much easier. Problems always arise when installer get creative. Make a wiring diagram of exactly what you did, even if it's a photocopy of the IM interconnect.

And label each bundle as to where it goes. In other words, if the bundle contains GPS cables to an indicator, label as such. If it's resolver inputs, tag it as resolver. l
Labels are a good thing to have.

And running a couple extra cables here and there for future use can save you a bunch of work in the future. Just make sure to properly coil, stow and...label.
 
Simple, audio lines are ALWAYS grounded at ONE end.

(I'm sure someone somewhere has an example that contradicts this but I've never seen it in PS engineering or Garmin manuals)
 
That's what I'm saying :D

Thanks for everyone's input. Definitely not looking to get creative. Just trying to understand.

And Chip - thanks. Excellent advice.
 
Its actually written in the specification. http://ftp1.digi.com/support/cabling/cabling_specifications.pdf

RS232 is "shielded" to provide a "tunnel" at ground potential the entire length of the data path. But, analog (voice/nav) shield is only grounded at one end to provide a "drain" so that stray voltage is carried to one place (ie. not looped back into the audio common). I don't understand the physics behind it. I must've been asleep in class the day they taught that.
 
If you were to look at the plug bodies on the unit where it shows shield ground, you will probably find that the shell doesn't actually contact any metal within the case. As to grounding only one end to the airframe, it's because without extensive bonding work between aircraft components, it is possible to have electrical potential between joined pieces. We have bonding requirements for various installations, with most being under 50 milliohms across the joint. Some go to less than 2.5 milliohms. Gets to be frustrating to complete an installation and not be able to meet the bonding requirements and have to pull it all apart and start again.
 
It's not a universal truth that you want to ground audio cables only on one side. In fact, most are not. The good thing about aircraft is that there is usually not the pervasive 60Hz coupling that ground base wiring is subject to. On the other hand, you may have a 400Hz whine if such power is used. The problem isn't so much the shield as what it's connected to.
 
Rule of thumb for where I worked:

Wire as the manufacturer specifies, and get the performance they specified, and the warranty.

Wire it your way, and both the above are not to be expected.

We did raise issues with manufacturers from time to time, and they modified the specifications, we rewired accordingly, and the warranty remained in effect. We had a pretty good lab, and harsh test conditions, but did not exceed the specs of the equipment.
 
It's not a universal truth that you want to ground audio cables only on one side. In fact, most are not. The good thing about aircraft is that there is usually not the pervasive 60Hz coupling that ground base wiring is subject to. On the other hand, you may have a 400Hz whine if such power is used. The problem isn't so much the shield as what it's connected to.
The audio noise problems I fixed were very often due to mike and phone jacks being grounded. Insulating them so that the system was grounded at one end fixed it. That alternator pumps electrons into the airframe, and they have to find their way to the battery negative and all the grounded electrical equipment. As it travels through the airframe it takes the easiest paths, and if an easier path is found through a cable shield, that's where some of it will go. It will induce current flow in the mike wiring that gets amplified. If it induces in the phone wire, it goes through the headset speakers and into the common ground with the mike and pollutes the signals that way.

Better alternator filtering, of both the field and output, can help, but the strobes can also feed noise into the systems. Old airframes are the worst, with tiny amounts of corrosion between skin and structural joints adding resistance that encourages the current to find other paths.

And it's not just avionics. Electrical oil pressure and temperature probes on the engine offer an alternate path to ground. On a poorly-grounded engine, alternator current will find its way into the probe, though the thermistor in it and along the wire to the gauge, where the extra current causes the gauge to over-read, and then it goes to the bus and battery, through the battery to ground, all because there was no decent path to ground in the first place. Cessna has a Service Bulletin outlining the installation of a small ground wire directly from the crankcase near the probe to the instrument case on the panel, to shunt that small flow and keep it out of the gauge.
 
Back
Top