Ground loops and audio hum

[k4kyv]

I have been re-routing some of my audio equipment in the shack, and have found the ground loop hum problem daunting to say the least.  My audio units are interconnected with balanced 500-ohm lines, and I have made it a point to do all grounding of equipment of each unit directly to the ground bus, and to avoid daily-chain grounds.  The balanced audio line is shielded, but the network of shields is connected to ground at only one point and left floating everywhere else.

Once I manage to minimise hum in the system, the slightest re-routing of audio lines to any one of the units (mic pre-amp, line amplifier, peak limiter, etc) causes hum to appear in the system again.  I use a patch panel for the main routing system, and sometimes patching around a unit simply by moving one patch cord over from one line to another causes unwanted hum to re-appear.

I think I have finally figured out the concept, but solving the issue is another problem - kind of like finding a cure for cancer.  With an array of interconnected equipment, you may achieve a 60~ a.c. null in the grounding system that corresponds to the most sensitive link in the audio chain, but changing anything in the lash-up causes the null to shift unpredictably.  Work for hours to  get the null back where the stray a.c. causes the least problem, and it takes disconnecting one unit, or changing one patch cord, and the null shifts again.

I happen to have on hand a half-dozen or so 400 V-A isolation transformers originally designed for use in TV service shops.  This morning I tried running everything in the shack off isolation transformers, allowing the 120 v.a.c power lines to all the equipment except the transmitters themselves to float, rather than be strapped to earth ground via the house neutral at the meter service entrance, and that seems to have substantially reduced the hum problem although it didn't completely go away.  I suspect the a.c. null is more controllable with the floating a.c. line than with the neutral connected to ground at the service entrance.

Of course, I did not disturb the third-wire safety ground, but according to the NEC neutral is supposed to be connected to earth ground at one point and one point only: at the service entrance.

The isolation transformers have apparently allowed for some improvement, but my whole station is still a perpetual work in progress as it has been for the past 49 years.

[k4kyv]

Exactly what do you mean by an isolated ground system?

I run a 3-wire underground cable from the house: two hots and a  neutral.  The neutral is isolated from the ground beyond the point where the two are tied together at the main service entrance to the house, and not connected to ground at the entrance to the shack.  I have an earth ground just outside which consists of a 8-ft ground rod, some  radials, and a buried #4 solid copper wire that runs from my ground rod to the main ground rod at the house service entrance.  I also have a 40m  counterpoise tacked under the shack to the floor joists to get rid of some rf problems I had on that band.  But that ground is completely solated from the neutral that comes into the shack.

I measure about 0.25 volts a.c. (IIRC) between the neutral and earth ground at the shack. My audio equipment runs off 110v, so that is pulled from neutral to the hot wire on one side of line, but I always treat the neutral the same as a hot wire, and depend on the safety (earth) ground for rf and lightning protection.  I noticed when I shorted the neutral to ground with a  clip lead to see just what would happen with the neutral grounded at the shack, and I got additional hum in the audio equipment.  I believe NEC says to strap neutral to a separate earth ground at the entrance to a separate building, but I just treat my shack as if it were another room in the same house.

I placed the isolation transformers between the 110v mains and my audio rack.  I used two, because one of the circuits is live only when the transmitter is turned on, while the other one is live all the time, to power the vfo and receiver which normally stay on all  the time except during T-storms.

In addition to isolating the 110 volt lines, the transformers should offer some surge protection in the event of a major lightning strike, since the transformer core will saturate at some point and limit the output voltage at the secondary.

So what exactly what are you calling an isolated ground system?  Would it be any different from what I just described?

[KB5MD]

I feel your pain Don.  I have the same problems, with AC hum and RFI.  It would be interesting to see how others have treated the problems and to what degree of success has been achieved in getting rid of them.  I'm going to try your idea of isolation transformers.  My Gates audio console is terrible for picking up RFI and you can forget about using openwire line in my shack! 
"Enquiring minds want to know!"

[Bill, KD0HG]

In broadcasting studios, it's common to use a "balanced, push-pull" AC power source to prevent ground loops.

Take an AC voltmeter or scope probe and look at what's between neutral and ground, which are supposedly tied together. It's scary what you'll see, especially in an office with many computers. Don, I see that you've already done that.

The idea is to use center-tapped isolation trannys, like Don is using, to feed the audio rack, but in a different way. The difference is that the secondary center-tap of the isolation transformer is grounded to steel building frame or directly to its *own* ground rod (NOT to any existing ground).. And that's where the 3rd ground prong of the audio rack outlets connects to as well. The 120 VAC secondary of the isolation transformer connects to what was once the "hot" and "neutral" prongs on the audio rack outlets.

This provides a truly balanced, instead of single-ended, power feed to sensitive audio components.

You also can add audio isolation transformers where needed. The old-style Bell System 600-600 ohm transformers are perfect, they've got a really good FM-grade audio response when properly matched.

[W1DAN]

Hi Don and all:

Here are my thoughts in case you have not tried any of this. But I know you have much experience as you have been a ham longer than I have been alive (I used to listen to you as a kid before I got my ticket) For safety reasons, I am scared of cutting off any ground lugs on power cord plugs. Insurance companies may have a problem with this.

If you do not have this already, I wonder if you may wanna try the a star system of grounding of your gear instead of tying things to a long pipe or bar. If your transmitter is at the end of a buss bar in a series situation, when you transmit and draw power, this can increase your (ground loop) voltage generated along the ground system due to any small amount of resistance in the buss bar (think voltage divider).

Is your gear solid state or tube? Active balanced in and out or transformer isolated? Voltage fed or terminated? (when it is active balanced and not floating, I sometimes like the use of the telco audio transformers as stated above if I have ground loop problems). Some audio gear's grounding is poorly implemented within the box of the audio device itself. In extreme cases the gear's internal grounding needs to be modified.

For RF immunity at my shack using balanced twisted pair audio cable, I normally tie the shield of the audio cable at both ends for RF immunity. In the TV station I work at we always tie only one end as you do, but we have little RF there. If I were in an RF environment and had hum, I'd try a 0.01 cap from the open end of the shielded cable's ground wire to the gear's ground. This will allow for an RF ground path, but not one for audio or 60hz hum.

Also if the hum is pure 60 cycle, it may be the magnetic field of a power transformer from one piece of gear in close proximity to audio transformers or electronics of another piece of gear. Re-locating the gear may be needed. And as usual, I keep everything at +4dbu or even +8dbu (or dbm).

Just my $0.02 worth.

Good luck!
Dan
W1DAN

[ka3zlr]

Great Subject Discussion,

 I don't know if any of the system moderators are reading but a couple of these postings would be very useful and should be added to the AM handbook on site. IMO.

This is a subject that's lightly discussed in alot of the Average handbooks.

[k4kyv]

In broadcasting studios, it's common to use a "balanced, push-pull" AC power source to prevent ground loops.

The idea is to use center-tapped isolation trannys, like Don is using, to feed the audio rack, but in a different way. The difference is that the secondary center-tap of the isolation transformer is grounded to steel building frame or directly to its *own* ground rod (NOT to any existing ground).. And that's where the 3rd ground prong of the audio rack outlets connects to as well. The 120 VAC secondary of the isolation transformer connects to what was once the "hot" and "neutral" prongs on the audio rack outlets.

This provides a truly balanced, instead of single-ended, power feed to sensitive audio components.

You also can add audio isolation transformers where needed. The old-style Bell System 600-600 ohm transformers are perfect, they've got a really good FM-grade audio response when properly matched.

Are you talking about isolation transformers with a balanced midtap on each 120v winding?  Those must be special units built just for that purpose.  The TV service isolation xfmrs I am using have taps only for adjusting line voltage from about 100v to 130v, so my 120v lines are now floating, with no ground connection at all. 

I do have a couple of larger ones that are designed to run on 120 or 240 volt lines.  The way they do it is by splitting the primary and secondary into two identical windings each.  Wire the split windings in series for 240, or in parallel for 120v.  I'll try using the series connections for 120v and let the point where the secondaries are tied together serve as midtap.  The transformer should run cooler @ twice the normal turns/volt, but the voltage regulation will be degraded due to quadrupling both the leakage reactance and ohmic resistance in each winding - something that shouldn't matter much in the audio rack where the load is pretty constant in any case.

Another possibility might be to use a small 1 to 3 amp variac across the output of the untapped isolation xfmr, grounding the variable tap, to make the variac serve as a midtapped autotransformer.  One could actually adjust for exact balance by varying the setting of the variac around the region of 60v output.

My radial ground system, which consists of 16000 ft of #12 bare copper forming 120 radials at 133 ft each, should serve as an excellent a.c. ground, connected to the shack via the braid on the RG-213 that feeds the remote antenna tuner.  It wouldn't do much good as a lightning or rf ground for the shack, although the tuner components right at the base of the tower have survived several direct hits to the tower during severe lightning storms, with no visible damage ever, while light switches and a.c. power switches to equipment in the shack and house were welded closed, and cordless phones and answering machines were wiped out, along with a couple of VCR's.

I have several UTC LS series audio transformers with three 600-ohm windings each, that might make good audio isolation transformers if needed.

[WA1GFZ]

The number one rule of isolation transformers is to reestablish a ground system on the secondary side. They work great for common mode line crud. All this sounds simple until you check the NEC bounced against your home owners insurance.

As a student of indirect effects lightning the rule is provide a current path or be prepared to handle the voltage.

As a direct effects novice I suggest a firm belief in God.
 

[n2bc]

Regarding Don's post yesterday,  this point: 

"I believe NEC says to strap neutral to a separate earth ground at the entrance to a separate building, ..." 

This didn't sound right.  I'm not an expert so I asked a friend who is. He's an inspector for the NYSFPA, NY State Fire Protection Assn.  The National Assn, NFPA, IS the NEC.

At least in residential wiring the statement is incorrect.  Neutral and ground are only bonded at the service entrance and both must be 'passed' to any branch circuit whether that circuit is in the same building or a different building fed from the same entrance panel. 

For example. If you want to add a branch to your shack, in the basement or a nearby building, you would install a two pole breaker in the main panel, use a run of properly sized 3-wire with ground cable.  At the branch panel the ground would go to a bus bonded to the box, the neutral would go to a bus isolated from the box, and the 'hot' wires to the main lugs. 

One key to the above is 'residential wiring'.  There are many other configurations spelled out for computer installations and 'sensitive electronic equipment' but I will not pretend to understand them. However, they are not applicable to residential.

Regarding additional ground rods. It is also an NEC requirement, again for residential, that all the grounds be tied together & back to the entrance panel ground, and that the conductor be at least #4 (possibly larger depending on local codes and the size of the residence's service). The #4 or larger requirement precludes using an existing run of 12GA romex... Essentially you have to have a dedicated & properly sized conductor running to the entrance panel ground. 

For a branch run to an out-building, there must be properly installed ground rods (two 8' rods 6' apart) connected with a continuous loop of #4 or larger at the branch circuit breaker box.  These 'return' to the service panel on the ground conductor of the 3-wire with ground cable.

None of this is aimed at solving the initial hum problem.  But any solution - again, at least in our homes, should start from and stay within code requirements.

73, Bill  N2BC

[Bill, KD0HG]

Are you talking about isolation transformers with a balanced midtap on each 120v winding?  Those must be special units built just for that purpose

Yes, I suppose they're an odd-ball design. The ones used in audio facilities also have a copper electrostatic shield between primary and secondary.

Again, the idea is to provide the audio rack with a truly balanced push-pull AC feed with neither side connected to ground. I suppose the secondary center tap is optional, but good for a safety ground. But you don't want to carry the ground and neutral that exists on the primary over to the secondary side.

Our Denver radio studios have isolated ground outlets, like hospitals use. In them, the third grounding prong isn't carried back on the green or the conduit back to the load center as in a residence. Instead, it connects to the building steel frame, which is a very low Z and separate ground. Again, that isolates the noisy AC system ground and loops from the equipment cabinets. Such outlets are usually color coded orange so you know they have an isolated ground.

Not a violation of the NEC at all. The neutrals of the outlets are still carried back to the load center and its required single-point tie between neutral and ground. There is no code violation with isolating electronic equipment with a transformer at an outlet and grounding its chassis separately. Electronics, cable TV, telephone, antenna and rooftop lightning rod grounds aren't considered part of the electric system. Ground 'em separately where you want to.

[k4kyv]

The cable that runs from my house to the shack is a 3-conductor drop cable designed for direct burial between the pole transformer and a residence.  A friend of mine who worked at the power company gave it to me.  Aluminium conductors, about 1/2" in diameter each, two for hot and one for neutral, all insulated.  I buried mine about 2' deep in plastic conduit, not wanting to trust the integrity of the system to the thin plastic insulation over the conductors.  I use a separate ground lead back to the service entrance made of solid #4 or #6 copper, buried a couple of inches below the surface of the ground like a ground radial.  The antenna radial ground system is bonded via the shield of coax  running to the remote antenna  tuner, plus several 20's radials at the grounding point at the shack and at the house.  The original 1930's-era wiring in the house is two wire with no ground, but later additions are standard 2-conductor with ground.

At the shack, I notice a substantially measurable a.c. voltage difference between the neutral and ground, and strapping the two together increases the hum level in audio equipment.  I'm pretty sure the NEC allows, but maybe doesn't require, the neutral to be strapped to ground at the entrance to a separate building, but in any case I treat mine exactly the same way as if it were an additional room in the house.

I haven't tried cad-welding, but I find that silver alloy brazing rods used for bonding copper plumbing together work well.  My radial ground system is bonded together with the stuff, and after a quarter century all connections appear intact whenever I dig down to inspect.  I once years ago used ordinary lead/tin solder to bond radials on an inverted-L to the common point, and the solder would deteriorate after about a month.  I literally had to routinely go out every few weeks and resolder all the  radials back to the ground lead.  The solder would turn to a white powder and the copper wire radials would fall away from the heavy ground wire. Lead solder is supposed to be illegal for bonding copper plumbing together because the lead may leach into the water lines, but I see it used all the time by supposedly licensed plumbers.  The silver allow requires a brazing torch; a propane torch doesn't get hot enough.  I use "Mapp gas" with a torch that works just like propane.  The alloy doesn't require any flux; copper soaks up the molten alloy like a sponge soaks up water.  The copper just has to  be reasonably clean with no scaly crud on its surface.

All equipment in the shack is grounded at the chassis and the local earth ground at the building.

All my 500/600-ohm audio lines float free from ground independently of the shield, but I have seen circuit diagrams showing midtaps on the 500/600 windings of the input and output transformers of the audio equipment, which are grounded at both ends of the line.  At the broadcast station where I worked ages ago, all the audio lines in the studio floated, with no grounded midtaps at any of the transformers.

[WA1GFZ]

When I wired my meter socket I asked my inspector if I should connect the ground to the meter socket. He told me to do the single point ground at the breaker panel. Inside my GE panel there is a strap bolted between the two ground buss bars. It can be  removed if it was mounted as a remote panel.
Don, The torch Home Depot sells with the push button start burning map gas will burn a hole through aluminum. I found out while soldering a pipe near a stud.
It has no problems dealing with the new silver solder. I did the whole house without a single leak and less that a tank of gas. Well worth the $40 I paid for it. 

[k4kyv]

You also have to be careful with copper.  When I was brazing my radial system together I had never used silver brazing rods or a Mapp gas torch before, and a time or two I ended up melting the end of the #12 copper wire into a round blob.  I quickly figured out that you want it to reach a faint, dull red glow, but never bright orange.  I have never tried soldering or brazing to aluminium.

The silver brazing makes a rock-hard connection that quickly turns black.  The ones exposed to the weather at the grounding switch, as well as the ones covered with soil where radials are bonded to the grounding strap, look exactly the same now as they did a quarter century ago.  I don't think you can beat that.  I have even managed to braze a ground wire directly to a copper-clad steel ground rod without too much difficulty.

A strange thing about copper is that after brazing, when it cools down, the heated copper becomes very soft and pliable, almost like chewing gum.  But bend the wire back and forth a couple of times and it instantly hardens back up to its normal rigidity.

[WA1GFZ]

Good Tip Don, I want to add some more ground wires in the yard and it would be cool to terminate them to the service entrance rod rather than coming inside to pick up the main buss. I just have the cheezie copper clamps and notice they look  almost green in the salt air.