Print

Please login or register.
1 Hour 1 Day 1 Week 1 Month Forever
Login with username, password and session length

[Keith Broughton]

Came across this device.
Any idea as to how it actually works?
http://www.ebtechaudio.com/humxdes.html

[Geoff Doane]

Came across this device.
Any idea as to how it actually works?
http://www.ebtechaudio.com/humxdes.html

I think we've covered it here before, but essentially it is just places two parallel diodes, one in one direction, the other reverse, in the ground connection.  That allows 0.6V of ground differential to be present without any current flowing.  If you get a real fault, the diodes conduct, clamping the equipment ground to ±0.6V .  If you get a REALLY big fault, the diodes short and hopefully you pop a breaker somewhere.  If the diodes opened, that would be a problem, but I've seen a lot of failed diodes, and I think they ALL shorted.

Still the lack of UL or CSA certification is somewhat concerning.

GTD

[Mike Sokol]

I think we've covered it here before, but essentially it is just places two parallel diodes, one in one direction, the other reverse, in the ground connection.  That allows 0.6V of ground differential to be present without any current flowing.  If you get a real fault, the diodes conduct, clamping the equipment ground to ±0.6V .  If you get a REALLY big fault, the diodes short and hopefully you pop a breaker somewhere.  If the diodes opened, that would be a problem, but I've seen a lot of failed diodes, and I think they ALL shorted.

Still the lack of UL or CSA certification is somewhat concerning.

GTD

I've done a bunch of testing on these, and you're exactly right. It's two silicon diodes in parallel with reversed polarity. There's also a small capacitor across the diodes to act as a high-frequency shunt, I believe. These do work to stop ground loop induced hum, but only with small differential voltages under 0.5 volts or so. I know this because my Glo-Melt transformer on my ground-loop demo rig allows me to create any ground loop I want up to 3 volts. I've personally measured a 5 volt ground loop differential, but that was in an industrial building. And certainly if someone flips the ground and neutral connections in an outlet you can get similar ground voltage differences of up to 5 volts very easily. I actually designed and built something very similar maybe 35 years ago, but it used 5 volt Zener diodes that were rated to 40 amps or more. That way I could stop hum in a system with up to 5 volts ground loop difference. The Hum-X is only good for stopping hum from small ground loop differential voltages up to 1/2 volt maximum.

The key point seems to be what happens if you introduce a dead short in the system? Will the diodes fail in open or short mode? I've blown up a few diodes that went open, but just like you most of the time they just short. Still, the Hum-X has pretty small diodes that I would be worried about failing catastrophically if plugged into a very stiff AC power supply. On my demo unit I've shorted the hot-to-ground with a screwdriver a number of times (10 maybe) and the 20-amp circuit breaker tripped without blowing up the Hum-X. I was on the end of an extension cord which certainly limited the peak currents available in a short, and maybe that's why my unit worked perfectly. So I can't guarantee what would happen if a Hum-X was plugged directly into a receptacle that had a short run directly into a 200-amp service panel. Who knows?

I think there are far better ways to reduce ground loop hum that doesn't resort to putting semiconductors in the EGC safety ground path. Ebtech also makes an audio transformer isolation box which works very well. As does Whirlwind and Radial. But you can figure that they're going to cost maybe $50 a channel or so if quality components are used. So figure somewhere around $100 for a stereo audio isolation box. But that's the kind of thing I put on all my troublesome systems to stop hum. In many churches that's a far cheaper solution than rewiring the entire electrical system.

[John Roberts {JR}]

Yes we have talked about this before.

A diode will generally fail as a short circuit and provide a safety ground bond, as long as the diodes have enough thermal mass to survive the transient heating (i.e. not vaporize). When I researched this back in the early '80s I used small epoxy diode bridges that gave me two diode drops in each direction and enough thermal mass to live long enough to take out the mains branch breaker/fuse.

I had my agency guy approach UL about this while working at Peavey (late 80s/early 90s) and they were receptive to opening a file, but I was not receptive to paying thousands of dollars to get approval for a work around that isn't actually needed when proper signal and grounding practices are followed (see pin 1 problem).

I wrote to the ebtech engineers privately and shared my observations that the diodes need to substantial to provide an adequate safety ground bond. I don't recall ever getting a response.

AFAIK this is not UL approved as a safety ground bond, at least it wasn't last time I checked.  I wouldn't use it unless you also have a GFCI protecting the path, or have taken it apart and confirmed that the diodes inside are large enough to survive a mains fault (diode bridge or say at least 3A to 4A diodes).   

JR   

[Keith Broughton]

Thanks for the info.
I have no intention of using them but wanted to know how they operate in case I come across one in a guitar rig or some stage gear.

I did do a "humx " search but no results.

[Mike Sokol]

Thanks for the info.
I have no intention of using them but wanted to know how they operate in case I come across one in a guitar rig or some stage gear.

I did do a "humx " search but no results.

Remember you can use a standard clamp-ammeter to search for ground loop currents in a PA system. Because the loop is outside of the XLR cable, you don't need split out the twisted pair from the shield. You just clamp around the entire cable and look for current. My experiments show that typical ground loop voltage to current ratios are around 1 amp per volt for normal cable lengths, and many pieces of gear with the pin-1 problem will hum noticeably with currents above 100 mA. Since most inexpensive clamp-ammeters will read down to 10 mA, it's pretty simple to find where the hum is coming from without disconnecting anything.

[John Roberts {JR}]

Remember you can use a standard clamp-ammeter to search for ground loop currents in a PA system. Because the loop is outside of the XLR cable, you don't need split out the twisted pair from the shield. You just clamp around the entire cable and look for current. My experiments show that typical ground loop voltage to current ratios are around 1 amp per volt for normal cable lengths, and many pieces of gear with the pin-1 problem will hum noticeably with currents above 100 mA. Since most inexpensive clamp-ammeters will read down to 10 mA, it's pretty simple to find where the hum is coming from without disconnecting anything.

This is a little like phase and polarity, when you say a speaker is wired out of phase, everybody knows what you mean, but technically it's opposite polarity.

As I've shared before (I'm sorry) ground loops are one turn windings, like a transformer secondary windings that convert whatever magnetic flux they intersect, into current/voltage flowing in those loops.

Of course it's remotely possible there is some massive magnetic flux present, in every suspected "ground loop". I expect the problem is the common "ground voltage potential differences" and the current flow is fully characterized by ohms law (I=E/R).

One simple experiment to confirm that it is, or isn't a magnetic loop (my bet), is take a long extension cord, then plug it's two ends together, and into nothing else. Spread out the extension cord to form a large area loop, to capture as much magnetic flux as possible. Then measure the current flowing in that loop. 

I expect many will keep calling them ground loops because it rolls off the tongue, and that parallel path between two ground nodes does ultimately proscribe a loop, but IMO calling it a ground loop is not an accurate description of the underlying electrical mechanism.       

JR

PS: Perhaps I'm just being pedantic about terminology, but "ground loops" are a real design issue inside products where the magnetic flux leaking from a power transformer can corrupt audio signal paths if wiring loops are present (or PCB layout loops, or whatever).

[Mike Sokol]

This is a little like phase and polarity, when you say a speaker is wired out of phase, everybody knows what you mean, but technically it's opposite polarity.

Yes you are correct, but the entire industry simply calls this "ground loop hum", and even my early ground impedance tester from the UK is called a GLIT (Ground Loop Impedance Tester). I just want to separate out the idea of the voltage difference between two "grounds" is what causes "ground loop" current to flow if there's a copper connection to conduct the current. And that current is what sneaks around inside of the gear with a pin-1 problem to create hum. I've never suggested that it's a magnetic coupling thing, nor does it come via the twisted pair. It seems to be caused by the shield being terminated on both ends, and the receiving gear having a return path that routes this current through the circuit board, rather than a direct bond to the AC power's EGC connector.

[Stephen Swaffer]

Mike,

Do you know the diode numbers? Diodes are cheap enough, it might make an interesting experiment after a frustrating week to see if the diodes would actually handle a serious ground fault?

[Len Zenith Jr]

Mike,

Do you know the diode numbers? Diodes are cheap enough, it might make an interesting experiment after a frustrating week to see if the diodes would actually handle a serious ground fault?

Here is a little excerpt from a guy who reversed engineered a humX:

"..but inside that slick little adaptor
is a very primitive (but clever) little circuit.  In a
standard AC receptacle, you have three contacts -
hot, neutral, and ground.  The hot and neutral lines
pass straight through the adaptor, although there's
a tap from the hot line (with a resistor) to power the
LED.  The mojo is in the ground line, which has
a pair of 6A/1kV diodes between the wall ground
connection and the adaptor's output socket ground.
These diodes are connected in parallel, with
the anode on one diode connected to the cathode
on the other.  (And vice versa, obviously.)  The
pair of diodes has a 1K 1/2 W -resistor in parallel
with it; as far as I can see, this resistor serves
to produce a "correct" indication when an outlet
tester is plugged into the adaptor, since without
it, the outlet tester will give an "open ground"
indication."

[ProSoundWeb Community]