What's new
What's new

HF start on TIG popping GFCI outlets

rke[pler

Diamond
Joined
Feb 19, 2002
Location
Peralta, NM USA
I have an older Miller 180SD that's never caused a problem, but since moving my shop a couple of years ago it seems that the HF TIG start is popping the GFCI outlet breakers in the rest of the shop. The welder has a dedicated 50A line from the 100A subpanel and the GFCI outlets are on completely separate runs so I'm thinking that it's some RF component causing the trips. I don't recall the trips when running the welder in stick mode. The house GFCI isn't tripping and I never say this at the old house so I'm thinking the new GFCI outlets are somehow more sensitive to noise.

So I'm thinking of making a reactor for the supply line, maybe get rid of some of the noise the welder can throw on the line. Not sure if I need to hit just one of the hot legs or both - thoughts? I'll likely use 3-4 wraps of 6ga around an old transformer core, should give a few hundred micro Henry. Alternatively I could buy a 50A line reactor off eBay for ~$100.

Comments?
 
I have an older Miller 180SD that's never caused a problem, but since moving my shop a couple of years ago it seems that the HF TIG start is popping the GFCI outlet breakers in the rest of the shop. The welder has a dedicated 50A line from the 100A subpanel and the GFCI outlets are on completely separate runs so I'm thinking that it's some RF component causing the trips. I don't recall the trips when running the welder in stick mode. The house GFCI isn't tripping and I never say this at the old house so I'm thinking the new GFCI outlets are somehow more sensitive to noise.

So I'm thinking of making a reactor for the supply line, maybe get rid of some of the noise the welder can throw on the line. Not sure if I need to hit just one of the hot legs or both - thoughts? I'll likely use 3-4 wraps of 6ga around an old transformer core, should give a few hundred micro Henry. Alternatively I could buy a 50A line reactor off eBay for ~$100.

Comments?

At least modest RFI/EMI upline filtering is hard to class as anything but a "very good idea". "Drive isolation" transformers are good as well.

That said, I'd suggest you also have a closer look at your GFCI system and individual components.

I had this problem in spades wherein the PO had daisy-chained the kitchen and two baths outlets all off a lone GFCI in an upstairs bath. He'd also done all the garage outlets and electric door opener off a single GFCI master daisy-chain in the garage/annex. Done that way around 30 years ago, so that may have been an economic necessity in that era.

Removing the daisty-chains - a b****y "antenna effect" those can be - and replacing with individual and reasonably affordable (as ten-pack, anyway) Cooper Shock Sentry GFCI solved the problem.

NOW the only master-slave situations are in my quad outlets intended for multi-media or guests bringing 'puters and stuff with low-power but rather numerous "wall-warts".

Hardly EVER get a nuisance-trip now, only the uber-rare legitimately justified ones.

And.. that includes the garage/annex now shop where some crude and rude stick-welding plus SERIOUSLY rude DC Drive and destructive test or excursions into commutator arc-flash territory has been going on.

No "false" GFCI tripping, even on 120 V outets less than ten feet from the sound and light show.

I like that. Insist on it, even.

Some of the very closest outlets to the "power" test-bench happen to power my ISP's "commercial service" fixed-IP NTE, others power some of my OWN routers!
 
If on different circuits, they are clearly not reacting to "real" leakage current in things plugged into them. Possibly the stuff plugged into them has filters to ground and is passing some HF current to ground, or else, as is pretty likely the GFCIs are just reacting to the HF directly.

Either way, put a real filter , not just an inductor, on the welder line, and the problem SHOULD go away.

There is also a chance that the welder has something going on with it now, that was not before.

Or that the way the place is wired is causing the issue.... It seems to be associated with the new shop, and that might be telling you something. The green wire grounds might be joined up in odd ways, or even not properly bonded at the service panel. it is worth taking a look, just to be sure.
 
Check your breakers type, if the shop wiring is newer, they may be arc fault type and not GFI type.
These were required in many jurisdictions in the 2014 electrical code cycle, mandatory in the 2017 cycle nationwide.

These are notorious for nuisance tripping, even from noise from switching power supplies for LED lamps. Some brands are more sensitive than others. If this is what you have it's not likely that filtering will work for you. Some are know to trip from a hand held two way radio.

I would start at inspecting the breaker types before trying filtering.

SAF Ω
 
Either way, put a real filter , not just an inductor, on the welder line, and the problem SHOULD go away.
HF start? Out at the tip?

One might need to be welding inside a decent Faraday Cage to tame THAT little Devil if the GFCI goods are "listening" over their attached power conductors.

I very nearly just ripped them all OUT of the shop area until the Cooper's tamed the problem.

Panel is old skewl Square-D QO, FWIW. The Arc CB's and GFCI CB's went back into their as-purchased boxes "a while" ago.

Praise be to old panels that JFW and "Grandfathering" to let them continue to do.
 
This is all new construction 3 years old. I'm not seeing breakers tripping, just the GFCI outlets - the shop circuits are wired with a GFCI at the first outlet and the rest of the outlets in the series behind it. I think there's an arc fault breaking in the box and it's not tripping, certainly the 20 or so in the house box aren't (this is a subpanel off the main 200A panel 100' away in the house).

The only think I can figure is that the 50A line is inductively coupled to the runs to the convenience outlets, but that's not the case for at least one of the runs. The 50A run is in the walls and likely parallel to one of the convenience runs for maybe 20', but the rest of it is in conduit. The welder itself is a long ways from the nearest outlet that's tripping and only 6 feet from another that isn't tripping.

I'm thinking that just replacing the GFCI outlets with regular outlets is going to be the best choice. Cheaper than the alternatives, anyway.
 
This is all new construction 3 years old. I'm not seeing breakers tripping, just the GFCI outlets - the shop circuits are wired with a GFCI at the first outlet and the rest of the outlets in the series behind it.
There lies your "antenna" or inductive loop, then. Once "indivdualized', nothing but the spring-ocnacts, sensor-side of an empty outlet.

Previous "chain" every donw-chain outlet and the wire TO it were "listening" side of the sensor.
The larger the antenna, the greater the senstivity to small signals.

I'm thinking that just replacing the GFCI outlets with regular outlets is going to be the best choice. Cheaper than the alternatives, anyway.

For not much premum over the labor, I found a few boxes of the Coopers gave me both the nuisance-trip immunity AND better GFCI protection than I started with.

Your mere 3-year-old premises, it isn't the quality of the GFCI so much as that long chain picking up on even weak signals that confuse the GFCI.
 
HF start? Out at the tip?

One might need to be welding inside a decent Faraday Cage to tame THAT little Devil if the GFCI goods are "listening" over their attached power conductors.

I very nearly just ripped them all OUT of the shop area until the Cooper's tamed the problem.

Panel is old skewl Square-D QO, FWIW. The Arc CB's and GFCI CB's went back into their as-purchased boxes "a while" ago.

Praise be to old panels that JFW and "Grandfathering" to let them continue to do.


If HF start trips them, but WELDING does not trip them, with kilowatts instead of milliwats of arc, then antenna action is not likely the issue. That IS a decent test to make.... Stick weld and see if that also trips the GFCIs. If not, then start looking at the welder, and filters.


I
 
If HF start trips them, but WELDING does not trip them, with kilowatts instead of milliwats of arc, then antenna action is not likely the issue. That IS a decent test to make.... Stick weld and see if that also trips the GFCIs. If not, then start looking at the welder, and filters.

Given the chaining, swapping but ONE outlet in each of only a few places gets him going.

I actually appreciate having GFCI, so basically bit the bullet on 40+ year-old outlets being "due" anyway, and did the whole house as well as shop with "stand alones", NO chaining other than those same-box quads.

I do also have the Faraday Cages, BTW. Three, but modest sizes. ISP's NTE, my router, the 10 GigE patch panel and switch.
 
If HF start trips them, but WELDING does not trip them, with kilowatts instead of milliwats of arc, then antenna action is not likely the issue. That IS a decent test to make.... Stick weld and see if that also trips the GFCIs. If not, then start looking at the welder, and filters.

Just tried that and there were no trips, so it sure looks like it's HF sensitivity. At the same time I have some 20A GFCI outlets right next to the welder that aren't tripping so I think it's just a problem with some cheap GFCI breakers. Here's a pic of the ones giving me problems:

bad.jpg

and the ones that don't trip:

good.jpg

I'll pull them out and get manufacturers, kind of afraid to see what the electrician I fired did back in there...

The ones that are popping are Cooper, haven't gotten an ID on the 20A ones.
 
If you are running from a sub panel, check to see if it has an isolated and insulated neutral buss. Equipment grounding conductors must be connected to a buss separate from the neutral.

Do a visual check on your grounds, all must tie back to the main service with no breaks in connections.
 
The grey 15A ones look counterfeit, lacking labeling of the test and reset buttons.

The white 20A ones look to be Hubbel brand.

Hopefully your breakers weren't purchased at the same discount source. It's a big problem in the industry.
Been going on for years. Cut rate contractors don't help.

UL warns of potentially hazardous GFCIs (Release 11PN-53) | UL

Mike Holt - Combat Counterfeiting

WARNING: “Smart Electrician” GFCI receptacles by Huizhou Faith | Electrical Business

And one from Eaton who owns the Cooper name..
Anti-Counterfeiting

SAF Ω
 
The white 20A ones look to be Hubbel brand.

The one with the red and black buttons look like ones that I stopped using, with the Leviton label.

Hopefully your breakers weren't purchased at the same discount source. It's a big problem in the industry.
Been going on for years. Cut rate contractors don't help.

Been going on for not years, but decades.
 
You could disconnect everything past the problem gfi and see if it is the antena effect or a bad gfi. You need to disconnect it anyway to change it so 1/2 way through turn on the welder and test. Then if the new gfi trips you know it is the long antena picking up signal. Then just put all outlets as gfi.
 
The way GFI's are SUPPOSED TO WORK, is excellent in theory, but in practice, they're unfortunately faced with some very obvious challenges of physics.

The first concept, is differential symmetry- what's going OUT on one side, should equate to what's coming BACK on the other... which makes total sense when you're dealing with a power system that say... 240v, with center-tapped ground and no neutral current.

The 120/240v circuit concept, however, defiles that a bit. ONE SIDE is 'hot', and the other, which is 'neutral' is also 'grounded'. That means that your LOAD is an ANTENNA, and it doesn't matter what you do, you'll have a circuit which is never free from differential-mode noise susceptibility. NOTHING, and I mean ABSOLUTELY NOTHING will change that... ever.


The second premise of GFCI operation, is that if there's ground current, there's something wrong, so ANY presence of ground current should result in a trip, right?

Enter basic fact of electromagnetic physics: Put a thousand-foot spool of wire on the ground underneath an overhead powerline, and ground one end of the spool, the other end will offer you enough induced voltage to kill you where you stand.

At HF frequencies, this is even more apparent.
These are just plain and simple facts of physics, and no matter how clever the GFCI's circuit may be, it will NEVER overcome those facts- engineers trying to beat those facts will merely sacrifice a little of this or that aspect of sensitivity, to overcome the challenges that exist in the real world... like HF noise, ground potential differences from atmospheric activity, etc.,

A GFI is exceptionally good at three things:

1) Protecting incredibly stupid people from doing things that are blatantly stupid
2) Grooming generally ignorant people into being even moreso ignorant of casual danger, to such point that they migrate to category #1
3) Exposing both 1 and 2, and an occasional member of the more intelligent technical realm to substantially higher dangers of complacency.
An excellent ecample being the placement of a GFI on a sump pump in a basement that's prone to flooding. A GFI, once tripped by the ground potential rise from a lightning strike, causes a basement to flood, which prompts an individual to venture in for a look. The GFI is submerged, and the GFI has tripped, faithfully protecting the sump pump from getting shocked, it also dismisses the possibility of said pump being able to do it's job...

But the BACK of the GFI is still live, so the idiot who put it in, goes to reset it, and gets electrocuted going there.

Being somewhat of the radio-fascinated persuasion, I've seen plenty of bad GFIs, and I've replaced 'em with better... but IMO, they ALL seem to fit the dumpster best.
 
The building code here requires that a circuit around a kitchen counter or bath counter has GFI's. The ones made today have leds, so at least you get a night light.
 
The way GFI's are SUPPOSED TO WORK, is excellent in theory, but in practice, they're unfortunately faced with some very obvious challenges of physics.

The first concept, is differential symmetry- what's going OUT on one side, should equate to what's coming BACK on the other... which makes total sense when you're dealing with a power system that say... 240v, with center-tapped ground and no neutral current.

The 120/240v circuit concept, however, defiles that a bit. ONE SIDE is 'hot', and the other, which is 'neutral' is also 'grounded'. That means that your LOAD is an ANTENNA, and it doesn't matter what you do, you'll have a circuit which is never free from differential-mode noise susceptibility. NOTHING, and I mean ABSOLUTELY NOTHING will change that... ever.


The second premise of GFCI operation, is that if there's ground current, there's something wrong, so ANY presence of ground current should result in a trip, right?

Enter basic fact of electromagnetic physics: Put a thousand-foot spool of wire on the ground underneath an overhead powerline, and ground one end of the spool, the other end will offer you enough induced voltage to kill you where you stand.

At HF frequencies, this is even more apparent.
These are just plain and simple facts of physics, and no matter how clever the GFCI's circuit may be, it will NEVER overcome those facts- engineers trying to beat those facts will merely sacrifice a little of this or that aspect of sensitivity, to overcome the challenges that exist in the real world... like HF noise, ground potential differences from atmospheric activity, etc.,

A GFI is exceptionally good at three things:

1) Protecting incredibly stupid people from doing things that are blatantly stupid
2) Grooming generally ignorant people into being even moreso ignorant of casual danger, to such point that they migrate to category #1
3) Exposing both 1 and 2, and an occasional member of the more intelligent technical realm to substantially higher dangers of complacency.
An excellent ecample being the placement of a GFI on a sump pump in a basement that's prone to flooding. A GFI, once tripped by the ground potential rise from a lightning strike, causes a basement to flood, which prompts an individual to venture in for a look. The GFI is submerged, and the GFI has tripped, faithfully protecting the sump pump from getting shocked, it also dismisses the possibility of said pump being able to do it's job...

But the BACK of the GFI is still live, so the idiot who put it in, goes to reset it, and gets electrocuted going there.

Being somewhat of the radio-fascinated persuasion, I've seen plenty of bad GFIs, and I've replaced 'em with better... but IMO, they ALL seem to fit the dumpster best.

Seriously good job of distilling it down into a single post, Dave. Worth printing out and hanging on the wall as a remninder that might avoid serious grief.

Warts and all - and the Cooper's I use have their share (made in China, but then what ISN'T?) - I still prefer 100% of the outlets that I CAN entice to work with GFCI have them. "Whole house", not just wet zones.

The safety gain may be slight.Very much less, as you point out, than those who came to mandate such goods might have wished for.

Even so, the nuisance has not been onerous, given close attention to the "rest of" site bonding and grounding. At the moment, a Harry Homeowner grade mid-sized Campbell-Hausfield air-compressor's shut-down event is the only routine offender, and even that but once in a great while. Easily corrected, just not worthy of the time before it is upgraded to a better one, all around.

Meanwhile? Band-aids.

On days I can't tolerate the silly device inserting interruptions into the air supply needed to support some project?

I just grab a high-grade extension cord and operate if off outlets in a different zone, 30 or more feet away.

That doesn't "fix" a damned thing. But it gets the job done.

:)
 
The way GFI's are SUPPOSED TO WORK, is excellent in theory, but in practice, they're unfortunately faced with some very obvious challenges of physics.

The first concept, is differential symmetry- what's going OUT on one side, should equate to what's coming BACK on the other... which makes total sense when you're dealing with a power system that say... 240v, with center-tapped ground and no neutral current.

The 120/240v circuit concept, however, defiles that a bit. ONE SIDE is 'hot', and the other, which is 'neutral' is also 'grounded'. That means that your LOAD is an ANTENNA, and it doesn't matter what you do, you'll have a circuit which is never free from differential-mode noise susceptibility. NOTHING, and I mean ABSOLUTELY NOTHING will change that... ever.


The second premise of GFCI operation, is that if there's ground current, there's something wrong, so ANY presence of ground current should result in a trip, right?

Enter basic fact of electromagnetic physics: Put a thousand-foot spool of wire on the ground underneath an overhead powerline, and ground one end of the spool, the other end will offer you enough induced voltage to kill you where you stand.

At HF frequencies, this is even more apparent.
These are just plain and simple facts of physics, and no matter how clever the GFCI's circuit may be, it will NEVER overcome those facts- engineers trying to beat those facts will merely sacrifice a little of this or that aspect of sensitivity, to overcome the challenges that exist in the real world... like HF noise, ground potential differences from atmospheric activity, etc.,

A GFI is exceptionally good at three things:

1) Protecting incredibly stupid people from doing things that are blatantly stupid
2) Grooming generally ignorant people into being even moreso ignorant of casual danger, to such point that they migrate to category #1
3) Exposing both 1 and 2, and an occasional member of the more intelligent technical realm to substantially higher dangers of complacency.
An excellent ecample being the placement of a GFI on a sump pump in a basement that's prone to flooding. A GFI, once tripped by the ground potential rise from a lightning strike, causes a basement to flood, which prompts an individual to venture in for a look. The GFI is submerged, and the GFI has tripped, faithfully protecting the sump pump from getting shocked, it also dismisses the possibility of said pump being able to do it's job...

But the BACK of the GFI is still live, so the idiot who put it in, goes to reset it, and gets electrocuted going there.

Being somewhat of the radio-fascinated persuasion, I've seen plenty of bad GFIs, and I've replaced 'em with better... but IMO, they ALL seem to fit the dumpster best.

Frankly I disagree with almost everything in the post.... have you ever even SEEN a GFCI?

To begin with, resetting is accomplished from the front, in case you have never seen one...no need to go to the insides, as your, ummm, "discussion" seems to imply. You do not touch any conductors, whether or not your feet are in water... or just on the usual damp concrete. So the chances of electrocution are slim even in that case... and if it really IS submerged, all bets are off anyway. The stupidest people still know not to get in water with live wires...... not to mention other technical issues with the scenario....

THAT is your argument against the GFCI? Really?

GFCIs serve a good purpose.... not perfectly, but decently. They protect against faults in equipment, for one thing, faults that you would otherwise not even know about until the next test of that piece of equipment.

As long as you have two wires, you have differential signals. Get used to it. GFCIs generally reject those signals, but perhaps not at 50 MHz, where spark energy often is.

Add ground, and now you have common mode signals, which are what the GFCI is intended to sense.

If your plug-in unit is blasticg RF all over the place, then you get what you deserve.... problems. The GFCI does not work at those frequencies, but may sense the results from them, so blasting RF may cause troubles. Surprise, I guess.....
 








 
Back
Top