Motor contactor drop out

[paulengr]

Yes, but imagine a power dip for a few cycles... where the contactor does not drop out. Its possible to have a significant current inrush afterwards.

Let’s be realistic here. The transformer core will be saturated at startup so current is limited only by the short circuit resistance of the transformer. At that point essentially all the energy is going into the primary anyways so you can do whatever you want on the secondary because it doesn’t matter. The transformer is already dead shorted on the primary side. The breaker/fuses must be set higher than that.

Not buying the holding in argument. Contactors must be designed to meet CBEMA so total loss of 2 cycles is required. Of course nothing “wrong” technically with dropping out. It only takes a cycle or 2 to mechanically drop out and one cycle for the arc to quench for larger breakers and contactors. The old “5-12” cycle information is true for oil quenched breakers and maybe mercury contactors but hardly any are left.

And if they did all auto restart at the same time that’s clearly a really poorly designed system. Same with leaving breakers at factory settings. You cannot engineer out the stupid factor. I’m surprised this even works. Most manufacturers set everything as low as possible to force you to set it.

As to the “expense” of electronics the cutoff continues to drop but starting around 100-400 A with some breaker lines the trip units are actually cheaper as electronic ones. With thermal mags the trip unit gets bigger and bigger with increasing size. With electronic breakers, overloads, etc., the electronics is the same and only the CTs get bigger. The CTs themselves have secondaries in the milliamp range. The trip unit cost is essentially fixed no matter the size of the breaker. Even back 15 years ago a GE Spectra RMS was cheaper than a thermal mag starting around 250 A. At 800 A and above there is a big bump in costs but not because of the electronics but the fact that mechanically we have reached the limits of a single stage spring mechanism. That’s why insulated case breakers first became a “thing” at 800 A and above but I’ve worked on up to 2500 A molded case breakers. I see one once in a while as the output breaker on larger generators. Even AB now sells electronic overloads cheaper than the cost of three heaters, around $100. Not all breaker manufacturers are cheaper with electronic due to hold over old school pricing and margins but it keeps getting better all the time. It’s a dirt cheap microcontroller with a couple CTs and some cheap interfacing. The two most expensive parts these days are the output coil and the CTs.

 

[paulengr]

Yes, but imagine a power dip for a few cycles... where the contactor does not drop out. Its possible to have a significant current inrush afterwards.

Let’s be realistic here. The transformer core will be saturated at startup so current is limited only by the short circuit resistance of the transformer. At that point essentially all the energy is going into the primary anyways so you can do whatever you want on the secondary because it doesn’t matter. The transformer is already dead shorted on the primary side. The breaker/fuses must be set higher than that.

Not buying the holding in argument. Contactors must be designed to meet CBEMA so total loss of 2 cycles is required. Of course nothing “wrong” technically with dropping out. It only takes a cycle or 2 to mechanically drop out and one cycle for the arc to quench for larger breakers and contactors. The old “5-12” cycle information is true for oil quenched breakers and maybe mercury contactors but hardly any are left.

And if they did all auto restart at the same time that’s clearly a really poorly designed system. Same with leaving breakers at factory settings. You cannot engineer out the stupid factor. I’m surprised this even works. Most manufacturers set everything as low as possible to force you to set it.

As to the “expense” of electronics the cutoff continues to drop but starting around 100-400 A with some breaker lines the trip units are actually cheaper as electronic ones. With thermal mags the trip unit gets bigger and bigger with increasing size. With electronic breakers, overloads, etc., the electronics is the same and only the CTs get bigger. The CTs themselves have secondaries in the milliamp range. The trip unit cost is essentially fixed no matter the size of the breaker. Even back 15 years ago a GE Spectra RMS was cheaper than a thermal mag starting around 250 A. At 800 A and above there is a big bump in costs but not because of the electronics but the fact that mechanically we have reached the limits of a single stage spring mechanism. That’s why insulated case breakers first became a “thing” at 800 A and above but I’ve worked on up to 2500 A molded case breakers. I see one once in a while as the output breaker on larger generators. Even AB now sells electronic overloads cheaper than the cost of three heaters, around $100. Not all breaker manufacturers are cheaper with electronic due to hold over old school pricing and margins but it keeps getting better all the time. It’s a dirt cheap microcontroller with a couple CTs and some cheap interfacing. The two most expensive parts these days are the output coil and the CTs.

 

[jim dungar]

Even 120/208 and 138/240?

Wow, are you really going to design hospitals with large feeders >1000A at only 208V? What hospital equipment is designed for 240Y/138 systems?

A low peak current limiting fuse is not enough?

No they typically are not. Most arcing faults are blow the current limiting point of large fuses, particularly over 800A.
But regardless, there is no way around NEC 240.67

About the same safeguards as someone closing a breaker into a fault.

It sounds like you prefer administrative controls over engineering controls.
The worker can be very diligent in their electrical safe work practices, but how does that prevent the purchasing department, the store room, or an outside contractor from changing brands of something often viewed as a commodity?

 

[mbrooke]

Wow, are you really going to design hospitals with large feeders >1000A at only 208V?

Why not? Routinely done up to 5000amps. (see attached photos for some 2,500amp US services)

What hospital equipment is designed for 240Y/138 systems?

Just about everything in the Republic of the Philippines. Single phase equipment is connected line-line. Most fun electrical design you will ever do.

No they typically are not. Most arcing faults are blow the current limiting point of large fuses, particularly over 800A.
But regardless, there is no way around NEC 240.67

Noted... I'd say is more the magnetic forces rather than paschens at play. But hey, now I have one more excuse to use more PDC (power distribution centers) in a building

It sounds like you prefer administrative controls over engineering controls.
The worker can be very diligent in their electrical safe work practices, but how does that prevent the purchasing department, the store room, or an outside contractor from changing brands of something often viewed as a commodity?

Its up to the management to make sure the replacement is correct. Its no different than hiring a micky mouse contractor that takes out a failed QOB 22kaic bolt on and puts in a 10kaic QO in the boiler room's NQOD. Or putting in lower filtration filter in an OR AHU unit. I could go on...

 

[mbrooke]

3,000 amp 120/208

 

[mbrooke]

PDCs/PDUs- I love these. Sprinkled around the building, they are great to way to distribute power and save on LV runs. And yes, the X-Ray unit does not mind 208 volts. Neither the MRIs or Tomography.

 

[mbrooke]

Let’s be realistic here. The transformer core will be saturated at startup so current is limited only by the short circuit resistance of the transformer. At that point essentially all the energy is going into the primary anyways so you can do whatever you want on the secondary because it doesn’t matter. The transformer is already dead shorted on the primary side. The breaker/fuses must be set higher than that.

Explain, confused here.

Not buying the holding in argument. Contactors must be designed to meet CBEMA so total loss of 2 cycles is required. Of course nothing “wrong” technically with dropping out. It only takes a cycle or 2 to mechanically drop out and one cycle for the arc to quench for larger breakers and contactors. The old “5-12” cycle information is true for oil quenched breakers and maybe mercury contactors but hardly any are left.

Good to know. So in theory, a 345kv line fault on high speed clearing should still cause the contactor to drop out?

And if they did all auto restart at the same time that’s clearly a really poorly designed system. Same with leaving breakers at factory settings. You cannot engineer out the stupid factor. I’m surprised this even works. Most manufacturers set everything as low as possible to force you to set it.

I agree if all motors restarted- but sometimes you want some motors to re-start automatically.

As to the “expense” of electronics the cutoff continues to drop but starting around 100-400 A with some breaker lines the trip units are actually cheaper as electronic ones.

Any idea of which lines 225 amps and below are coming with electronics? Or coming standard?

With thermal mags the trip unit gets bigger and bigger with increasing size.

Agree

With electronic breakers, overloads, etc., the electronics is the same and only the CTs get bigger. The CTs themselves have secondaries in the milliamp range. The trip unit cost is essentially fixed no matter the size of the breaker. Even back 15 years ago a GE Spectra RMS was cheaper than a thermal mag starting around 250 A. At 800 A and above there is a big bump in costs but not because of the electronics but the fact that mechanically we have reached the limits of a single stage spring mechanism. That’s why insulated case breakers first became a “thing” at 800 A and above but I’ve worked on up to 2500 A molded case breakers. I see one once in a while as the output breaker on larger generators. Even AB now sells electronic overloads cheaper than the cost of three heaters, around $100. Not all breaker manufacturers are cheaper with electronic due to hold over old school pricing and margins but it keeps getting better all the time. It’s a dirt cheap microcontroller with a couple CTs and some cheap interfacing. The two most expensive parts these days are the output coil and the CTs.

What about the cost of a typical >600amp breaker vs a fused switch?

 

[mbrooke]

Not buying the holding in argument. Contactors must be designed to meet CBEMA so total loss of 2 cycles is required.

Thank you for the new term! Learned something new today.


http://www.powerqualityworld.com/2011/04/cbema-curve-power-quality-standard.html

https://www.sceg.com/docs/librariesprovider5/pdfs/powerqualitystandards.pdf

Got me thinking about vectors too.

 

[paulengr]

Explain, confused here.

Read up on transformer inrush. Just like motor inrush is independent of load so is transformer.

Good to know. So in theory, a 345kv line fault on high speed clearing should still cause the contactor to drop out?

If it is long enough. Transmission level and distribution for that matter tends to be network distribution not radial.

Any idea of which lines 225 amps and below are coming with electronics? Or coming standard?

That’s pretty rare right now because at that point thermal magnetic beats microprocessor cost wise. I’d almost think you would have to buy a 250 A frame then dial it down. I know Masterpacts and Siemens will go to 50% of frame size so that gets you to 125 A. GE Spectra RMS went to 100 A bit no idea with current breakers. Electronic trip is generally 250-400 A and above.

What about the cost of a typical >600amp breaker vs a fused switch?

A lot depends on design. Generally breakers are more but it depends more on whether it’s MCCB, ICCB, or ANSI, or air vs vacuum vs SF6, generally in that order of increasing prices.

 

[mbrooke]

Read up on transformer inrush. Just like motor inrush is independent of load so is transformer.

Not what you have residual magnetism. The paper linked earlier is very eye opening for me.

If it is long enough. Transmission level and distribution for that matter tends to be network distribution not radial.

Transmission is meshed, distribution is run normally open and electrically radial. Exception is Con Edison, Com Ed and dense urban areas.

That’s pretty rare right now because at that point thermal magnetic beats microprocessor cost wise. I’d almost think you would have to buy a 250 A frame then dial it down. I know Masterpacts and Siemens will go to 50% of frame size so that gets you to 125 A. GE Spectra RMS went to 100 A bit no idea with current breakers. Electronic trip is generally 250-400 A and above.

Thats what I've seen.

A lot depends on design. Generally breakers are more but it depends more on whether it’s MCCB, ICCB, or ANSI, or air vs vacuum vs SF6, generally in that order of increasing prices.

Any numbers of the top of your head for say a 2000amp 240 volt switchgear with breakers vs fuses with off the shelf specs? I know thats asking "how much does it cost to re-wire a home" but roughly- ie a home rewire is about $10,000-$15000.

 

[mbrooke]

There may be a way around it:

(B) Method to Reduce Clearing Time. A fuse shall have a
clearing time of 0.07 seconds or less at the available arcing
current, or one of the following shall be provided:

0.07 seconds is key. But they don't really go into defining what they mean by arcing current or how to find it.

 

[gar]

191028-1203

mbrooke:

An arcing current is any current level that supports an arc.

This is a function of gas, pressure, length, and having been initially ionized.

I don;t know the lowest arc current possible, but it is likely under 1 A.

 

[mbrooke]

191028-1203

mbrooke:

An arcing current is any current level that supports an arc.

This is a function of gas, pressure, length, and having been initially ionized.

I don;t know the lowest arc current possible, but it is likely under 1 A.

Correct, but according to Pashcen's law you need about 340 volts peak in none ionized atmospheric pressure to sustain an arc.


Further, utilities have known for some time 208 volts is not a concern in regards to arcing:

 

[mbrooke]

.....

 

[paulengr]

I agree that somewhere in the neighborhood of 200-300 VAC arcs are not sustainable depending on the conditions. IEEE tested different configurations and has confirmed down to 208 VAC and above 2 kA arcs are sustainable and above 1.2 cal/cm2. Conditions where you have a set of vertical busbars terminating into an insulating barrier like into the top of a breaker in an MDP are the worst. In open air for instance most utility equipment it’s not going to happen but there is a lot of hand waving going on in the NESC table.

OSHA investigated a case in 2009 where an electrician died from burn injuries from an arc flash on a 240/120 temporary construction panel in Georgia.

 

[mbrooke]

I agree that somewhere in the neighborhood of 200-300 VAC arcs are not sustainable depending on the conditions. IEEE tested different configurations and has confirmed down to 208 VAC and above 2 kA arcs are sustainable and above 1.2 cal/cm2. Conditions where you have a set of vertical busbars terminating into an insulating barrier like into the top of a breaker in an MDP are the worst. In open air for instance most utility equipment it’s not going to happen but there is a lot of hand waving going on in the NESC table.

OSHA investigated a case in 2009 where an electrician died from burn injuries from an arc flash on a 240/120 temporary construction panel in Georgia.

IMO, I think a lot of it boils down to how IEEE, NEC and NESC define arcing.

 

[paulengr]

Which leads me to ask- after 30 years of not opening, what have you seen fail more: Fused switches like bolted pressures or molded case circuit breakers? What about power circuit breakers?

OK, a little background. I work for the largest motor shop in the Carolinas in their field service group. I have almost every kind of customer you can imagine. We do breaker and switchgear maintenance. I'm the guy that gets called when something seems to be wrong and the local guys can't, or won't, handle it. So for your scenario I either get the service call AFTER someone tried to open the equipment and the breaker/disconnect burned up on them and now they're trying to figure out what to do, or once in a while beforehand when the maintenance guys know enough to suspect what will happen but are looking for someone that can magically make it all better, or at least take the risk of pulling the handle when they aren't willing to do it themselves.

In answering your question, MAYBE a 15 or 20 A MCCB will successfully open but that's about it. In most parts of the country there was a huge capital expansion and the entire electrical distribution system was essentially built for the most part in the 1960's and 1970's so it's all about 50 years old. Some newer parts exist to be sure but a lot of it is that old. So I see this kind of stuff pretty frequently, but even the 1980's and 1990's stuff is in pretty rough shape these days.

First we're not talking about tripping here, just opening. If we're talking tripping I feel pretty good about the fuses. It's an electro-chemical reaction of sand and metal or borax, none of which are particularly reactive or "go bad" over time. Fuse clips yes, fuse bodies, fuse caps, and other metal parts, sure. But fuses themselves, it's pretty rare that they actually fail to work. Trip mechanisms in breakers on the other hand...good luck with that.

The problem is that with every one of those mechanisms you are going to have corrosion, fretting, and other things that essentially weld or glue the contact tips together. That's the minor problem. The bigger problem is that except for very small breakers and disconnects they all use some kind of greased bearing. The whole idea is that the grease separates the metal surfaces of the bearing and prevents metal-on-metal contact. The whole idea is that the grease is just a very thick oil that moves very slowly on it's own, so it takes a long time to push all the grease out of the joint. But after 30 years, the grease long was long since pushed out of the joint either by spring pressure or gravity. A second problem is that the vast majority of breakers are lubricated with Mobil 28 or something even older. This is a PAO oil with a clay thickener. The clay is fine but the PAO oil long, long since evaporated completely out after 30 years. So even if we can overcome metal-on-metal contact the grease that is in the rest of the bearing is basically a hard ball of clay bound together with a small amount of wax. This stuff gets so hard it takes a hammer and chisel to remove it.

So let's talk about what really happens in this situation. Because I've experienced it first hand multiple times. First off it is best if at all possible to shut the power off. Now get a really good grip on the handle and pull as hard as you can on that disconnect or MCCB. It is HIGHLY LIKELY that the handle is going to break off. If it opens at all the bearings are likely to break out of their housings. MCCB's often just split in half and disintegrate. Hopefully the contact tips don't stick together and shear right off. If you do it under power, LOOK OUT because it will arc, and it will destroy itself. Now with the ICCB's and the power breakers because those use springs or some other source of power other than my arm, they usually don't make it this far. They either open very slowly, destroying themselves in the process if they are under power, or they are just seized up and never move.

Homeowners frequently never experience this in their houses mostly because very small MCCB's will just use brass bushings or in older equipment jeweled bushings for movements. MCCB's are sealed for life so as long as the pressure isn't too high bushings are the preferred bearing mechanism because there is no grease or oil to worry about. However eventually the size gets large enough where this is no longer effective. Data in old IEEE 493 surveys shows that failure rates on 600 A and larger frame MCCB's are 3 times higher. Failure rates on MCCB's are an order of magnitude higher compared to disconnects but in this particular scenario we are only considering the opening mechanism so the reasons for this difference don't matter.

There are things that can be done to improve this. Fluorosilicone breaker grease (Molykote 3451) essentially doesn't dry out. It won't prevent metal-on-metal contact if the breaker is never exercised but it solves part of the issue of lack of maintenance. Also switching to magnetic actuators which is a newer design can reduce the number of moving parts down to a single part that runs on sliding (bushing) bearings. But by way of example S&C makes a fused disconnect where the disconnect is a vacuum interrupter that is mounted in a stainless steel box. The box is filled with SF6 and then plug welded shut. The disconnect operates via stainless steel bellows that go through the box seals. Similarly I believe Siemens makes GIS switchgear where the breakers themselves are again vacuum interrupters in a pressurized SF6 container along with the disconnect. The mechanism itself is all out in the front. The recommended maintenance procedure is to do breaker testing every 10 years (testing the vacuum integrity, etc.), then do some sort of mechanical PM (greasing or replacing parts) at the 20 year mark, finally scheduling the whole thing for replacement at 40 years. This stuff is obviously VERY expensive. BUT even with the top of the line equipment that goes into special military and utility critical systems, they still have to do maintenance on it 3 times in it's entire 40 year life. And that's infinitely more maintenance than running it for 30 years and then attempting to open it.

So the simple answer is that even considering the best of the best of the best equipment that's out there, where even basic PM's are extremely difficult or impossible to perform because of the inherent cost or danger, it still requires something more than 30 years of total neglect.

 

[jim dungar]

IMO, I think a lot of it boils down to how IEEE, NEC and NESC define arcing.

The NEC references to energy reduction are related to arc flash and worker safety. You need to look at NFPA70E. Probably the most popular calculation method is IEEE1584.

 

[mbrooke]

Noted and noted.

All I have to say is WOW! Great post and excellent advice.

Can you parallel list the maintenance for both fused gear and MCCBs?

 

[paulengr]

IMO, I think a lot of it boils down to how IEEE, NEC and NESC define arcing.

When talking about at least NEC and NESC, there is no definition issue. At lower voltages and currents and what I'm talking about here is what is going on inside of neon and fluorescent lamps, we can have controlled discharges. Let's just call those sparks. For safety and power distribution purposes they don't really matter. But as voltage and available short circuit current rise at some point we enter a condition where voltage stops going up or at least goes up very, very slowly but we can push almost unlimited current through the air. This is the power arc or just arc. That's what NFPA 70E and NESC are concerned about when it comes to equipment and personnel safety. NEC has nothing at all to do with it other than requiring arc flash stickers and AFCI's. If you want to play around at the low end where we have streaming, glow discharges, and all kinds of other strange effects, then I'd tend to agree but that's not where most of the standards are concerned.

If two conductors are close enough together and depending on other things like voltage, available fault current (system impedance), and air temperature, we form an arc through the air. This happens almost instantly. Once it does the core of the arc is a very thin (a few millimeters at most) column of air that is highly conductive plasma. The air is so hot that effectively all the electrons are blown off their nuclei and just kind of float around in a highly conductive cloud. The actual arc is very limited in size because plasma is very magnetic so it pulls back onto itself. There are physics calculations for how big it gets but this doesn't normally concern us at all. The arc is almost completely self sustaining. Once the current drops to zero, the arc extinguishes. The air starts to cool down. Then as voltage rises again (in an AC system) and exceeds a threshold generally around 100-200 V, the arc restrikes and goes again, over and over again if the arc is self-sustaining.

So the arc is not really continuous current. If we are working at control/residential voltages (100-250 V), it actually looks more like a square wave since most of the time the arc is extinguished. This makes it hard for the arc to stay ignited. And in terms of AC RMS current drawn, it is a lot smaller than it would be in the case of a dead short through a conductor (a bolted fault). On the other hand for over 1000 VAC the time that the arc stays extinguished is a very small part of the total arcing time so for all intents and purposes, it might as well be almost the same as a bolted fault. So for calculation purposes in the old IEEE 1584-2002 standard, we just set it equal to the bolted fault. In the current (2018) standard the calculation is a little more complex and a little more exact so the arcing current as a percentage of bolted fault current gradually increases.

These formulas are vastly better than the old ones which didn't have much in the way of scientific evidence for them. The current IEEE 1584-2018 standard is based on somewhere over 3,000 tests and the margin of error on the calculated result is around +/-15%. The old (2002) standard had quite a bit of data on error rates and details into how the calculations actually work. Among other things the experimental tests after grinding through all the math tended to have a "bimodal" distribution where the results were one of two numbers. The calculation procedure used a somewhat arbitrary way to fix this (run the calculation the normal way, then with the current calculated at 85% of the original value) but the new one gets rid of this extra fudge factor because the distribution is no longer bimodal. The new (2018) standard doesn't have quite as much detail but there is enough there to know that it is better than the 2002 numbers.

There really is no mystery at all when it comes to arcing faults. It's just that bolted faults and the associated calculations have been around almost since Ohm's Law and when physicists were still working out how it all works in the 19th century. Arcing current calculations have only been around for about 25 years because for practical cases we had to spend a lot of money on a lot of experiments to nail down all the details. But that work has been completed so now there should be very little mystery about it.