Surge prtector failier

[DREPANDC]

We were asked to a customers site to investigate several surge protectors that failed. The failer was charaterized by sever heat build up in the MOV board, melting of the plastic enclosure, smoke and damage to the desk top.

The devices were pluged into modular furniture. We suspected an open neutral and checked all internal furniture connections, branch circuit connections and visually checked the source panelboard.

We contacted APC (surge protector manuf.) and verified that this type of failer has been repoerted before and is cause primarily by an open neutral on a multiwire circuit creating an over voltage.

System is a seperatly derived via a 120/208v 3p4w 75KVA xfmr.

To date we have not found the root cause or evidence of an open nuetral.

There are two items we discovered that concern me.

1) The branch circuits have a seperate #12 equip. grounding conductor that terninates at the panelboard on a buss with non factory screw connections to the enclosure. There is no eq. gnding. conductor from the panel to the xfmer and we may have a high resistance ground. We are planning to test the resistance to ground via a clamp on ground resistance tester (AMEC 3711)

2) With the furniture completely disconnected (no load and phase, neutral and ground conductors disconnected) we have 6-8V betwenn N and G. This is true at the furniture connection and at the panel.

Any thoughts?

 

[George Stolz]

To date we have not found the root cause or evidence of an open nuetral.

What were your voltage readings from each phase to neutral, and each phase to the grounding conductor? It sure sounds like an open neutral.

1) The branch circuits have a seperate #12 equip. grounding conductor that terninates at the panelboard on a buss with non factory screw connections to the enclosure.

Can you describe the "non-factory connections"?

There is no eq. gnding. conductor from the panel to the xfmer and we may have a high resistance ground.

Is there a system bonding jumper installed in the panel, connecting the neutral to the feeder/branch circuit EGC's? Per 250.30(A)(1), it can be located anywhere from the transformer to the first disconnecting means.

We are planning to test the resistance to ground via a clamp on ground resistance tester (AMEC 3711)

For what purpose?

...we have 6-8V betwenn N and G. This is true at the furniture connection and at the panel.

Any thoughts?

It sounds like there is no system bonding jumper for the seperately derived system. I don't believe this could reproduce an open neutral condition, so you likely have two problems, and happened to stumble across the lack of a bonding jumper in the process.

Without that jumper in place, a ground fault will not open an overcurrent device - this needs immediate attention.

If that is the case, it denotes a serious mistake on the installer's part, and would explain why a bad neutral connection exists in the system. Sloppy folks seldom make one isolated mistake.

 

[don_resqcapt19]

Did they all fail at the same time? I have seen that failure on those devices with no external problems in the wiring system.
Don

 

[DREPANDC]

What were your voltage readings from each phase to neutral, and each phase to the grounding conductor? It sure sounds like an open neutral.

118 to 119 for P-N and P-G - If there was an open N it was intermitent, we found no evidence of that.


Can you describe the "non-factory connections"?

From the furniture there is a whip that snaps into the furniture (factory connection) and the othere end connects to the branch jbox via faceplate and wire nuts


Is there a system bonding jumper installed in the panel, connecting the neutral to the feeder/branch circuit EGC's? Per 250.30(A)(1), it can be located anywhere from the transformer to the first disconnecting means.
Yes - in the xfmr, looks fine, XO, frame, N, and Grnd Elec Conductor all bonded corectly

For what purpose?
To check to see if we have 2 problems, an intermitant loss of netral as well as a high resistance path to ground


It sounds like there is no system bonding jumper for the seperately derived system. I don't believe this could reproduce an open neutral condition, so you likely have two problems, and happened to stumble across the lack of a bonding jumper in the process.

Sound possible will investigate further

Without that jumper in place, a ground fault will not open an overcurrent device - this needs immediate attention.

If that is the case, it denotes a serious mistake on the installer's part, and would explain why a bad neutral connection exists in the system. Sloppy folks seldom make one isolated mistake.

I agree - thanks

 

[DREPANDC]

Did they all fail at the same time? I have seen that failure on those devices with no external problems in the wiring system.
Don

They have occured randomly over the last 6 months on 3 or four different days, with multiple fails on any one day. Information is suspect, 10 different people 10 different stories.

 

[petersonra]

They have occured randomly over the last 6 months on 3 or four different days, with multiple fails on any one day. Information is suspect, 10 different people 10 different stories.

MOVs are known for catastrophically failing.

 

[brian john]

We had a customer?s site with the exact problem and had trouble tracing it down, I had the electricians disassemble the systems furniture. Seems it only happened when one guy placed his feet on his desk and the when he wanted more coffee pushed his chair back with his feet on the edge of the desk. This pressure forced the system furniture apart stressing the connector and opening the neutral termination in the flexible connector.

The fix was no more coffee for that employee. What we actually did was order new connectors, seems the contractor that installed the system furniture utilized used connectors, some of which had broken shells and would not adequately lock.

 

[dereckbc]

2) With the furniture completely disconnected (no load and phase, neutral and ground conductors disconnected) we have 6-8V betwenn N and G. This is true at the furniture connection and at the panel.

Any thoughts?

Right there is your problem. With no load that volatage has to be a big fat 0 with no exceptions. Something is open that shouldn't be. If this is derived from an isolation transforner, start there and verify Xo bond and work down stream.

 

[brian john]

We have investigated numerous problems like this and tracked them dowm to faulty system furniture connectors at the corners, where the factory connectors plug together. Took a kick, push, jolt or jar to open the neutral and then smoke.