Heat tracing GFCI neutral bonding

[phineascage]

Forgive me if this has been covered before. I have so far not found the answer to my question. I am struggling with a GFCI type breakers feeding heat trace circuits that immediately trip when energized. These are Square D Series 2 Type QO breakers. The heat tracing is Raychem 5BTV2-CT self regulating heating cable. I have tested and retested the heat tracing as the manufacturer suggests, megging the circuit at various voltages up to 2500V and measuring continuity between the heat trace conductors. The insulation resistance test provided 1T ohm and the continuity between the conductors was around 30ohms for a roughly 20' run of the heat trace. These values are well within design specifications of the heat tracing. I am now talking to online tech support with Schneider/Square D and the best I can get from the person I'm talking to is that the grounding and neutral must be segragated in the sub panel. Our system consists of a delta-wye transformer 480/208VAC transformer feeding an outdoor heat trace panel. Initially the neutral was bonded to ground in this heat trace panel. I removed this ground connection and still received the same results, breaker immediately tripping when energized. I contacted Scheider again and this time was told the same thing but also that continuity between the neutral and the panel ground must be greater than 1M ohm. In our system I measure roughly 7 ohms from neutral to panel ground which can only mean that the neutral bonding to ground is happening upstream from the heat trace panel. I have asked for documentation spelling out grounding and neutral requirements for these breakers but the response I received was that this is an NEC question. I then asked for documentation spelling out the greater than 1M ohm continuity requirement between neutral and ground for these breakers and the response I received with this was that this is an industry standard and not spelled out at least in their documentation. Have any of you guys run into this sort of issue before? I know the heat trace circuit and breaker are fine. I have operated the heat tracing from a 120VAC GFCI receptacle with no problems. I have swapped the circuit feed from one Square D breaker to a spare unused one in the panel and received the same immediate trip. I can disconnect the circuit at the breaker and it will not trip. The test button on the breaker also trips the breaker. I am by no means a good electrician so I was hoping to get some help here. I appreciate anyone reading through this long winded description. If anyone has any questions, please ask. Thank you.

 

[don_resqcapt19]

Forgive me if this has been covered before. I have so far not found the answer to my question. I am struggling with a GFCI type breakers feeding heat trace circuits that immediately trip when energized. These are Square D Series 2 Type QO breakers. The heat tracing is Raychem 5BTV2-CT self regulating heating cable. I have tested and retested the heat tracing as the manufacturer suggests, megging the circuit at various voltages up to 2500V and measuring continuity between the heat trace conductors. The insulation resistance test provided 1T ohm and the continuity between the conductors was around 30ohms for a roughly 20' run of the heat trace. These values are well within design specifications of the heat tracing. I am now talking to online tech support with Schneider/Square D and the best I can get from the person I'm talking to is that the grounding and neutral must be segragated in the sub panel. Our system consists of a delta-wye transformer 480/208VAC transformer feeding an outdoor heat trace panel. Initially the neutral was bonded to ground in this heat trace panel. I removed this ground connection and still received the same results, breaker immediately tripping when energized. I contacted Scheider again and this time was told the same thing but also that continuity between the neutral and the panel ground must be greater than 1M ohm. In our system I measure roughly 7 ohms from neutral to panel ground which can only mean that the neutral bonding to ground is happening upstream from the heat trace panel. I have asked for documentation spelling out grounding and neutral requirements for these breakers but the response I received was that this is an NEC question. I then asked for documentation spelling out the greater than 1M ohm continuity requirement between neutral and ground for these breakers and the response I received with this was that this is an industry standard and not spelled out at least in their documentation. Have any of you guys run into this sort of issue before? I know the heat trace circuit and breaker are fine. I have operated the heat tracing from a 120VAC GFCI receptacle with no problems. I have swapped the circuit feed from one Square D breaker to a spare unused one in the panel and received the same immediate trip. I can disconnect the circuit at the breaker and it will not trip. The test button on the breaker also trips the breaker. I am by no means a good electrician so I was hoping to get some help here. I appreciate anyone reading through this long winded description. If anyone has any questions, please ask. Thank you.

Assuming a grounded system the resistance between the system neutral to the panel enclosure should be almost zero. Even the 7 ohms is much too high.
There should be a system bonding jumper at a single point in the system where the neutral is connected to earth. At every point on the load side of the system bonding jumper, there should be no direct physical connection between the neutral and the EGC. The resistance between the neutral and the EGC or electrical enclosures should be very very low.

 

[phineascage]

Assuming a grounded system the resistance between the system neutral to the panel enclosure should be almost zero. Yes, grounded system. 480VAC 3 phase MCC feed to a 480/208 dry type ACME delta-wye transformer to heat trace panel. Continuous grounding at MCC, transformer casing, heat trace panel frame and there ground bonded to neutral. At the heat trace panel and per recommendation of Schneider I removed the ground bond from the panel neutral bus. At this time measuring between the neutral bus and ground continuity still exists at ~7ohms. Per Schneider, measuring continuity at this point should yield greater than 1M ohm for their GFCI breakers to function correctly. This is where I get confused. Elsewhere on the site in other heat trace systems with different manufacturers GFCIs the neutrals clearly have continuity to ground and the breakers function as intended. Does this make sense? Ungrounded neutral for heat trace dedicated GFCIs to function properly and not nuisance trip?

Even the 7 ohms is much too high.
There should be a system bonding jumper at a single point in the system where the neutral is connected to earth. At every point on the load side of the system bonding jumper, there should be no direct physical connection between the neutral and the EGC. The resistance between the neutral and the EGC or electrical enclosures should be very very low.

 

[synchro]

I am struggling with a GFCI type breakers feeding heat trace circuits that immediately trip when energized. These are Square D Series 2 Type QO breakers. ... Our system consists of a delta-wye transformer 480/208VAC transformer feeding an outdoor heat trace panel. ... I know the heat trace circuit and breaker are fine. I have operated the heat tracing from a 120VAC GFCI receptacle with no problems.

It's possible that the capacitance between 1.) the conductors and resistive material inside of the cable, and 2.) the surrounding grounded metal braid of the cable is causing a leakage current path for a common-mode current that's tripping the 208V GFCI breakers.

Is the controller for the heat trace cable interrupting both phases of 208V to turn it off, or just one of the phases? If both phases are shut off together (for example with a 2-pole contactor) then you can ignore the comments below.

If just one side of the 208V supply is turned off, then the two conductors and the resistive core of the heat trace cable will be at 120V_AC to ground when the heat is turned off (due to the relatively small resistance of 30 ohms between the heat trace conductors). Therefore the entire distributed capacitance of the cable to ground will be at 120V_AC.

When 208V is applied across the heat trace conductors to turn it on, the in-phase component on the conductors relative to ground is cos(60°) x 120V_AC = 60V_AC. So this will result in a smaller leakage current to ground than if the entire cross section of the cable was at 120V_AC to ground, which is the case when only one side of the 208V is switched off as was mentioned above.

When the heat tracing is operated from a 120V_AC GFCI receptacle, however, the voltage between the resistive material and ground will vary from 0V (neutral side) to 120V (hot side) over the region between the two conductors. And so we'd expect about half of the capacitive leakage current compared to when both conductors are at 120V to ground (the case when only one phase of the 208V is switched to turn it off).

As a test, try operating the heat trace cable from the 120V_AC GFCI receptacle but without the neutral connected to the cable. This would be the same operating condition you'd have if only one phase of the 208V was switched off. If the 120VAC GFCI trips, then if you're switching just one side of the 208V off, it confirms that you need to switch both sides off.

By the way, do you need 5mA GFCI protection vs. 30mA GFEP ?

 

[phineascage]

It's possible that the capacitance between 1.) the conductors and resistive material inside of the cable, and 2.) the surrounding grounded metal braid of the cable is causing a leakage current path for a common-mode current that's tripping the 208V GFCI breakers.
The heat tracing is new, just installed. Per Schneider the breakers are appropriate for the application. Almost certain the feed to this particular circuit is THHN or similar 10AWG in overhead rigid piping. The system is fed 208VAC three phase but all circuits are individual 120VAC fed.

Is the controller for the heat trace cable interrupting both phases of 208V to turn it off, or just one of the phases? If both phases are shut off together (for example with a 2-pole contactor) then you can ignore the comments below.
There is no controller. The system is manual on all the time during freezing weather. A GFCI circuit breaker dedicated to this specific circuit is tripping.

If just one side of the 208V supply is turned off, then the two conductors and the resistive core of the heat trace cable will be at 120V_AC to ground when the heat is turned off (due to the relatively small resistance of 30 ohms between the heat trace conductors). Therefore the entire distributed capacitance of the cable to ground will be at 120V_AC.

When 208V is applied across the heat trace conductors to turn it on, the in-phase component on the conductors relative to ground is cos(60°) x 120V_AC = 60V_AC. So this will result in a smaller leakage current to ground than if the entire cross section of the cable was at 120V_AC to ground, which is the case when only one side of the 208V is switched off as was mentioned above.

When the heat tracing is operated from a 120V_AC GFCI receptacle, however, the voltage between the resistive material and ground will vary from 0V (neutral side) to 120V (hot side) over the region between the two conductors. And so we'd expect about half of the capacitive leakage current compared to when both conductors are at 120V to ground (the case when only one phase of the 208V is switched to turn it off). The heat trace circuit is normally fed 120VAC from the GFCI breaker that is tripping. The system is 208VAC but individual circuits are tapped off at 120VAC with a respective neutral to each. Feeding the heat trace circuit using a pig tail and extension cord plugged into a GFCI receptacle results in normal operations, eg. a heating cable. To me this is little different than it's normal feed from a 120VAC GFCI circuit breaker. I cannot understand why it works from the receptacle and not the breaker.

As a test, try operating the heat trace cable from the 120V_AC GFCI receptacle but without the neutral connected to the cable. This would be the same operating condition you'd have if only one phase of the 208V was switched off. If the 120VAC GFCI trips, then if you're switching just one side of the 208V off, it confirms that you need to switch both sides off.

By the way, do you need 5mA GFCI protection vs. 30mA GFEP ?The circuit breakers which I believe are GFCI and not GFEP, I don't fully know the difference and I base this statement on the way Schneider referred to these Type QO breakers (Ground Fault) are 30mA per the specs I could find on them and are integral to the original system design.

 

[phineascage]

It's possible that the capacitance between 1.) the conductors and resistive material inside of the cable, and 2.) the surrounding grounded metal braid of the cable is causing a leakage current path for a common-mode current that's tripping the 208V GFCI breakers.

Is the controller for the heat trace cable interrupting both phases of 208V to turn it off, or just one of the phases? If both phases are shut off together (for example with a 2-pole contactor) then you can ignore the comments below.

If just one side of the 208V supply is turned off, then the two conductors and the resistive core of the heat trace cable will be at 120V_AC to ground when the heat is turned off (due to the relatively small resistance of 30 ohms between the heat trace conductors). Therefore the entire distributed capacitance of the cable to ground will be at 120V_AC.

When 208V is applied across the heat trace conductors to turn it on, the in-phase component on the conductors relative to ground is cos(60°) x 120V_AC = 60V_AC. So this will result in a smaller leakage current to ground than if the entire cross section of the cable was at 120V_AC to ground, which is the case when only one side of the 208V is switched off as was mentioned above.

When the heat tracing is operated from a 120V_AC GFCI receptacle, however, the voltage between the resistive material and ground will vary from 0V (neutral side) to 120V (hot side) over the region between the two conductors. And so we'd expect about half of the capacitive leakage current compared to when both conductors are at 120V to ground (the case when only one phase of the 208V is switched to turn it off).

As a test, try operating the heat trace cable from the 120V_AC GFCI receptacle but without the neutral connected to the cable. This would be the same operating condition you'd have if only one phase of the 208V was switched off. If the 120VAC GFCI trips, then if you're switching just one side of the 208V off, it confirms that you need to switch both sides off.

By the way, do you need 5mA GFCI protection vs. 30mA GFEP ?

I am very tempted to find the neutral/ground bond at the 480/208 transformer transformer feeding the the heat trace panel and remove the ground bond. Of course leaving the system ground in place. I believe this would produce a floating neutral with some amount of voltage measurable to ground in the heat trace panel but would otherwise cause no problem or reason for concern.

Is this a common practice with GFCI breaker feeds to heat trace circuits? Schneider has stated so, to my total confusion since I have observed in operation the exact opposite, the only difference being a different manufacturer's GFCI breaker. In 2 cases at my site being Siemens and GE in complete contrast to my problem situation with Square D.

 

[synchro]

By the way, do you need 5mA GFCI protection vs. 30mA GFEP ?

phineascage:
The circuit breakers which I believe are GFCI and not GFEP, I don't fully know the difference and I base this statement on the way Schneider referred to these Type QO breakers (Ground Fault) are 30mA per the specs I could find on them and are integral to the original system design.

30mA trip breakers would be GFEP (Ground Fault Equipment Protection).
Are the Square D GF breakers being used single-pole or 2-pole?

 

[Krusscher]

I am very tempted to find the neutral/ground bond at the 480/208 transformer transformer feeding the the heat trace panel and remove the ground bond. Of course leaving the system ground in place. I believe this would produce a floating neutral with some amount of voltage measurable to ground in the heat trace panel but would otherwise cause no problem or reason for concern.

Is this a common practice with GFCI breaker feeds to heat trace circuits? Schneider has stated so, to my total confusion since I have observed in operation the exact opposite, the only difference being a different manufacturer's GFCI breaker. In 2 cases at my site being Siemens and GE in complete contrast to my problem situation with Square D.

You can use GFEP breakers which are less sensitive then GFCI but I don't think it would be wise to remove your ground to neutral bond...

 

[synchro]

I am very tempted to find the neutral/ground bond at the 480/208 transformer transformer feeding the the heat trace panel and remove the ground bond. Of course leaving the system ground in place. I believe this would produce a floating neutral with some amount of voltage measurable to ground in the heat trace panel but would otherwise cause no problem or reason for concern.

GFCI or GFEP breakers require a neutral-to-ground bond on their input side. That's so that a ground fault on their output side will complete a closed circuit around the current transfomer that's inside of the GF breaker, therfore allowing it to detect a ground fault current. If you float the neutral of the transformer, there still might be enough capacitance from the transformer windings to the core that a GF breaker could trip under some circumstances, but it is not the correct way to configure it and meet code requirements.

 

[phineascage]

GFCI or GFEP breakers require a neutral-to-ground bond on their input side. That's so that a ground fault on their output side will complete a closed circuit around the current transfomer that's inside of the GF breaker, therfore allowing it to detect a ground fault current. If you float the neutral of the transformer, there still might be enough capacitance from the transformer windings to the core that a GF breaker could trip under some circumstances, but it is not the correct way to configure it and meet code requirements.I'm thoroughly stumped. I know there is absolutely no real ground fault in the circuit yet the breaker continually trips immediately once energized. This also isn't isolated to one circuit. The others fed from this panel exhibit the same behavior. The only thing towards a solution is basically what Schneider says, remove neutral to ground bonding. I'm confident this would allow the circuit to operate, albeit questionably from a code or even reliability standpoint. However this is the solution offered by the breaker manufacturer. I'm in a precarious place..

 

[synchro]

Just make sure that the neutral which supplies the heating cables is coming from a load side of terminal of a GF breaker, and not from the neutral bar in a panel, etc. This may be obvious, but it is important.

If Schneider is saying to remove neutral to ground bonding, perhaps they were referring to any such bonding on the output conductors from the GF breaker (for example, if there was a N-G bond within a small subpanel the breaker might be supplying).

 

[phineascage]

Just make sure that the neutral which supplies the heating cables is coming from a load side of terminal of a GF breaker, and not from the neutral bar in a panel, etc. This may be obvious, but it is important.

If Schneider is saying to remove neutral to ground bonding, perhaps they were referring to any such bonding on the output conductors from the GF breaker (for example, if there was a N-G bond within a small subpanel the breaker might be supplying).There is a bond from the panel ground to panel neutral bus Panel ground is also tied to the center tap of the secondary or wye side of the panels feeding transformer. Schneider said to remove this this bonding in the panel between the neutral bus and panel ground. I did as recommended.
Continuity between panel ground and the panel neutral bus still exists which must mean there is additional neutral to ground bonding upstream of the panel, presumably in the 480/208 feeding transformer. Schneider additionally stated that for these breakers to function correctly that the panel neutral to ground resistance must be greater than 1M ohm which leaves me the only option to unbond the neutral upstream of the panel. I could not get any documentation from them explaining this. They only said this was common industry practice. I'm befuddled.