VFD Cable Code requirements 4

[buzzp]

Someone shoot me now, my boss wants to know where the requirement is to have cable going to VFD's being VFD cable (XLPE and perhaps shielding of some sort). I told him I know it is a general requirement in Article 110 (110.3). He was looking for a more specific requirement. I told him I don't think there is one. I couldn't find one...is there such a more detailed requirement in the NEC other than it violates basic engineer principals cause I know the insulation will break down over time. These will be for a 125HP and a 500HP VFD.

 

[GroovyGuy]

Hey Buzz,
I have never used a special cable for VFD applications. In Canada I have used standard unshielded TECK for everything up to 4.16kV, shielded for 6.8kV and up. In the US I have used standard unshielded MC on all 480V drives. The special drive cables, ie shielded and triplexed ground conductors are not warranted for 99% of cases.
I have never heard of the cable insulation breaking down due to a PWM voltage waveform.
GG
Note: On most LV drives I usually install a dv/dt filter at the drive output. This helps to protect the motor insulation if the feeder length is excessive.


"I have not failed. I've just found 10,000 ways that won't work." Thomas Alva Edison (1847-1931)

 

[jraef]

The NEC is not a design guide for best practices, it's a list of minimum requirements mostly for preventing fires. So as you have observed, there is nothing dictating the quality of an VFD installation, just a few rules surrounding conductor sizing etc. So if your boss must have a gun at his temple in order to justify doing things right, you are not going to find one for him.

BEST PRACTICE however is XLPE insulation on output conductors, plus some sort of shielding. Can you use unshielded cable? Sure. But you are creating relatively powerful local radio antennae. If you have nothing around the area that will be affected, nobody knows. But I for one have witnessed very dramatic effects of having no shielding, including a visit by military police in the desert near Las Vegas who triangulated on us while we ran several 5HP motors approximately 25 miles from a sensitive government installation. When industrial monitors were CRTs it was also a lot more noticeable, but I haven't seen a CRT in use in a while now.

The issue on the cable insulation has to do with thermoset plastic (PVC) used in conventional building wire like THHN. The insulation is injected into molds with the wire running through it as a liquid, and as such there can be microscopic bubbles in it. As it is set up with heat, some of those bubbles get trapped in the plastic. At 600V sine wave power these are irrelevant, but in the presence of high voltage standing waves that a PWM output can create, the peak potential between phases can easily exceed 2000V on a 480V drive. In that environment, those bubbles represent a weakness in resisting partial corona discharge and once it starts, the cable fails quickly. It may never happen, but luck is not a valid engineering strategy.


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[buzzp]

Thanks for the reply. These are 480V drives. There is no filtering on this drive. It is a unit the customer is giving us from another facility (this by itself screams liability to me, what if the drive don't work after it is moved?)
It sounds as though you aren't using any special insulation neither.
I have seen failed insulation on cables feeding PAM type drives (submersible motors in the oil field). However, it is unknown if it was due to the drive or something else. Thanks again.

 

[GroovyGuy]

hi Buzz,
The submersible drive is the 1% of problems that I did not detail in my earlier message. This is primarily a result of the long feeder lengths.
I would suggest that any triplexed and armoured cable (ie type MC) should be fine for your application. If the feeders are long, or if the motors are old, I would suggest a dv/dt filter be placed on the drive output. You can fit these into the existing drive cabinet (if there is room) or you can add them externally (ie in a can).

GG

http://www.transcoil.com/Products/V1k-dv-dt-Motor-Protection-Filter.htm

"I have not failed. I've just found 10,000 ways that won't work." Thomas Alva Edison (1847-1931)

 

[EdStainless]

Inside building we always use upgraded insulation and shielding to reduce noise, outside we don't care.
Though I can tell you that for long lengths (14,000' down a well) cable quality really matters and standard sizing rules don't cut it even with filtering.

= = = = = = = = = = = = = = = = = = = =
P.E. Metallurgy, Plymouth Tube

 

[buzzp]

Jraef, Your reply must have come right when I posted mine.
I think this drive is going to be about 150' away from the motor in this case. It is a drive the customer has offered up for us to use (which, to me has potential issues to begin with). There is no filtering on it at all from what I have been told. This facility is in the middle of no where(probably what you thought until the military police paid you a visit). They have PLC's spread throughout the facility.
I appreciate the feedback and there is obviously some different ideas on what is needed and what isn't vs. what is required. I realize the NEC is for safety, etc. vs. a design guide but my boss don't understand this. I usually like to be conservative on decisions like this but the boss man don't and will probably require me to use non-VFD cable. I am going to push Article 110 on him to see if I can convince him otherwise. Keep the debate going. Thanks again for all the feedback. I appreciate all of your opinions.

 

[waross]

I hope that at least you can use 1000 Volt rated XLPE cables.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[jraef]

Coincidentally I just had a situation pop up today where a contractor installed a VFD for a 75HP pump and it is about 80ft from the drive to the motor. In spite of my suggestions, they ran PCV conduit, THHN cable, no shielding, but figured that "since it was buried 24" deep it will be fine". It isn't. There is a pressure transducer on the pipe near the motor and every time the VFD fires up, the pressure values become erratic. they have a bypass starter around the VFD and when they run it in bypass, no problems. They have tried re-running the wires to the transducer with better shielding, they even wrapped foil around it! No help (of course). They happened to have an underground cable locator and ran it over the motor cables, it's screaming like a banshee. I have suggested pulling those motor cables back out and using shielded VFD cable, I am not popular today...


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[buzzp]

They even used PVC with no cable shield - that's bad. I will win the shielding argument if no other reason than using RMC/IMC conduit but it still isn't the best. I take pride in providing the best installation I can, within reason. Using the proper insulation and shielding on VFD's is within reason as far as I am concerned. Could you get away with sub-standard cable? Maybe, maybe not. You would know fairly quickly about the lack of shielding consequences (until they put in new equipment) but the insulation failure might be days or months down the road after we are long gone in which case my boss says we don't care - wow. Time to update the resume after only months on the job. Too many red flags popping up for me.

 

[Marke]

Hi Buzzp

Most VFD manufacturers recommend the use of a screened VFD cable, and in some countries there are recommendations from the authorities as a means of reducing interference with other equipment installed near the VFD installation.

The mechanism of the interference, is due to the PWM output waveform generated by the VFD and the capacitive coupling of high frequency currents into the frame of the motor due to the fast switching edges of the voltage waveforms and the capacitance between the motor windings and the stator of the motor.
The very vast switching waveform causes a very high charging pulse current into the from of the motor.
This current needs to return to the source, and the source, is the DC Bus of the VFD.
European drives typically have small capacitors connecting the frame of the VFD to the plus and minus rails of the DC bus (Top and Bottom of the IGBT output switches). This is commonly referred to as level 1 EMC filtering. If there is an earth conductor between the frame of the motor and the frame of the drive, there is a return path for this current. The problem is that the pulses are very very short and very steep slopes, so the energy is concentrated at high frequencies, typically centered around 150KHz with modern drives.
At 150 KHz, the current flows on the very surface of the conductor only. The impedance of the conductor is a function of the surface area of the conductor, not the cross sectional area of the conductor.

If the phase conductor from the VFD to the motor and the earth return from the motor to the VFD are equal cables, then we have a potential divider and this can result in transient voltages close to half the output voltage.
Reducing the impedance of the return path will reduce the voltage on the frame of the motor.
A screened cable with the screen correctly terminated at both the motor end and the VFD end, will provide an earth return path for the high frequency currents with an impedance that is typically in the order of 50 times lower than the phase conductor and so the motor transient voltage is reduced by a factor of 50.

The higher the transient noise voltage on the frame of the motor is, the greater the stray current flowing through all connected circuits and structures. The current will flow through all possible paths to find it's way back to the DC bus and that can include a return via the transformer and the input rectifier, plus EMC capacitors connected from earth to phase in numerous other electrical components such as timers etc.

The interference is primarily at frequencies below 1 - 2 MHz and is essentially conducted noise rather than radiated.
The screen in a screened cable is one way to provide a noise high way home. In many cases of severe interference, I have corrected the problem by providing an alternative path using a suitable conductor with a significant surface area. Typically, I use a very wide flat braid. I have used galvanized water pipe, aluminium strip, steel angle, almost anything that is conductive. The important factor is the surface area and I recommend the equivilent surface area of the screen of an appropriate screened cable slit open and laid flat.
Foil does not work effectively because it is too thin. I recommend the conductor should be at least 1.5mm thick.

Best Regards,
Mark

Mark Empson
Advanced Motor Control Ltd

 

[buzzp]

Thanks for the information Mark, it was helpful. If you separate the ground from the phase wires, it would seem this would increase the impedance. If you agree, it would seem this may potentially cause issues with a breaker not tripping in the event of a fault (USA - I understand Europe doesn't use grounding the way we do). I suppose it is somewhat of a fine line. I am trying to wrap my head around all the potential issues so excuse me if I misstated something and feel free to correct me if you think otherwise.
This particular customer did have the 500HP motor fail within a few weeks of it getting installed. There were two potential issues and they fixed them both - motor wasn't inverter duty rated (motor burned at end bearings, believe this is typical of failure due to VFD operation and insulation breakdown) and the cable was not cable but all single conductors in metal conduit. The manufacturer of the equipment put in an inverter duty rated motor and at the same time, our customer installed XLPE cable with some sort of shielding (I don't have specs or PN of cable yet). Where and how the shielding was terminated, I don't know. I guess it has been operating smoothly for a few years now. So, needless to say, the customer is zeroed in on what we are doing when we move this motor and VFD to their new facility. I believe their motor failure was most likely due to the insulation of the motor vs. any issues with the cable.
Just curious, in your experience have you seen a cable failure that could most likely be attributed to operating a PWM type VFD using standard insulation? I have no doubt that it happens.
Thanks to all for your replies!

 

[Marke]

Hello Buzzp

No, it is not the proximity of the earth return to the phase conductors that reduces the impedance, it is the surface area of the conductor. At 150KHz, most of the current flows in the outer 0.5mm of the conductor and nothing flows in the center.
If we take a standard circular conductor, at 50 Hz or 60Hz, the impedance is a function of the cross sectional area and length. At 150KHz, we could hollow out the enter of the conductor and leave a akin of say 1.5mm thick and the impedance would not change significantly, so per weight of copper, we want a thin conductor for minimum high frequency impedance. The screened cable is a low cost way of achieving a conductor with a high surface area for the return path.
I have solved numerous interference problems where the return path has taken a very different route than the phase conductors.
Issues that I have seen have been interference with the operation of flow meters and pressure transducers, I have seen numerous electromagnetic flow meters that have been damaged by the conducted noise currents. I have cured problems of interference with communications systems, radio and serial comms caused by the conducted emmisions.

I have not come across cable damage that I would associate with the output waveform of a VFD, but have seen motor failures that have been reported as first turn failures. More detailed analysis has shown this is not the case and have actually only seen two definite first turn failures due to VFD waveforms.

Bearing failure is a separate problem, once again the switching waveform causes currents to flow in the motor bearings via a number of different mechanisms.
The motor should be fitted with an insulated bearing at the non drive end, and a shaft brush at the drive end. There are different insulated bearing designs which have different results in reducing the bearing currents. Where the currents are inductively coupled, it is preferable to minimise the capacitance of the insulated bearing and in this case, the hybrid bearing is preferable.
A common mode choke on the output of the VFD will also help to reduce the bearing currents.
NB. This a separate problem from the conducted emissions interference issue, but it does also originate from the fast switching PWM waveform on the output of the VFD.

To minimise conducted emissions interference, you need a low impedance return path from the motor to the DC bus of the VFD.
To minimise bearing failure, you need an insulated non drive end bearing and a shaft grounding brush at the drive end.
A common mode choke on the output of the VFD is helpful in both cases.
To minimise the first turn failure in the motor windings, you need to limit the dv/dt applied to the cables at the drive end. This is best achieved with a true sinusoidal filter that provides common mode and differential mode sinusoidal filtering. dv/dt filters will act to reduce the transient voltage amplitudes.

Best regards,
Mark

Mark Empson
Advanced Motor Control Ltd

 

[waross]

Don't forget that there are two different possible ground currents.
1> The high frequency induced current caused by the PWM,
and
2> The line frequency current in the event of an earth fault in the motor.
The skin effect that Mark describes is a high frequency effect and the effect at 50 Hz or 60 Hz is slight.

Bill
--------------------
"Why not the best?"
Jimmy Carter

 

[jraef]

Motor fault current, to ground or bolted, is seen by and reacted to by the drive, but has little effect on the incoming line. The drive will trip off long before the drives rectifier draws enough current to even begin to trip a breaker, so the breaker doesn't really react to down stream faults. For the most part the only thing the breaker/fuses on the line side of a drive are there for is to prevent a fire in case the rectifier shorts out.

Mark,
Most installation specs I see require that the ground conductor have at least the same circular area as the power conductors, which I understood to be because of the HF common mode noise issue you described. Are you saying this is inadequate? Here, we require that the motor frame, the conduit and the drive ground connections bonded to the building ground. Would not all of those return paths be adequate?


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington

 

[buzzp]

I should of clarified that my concern with clearing a fault would be at 60Hz but then as pointed out, the drive will trip before the breaker or at least get into current limit, assuming it is after the VFD. Which part contributes more to the overall impedance at 60Hz, the capacitive or inductive part of the cable? If the distance between the hot and neutral is increased the impedance (C part) goes up. For VFD's, I suppose this isn't a real concern since it is fed from the VFD. If there is no VFD then this could create fault clearing issues.
Good discussion!

 

[Marke]

Hi Jeff

If you take the cross section of the phase conductor and flatten that out, you get the same cross section and much more surface area, so yes, you meet both criteria.
Some cable sources charge by the weight of copper, so the cost is not too much of a premium when a screened cable is used.

Bonding the motor frame and VFD to the building ground does work provided that the surface area along the whole length is adequate. The big problem that I find is that most electricians do not understand how important the surface area is and use standard conductors (pigtails) to do the bonding. I describe it as like using a 6 inch water pipe with a length of half inch. The flow is determined by the half inch, not the six inch.

EMC cable tends to have both a screen, sometimes a double screen, and three protective earth conductors internally.

I some regions, there is a requirement to use metal conduit, and bonded correctly, that provides a good high frequency return path. In most European influenced regions that I have experienced, the use of plastic conduit is more usual.

Best regards,
Mark.

Mark Empson
Advanced Motor Control Ltd

 

[jraef]

I agree on all of your points Mark. When I teach my VFD-101 class for electricians and get to this issue, I usually get a lot of pushback on using shielded (screened) VFD cable, mostly because of the perceived added cost burden. I always then add "Compared to what?" and get into the details. One of those details, when I tell them it's OK to use separate (XLPE) conductors in steel conduit is the subject of RIGOROUSLY adhering to proper bonding components and methods, which leads to admissions by most electricians that they are sloppy about that in many cases, relying on the simple lock nuts to "bite" though paint and coatings to bond to bare metal. That's one of the big fallacies in that. But I like your analogy to flattening out the cable to get the surface area needed, I'm going to plagiarize that, but I'll send you a nickel every time I use it...

PVC conduit is very popular here too, albeit often specified as forbidden for VFD outputs unless shielded cable is used when a consultant is aware of these issues. Unfortunately, electricians are quite often left to making these decisions on their own and if they are contractors, will tend to base their decisions on economic rather than engineering principals. "If they don't specify it, they get the cheapest thing I can legally use". I was involved in a local informal committee to provide input for changes to the next cycle of our NEC (National Electric Code) and raised this issue, saying the we should codify the available options when it comes to output wiring. I was almost shouted down by mostly contractors on the committee, all of which claimed that they have never seen problems. My response to that is always that "Luck is not a valid engineering strategy.", but it generally gets ignored.[pre][/pre]


"You measure the size of the accomplishment by the obstacles you had to overcome to reach your goals" -- Booker T. Washington