Parallelled VFD cabling for submersible pumps 4

[bentov]

A 25hp 460V (32fla) submersible in for repair needs a new cable. The old one is OEM special, SO type cord with 3#4 (load), 1#6 (ground) & 3#12 (seal fail & in wdg thermal + common), $30+ per foot. Though installed with VFD controls as a new package, this cable was not shielded. The pump is dual voltage, original cable large enough for 230V connection (64fla) but will not ever run that way.

We're thinking of using some readily available 12-14 SOOW (14 conductors, all #12), parallelling 3 for each load leg, 2 for the ground, 1 each for the seal fail/therm/common. Any undesireable consequences (other than due to lack of shielding) likely from this kind of parallelling on a VFD load?

 

[davidbeach]

Not legal if the NEC is the governing code.

 

[LiteYear]

A potential consequence given a VFD and a submersed cable is that if the grounds are not well positioned within the cable, the capacitance from the phase conductors to the sheath of the cable may be enough to conduct significant high frequency current into the surrounding fluid. This is potentially more serious than a lack of EMI related shielding, since currents near the switching frequency of the drive (probably a few kHz) can be significant when the capacitance of the entire cable is effectively paralleled by a conductive fluid.

 

[Skogsgurra]

As LiteYear says. Have had significant problems with level transducers being influenced by HF EMI in water close to pump.

Gunnar Englund

http://www.gke.org

--------------------------------------
Half full - Half empty? I don't mind. It's what in it that counts.

 

[bentov]

Thanks for the replies. The original OEM cable in question is the 4AWG/3-2-1-GC on page 5 here:

http://www.flygt.de/1127282.pdf

($29/ft on an OLD price list). Turns out one of the #12's is a Ground Check conductor that is not referenced with any protective component in the panel so should be safe to skip that one - no seal moisture sensor on this so remaining "control core" pair is for a N.C. in-winding thermal circuit.

So to substitute we need 6 conductors (3 load, 1 ground, and 2 control) correctly sized. I presume the OEM cable is adequate, so reference is the low volt amp rating (64a) - per NEC 400-5(a) #4 is good for 70a, so I believe the unit is 1.0 Service Factor, using fla x 1.1 for ampacity (so 64 x 1.1 = 70, whereas for SF 1.15, 64 x 1.25 = 80a so #4 would be too small).

So for the high voltage current we care about, 32 x 1.1 = 40, per 400-5(a) can use #8 BUT I cannot find any SOOW 8-6 (only 8-5, not enough conductors). I do find 12-10, per NEC #12 with 6 current carrying good for 20a so 2 parallel per load leg would work, would need to pair up the ground too though cause 12 not big enough . . . not sure about that (parallelling ground conductors, is that the "not legal" part davidbeach mentions?). 10-10 is also available so might be safer since 1#10 is OK for ground conductor size, could also get our Ground Check conductor back that way just to be thorough. So that covers ampacity, unless I'm missing something.

LiteYear, regarding conductor positioning, would you say then that our 10-10 plan would definitely result in transmission of higher capacitive currents vs the OEM? And would that pose some kind of hazard or VFD dysfunction, or mainly the potential signal interference per Skogsgurra? We can choose (& mark) conductors any way we please - what would be the preferred geometry?

And yes, there are 4-20mA level transmitters in the wet wells, though I don't yet know how near the power cables, or how well any of that ever worked to begin with.

Compounding the misery, this duplex storm drain is only a couple years old, very low hours, but the contractors spliced the cables (non-watertight)in a Christy box below ground level, so the first power outage/flood event soaked into the sub cable, wicked down to the motor unbenownst - VFD's tripped, so the old-school maintenance guy pulled the pumps and took them to his shop, "tested" both with 230V off a circuit breaker, ground faulted (burnt) both windings. While rewinding those we get the message to lengthen the cables about 5 foot so they can raise the splice box. I'd like to save the financially strapped city (where I live) the extra thousand bucks a pop for these cables if possible. Plus, don't know how things will work out till we get them back in and running so I can check operation of the controls . . . maybe what they really need is the NSSHOU/3E screened cables (page 8) that don't even appear in the price list I have - $60/foot? $100 . . . ?

 

[mesutphen]

You reference the NEC, so I assume you intend to abide by it. So, tagging along with what davidbeach said before, you cannot parallel these conductors and be in compliance with the NEC. Section 310.10(H)(1) states conductors "shall be permitted to be connected in parallel (electrically joined at both ends) only in sizes 1/0 AWG and larger..."

 

[ozmosis]

bentov
what is the distance between VFD and motor, and does the motor require the VFD to limit the level of dv/dt? If so, chances are you may need to have a dv/dt filter or sinus filter between VFd and motor. The reason I raise this is to be aware of voltage drop to the motor if the cable is long (an output filter will impact volts drop). This will also determine the type of cable to use if an issue.

 

[LiteYear]

LiteYear, regarding conductor positioning, would you say then that our 10-10 plan would definitely result in transmission of higher capacitive currents vs the OEM? And would that pose some kind of hazard or VFD dysfunction, or mainly the potential signal interference per Skogsgurra? We can choose (& mark) conductors any way we please - what would be the preferred geometry?

Assuming you mean SOOW 10/10, then to be sure I'd have to see the cross section. But looking at the 4AWG/3-2-1-GC, I'd say it would be hard to get much worse as far as capacitive coupling to the sheath goes. There is no earth conductor between the phase conductors and the sheath for more than half the cable, so I would expect fairly high capacitive coupling from phase to sheath. So if you're not having problems with that cable I wouldn't expect too much difference with 10/10.

The kind of problems I'd expect are both EMI issues such as Skogsgurra describes as well as possibly hazardous currents flowing in the surrounding fluid or touch voltages existing on the sheath of the cable. I doubt it would cause problems for the VFD - it probably wont care where those currents are flowing.

And finally, the preferred geometry is to surround the phase conductors with earth conductors. Basically, the closer you get to a shielded cable geometry, the better. A good compromise is to have interstitial earths between each phase conductor, with the earths closer to the sheath than the phases (eg. Type 241 mining cable).