NEC Changes For #14 Ampacity

[FionaZuppa]

Very nice. Your circumferences are off by a factor of 2, and hence the surface areas and the quantities derived from that. Since it is a consistent error, it doesn't affect any of the trends shown.

Cheers, Wayne.

ah yeah... i see it
will fix

 

[FionaZuppa]

NEC #'s

fixed watts/ft

fixed P/A

fixed P/A using NEC #14 ampacity for P/A

 

[FionaZuppa]

several oddities in NEC #'s. the last pic shows P/A fixed to same as NEC #14 P/A, so why do the NEC #'s for ampacity go done so much if 15A on #14 P/A is acceptable?
as shown, taking #14 to 20A (as it were in 2008 nec) takes the P/A # almost 100% higher than 2011 NEC #. but again, the 15A OCD restriction seems to be contradicted by the NEC table itself (such as 75 & 90C column ratings, and allowed exceptions, etc).

is #14 NM-B running at 19A a fire hazard under any allowed wiring scheme?

so. your thoughts?

 

[FionaZuppa]

when did NEC have 25/30/35 for #14 ??

see http://www.houwire.com/products/technical/article310_17.html

everything discussed thus far is "well-known" physics for a long time. i keep scratching my head wondering why all these wild differences between code versions, and lack of evidence to back up the #'s.

if i am THWN-2 in buried conduit with 4% fill and ambient is 40F, ampacity is 28.75A, i cant use 20A OCD unless its one of the exceptions??

 

[FionaZuppa]

from another table (http://www.panduit.com/heiler/SelectionGuides/WW-WASG03 Electrical Wire Sizes-WEB 7-7-11.pdf) i see that OD varies depending on insulation types, this will impact the #'s. so now need to compare P/A for the ampacities across 60/75/90C columns, etc. example, a THWN has a bigger OD than say THHW, thus for 60C ampacity the THWN will have more surface area and thus less P/A. so in essence NEC takes a NM-B rated at 90C into the THHW 60C column, thus also reducing its P/A down to THHW.

 

[wwhitney]

so now need to compare P/A for the ampacities across 60/75/90C columns, etc. example, a THWN has a bigger OD than say THHW, thus for 60C ampacity the THWN will have more surface area and thus less P/A.

The hottest part of the insulation will be that adjoining the copper wire, so I think it is appropriate to calculate P/A at that boundary, in which case the A involved doesn't depend on insulation thickness or outer diameter.

However, the insulation material and thickness will influence the maximum allowable P/A for a given maximum insulation temperature, due to the conductive thermal resistance of the insulation.

At first glance, it is unclear to me whether increasing insulation thickness increases or decreases the heat loss from the copper wire. The heat loss pathway is conductive through the insulation, and then either radiant or convective from the insulation to the surrounding air and environs. If the heat loss from the outer surface of the insulation is the dominant factor, then increasing the insulation thickness would give more surface area and presumably increase the heat loss. While if the heat conduction through the insulation itself is the dominant factor, then increasing the the insulation thickness would reduce the heat loss.

Cheers, Wayne

 

[FionaZuppa]

well, NEC i believe is only concerned about the outside of the wire. the insulation types i do not believe will get anywhere near 60/75/90C for listed NEC ampacities, thus i dont see the insulation temps being an issue for the insulation itself. a breakdown of insulation under the ignition temp may provide arc paths though.

the thicker the insulation the hotter the wire (metal) will be because the insulation thermal resistivity will go up, etc. this means hotter copper but cooler insulation, Temp(F)/sq.in at OD will go down. this is why i dont fully understand NEC ampacities, wire with Goxel are rated much higher w/o insulation damage. what was the primary motive of the NEC ampacity # making??

the primary modes of heat transfer are most definitely conductive to the OD, and then convective from there. even in the metal wire its conductive. most liquid "wires" would have convective as primary mode withing the wire itself.

 

[wwhitney]

well, NEC i believe is only concerned about the outside of the wire.

Who knows where the NEC got its numbers. As a performance issue, if the insulation is not to exceed, say, 75C, that would mean everywhere, so we had best look at the hottest point in the insulation. That will be the inner boundary of the insulation against the heat source, the wire.

That raises one other possibility for why the NEC's allowed P/A decreases with wire size. If for some reason larger wires were more likely to have discontinuities or other features that could lead to local hot spots, then the need to protect the insulation at those hot spots could lead to a reduction in the allowed P/A.

the thicker the insulation the hotter the wire (metal) will be because the insulation thermal resistivity will go up, etc.

Not necessarily. If the insulation is a good conductor of heat relative to how well the outer surface of the insulation can convect away heat, then thicker insulation could give you more convective area and lower the temperature of the wire. Obviously that is true in the limit case that the thermal resistance of the insulation is 0. Whether it is true in practice I have no idea.

Cheers, Wayne

 

[FionaZuppa]

Who knows where the NEC got its numbers. As a performance issue, if the insulation is not to exceed, say, 75C, that would mean everywhere, so we had best look at the hottest point in the insulation. That will be the inner boundary of the insulation against the heat source, the wire.

That raises one other possibility for why the NEC's allowed P/A decreases with wire size. If for some reason larger wires were more likely to have discontinuities or other features that could lead to local hot spots, then the need to protect the insulation at those hot spots could lead to a reduction in the allowed P/A.


Not necessarily. If the insulation is a good conductor of heat relative to how well the outer surface of the insulation can convect away heat, then thicker insulation could give you more convective area and lower the temperature of the wire. Obviously that is true in the limit case that the thermal resistance of the insulation is 0. Whether it is true in practice I have no idea.

Cheers, Wayne

its no different than our std fiberglass insulation. the thermal resistance goes up as the insulation gets thicker. when insulation gets very very thick you are then basically back to heat transfer via conductive mode and when radial insulation thickness goes to infinity then P/A approaches zero, thus thermal R is significant at this point and P/A becomes non-significant. this doesnt mean the wire temp goes to infinity, just means there's a R value for the thermal transfer and a equilibrium point at which the metal to insulation interface will converge to. the primary mode of heat transfer within solids is conduction.

as for the insulation performance, need to verify what the UL test specifies. my guess is the UL tests are measuring temp on the OD.

 

[mbrooke]

Just tossing a curve ball into this.

What gets hotter, the terminals or the copper itself during an overload?

 

[FionaZuppa]

Just tossing a curve ball into this.

What gets hotter, the terminals or the copper itself during an overload?

an overload means what? is an overload +5% above rated OCD ? i suspect any junction point, especially if its dissimilar metals, will have a slightly higher ohms/ft, but since the junction point is very short i dont know if the I^2R at the junction point would have higher P/A than the P/A of same length of wire.

the idea of junction temps wrapped up into the NEC ampacities is plausible, but this notion seems to carry a "whaaaat" in it when romex is bound to the 60C column, and, blatant restriction that #14 must be on 15A OCD(max), #12 on 20A OCD(max) (exceptions apply). the nec verbiage should have no regards for equipment OTD/OCD devices (like some motors have), just as nec has no regards for equipment wiring, thus the exceptions seem baffling. nec should only be concerned about the circuit (bc's here) from OCD to equipment.

i am all ears, tell me how NEC gets their numbers.......

 

[mbrooke]

My mistake, by overload a current in excess of the listed current carrying capacity.

When a shard neutral is misapplied, its usually the terminals that seem to have the most heat damage.

 

[FionaZuppa]

My mistake, by overload a current in excess of the listed current carrying capacity.

When a shard neutral is misapplied, its usually the terminals that seem to have the most heat damage.

the item you mention seems to not be an issue given the allowed exceptions. i was not asking for #14 to run on a 30A OCD, this has more to do with the derivation of the ampacity #'s, 90C NM-B restricted to 60C column, the blatant restriction on #14 and #12 for OCD, and the allowed exceptions. to me these four items seem to conflict with each other.

when is there a shared neutral with #14 or #12 wire? a MWBC?

 

[wwhitney]

its no different than our std fiberglass insulation. the thermal resistance goes up as the insulation gets thicker.

Sure, I agree, I am just pointing out that as the insulation gets thicker, the outer surface area of the insulation gets larger, so the the insulation can shed heat to the air more quickly. Depending on the relative magnitudes of the conductive heat flow through insulation versus the radiant/convective heat loss from the outer surface of the insulation, it could be that thicker insulation sheds heat better overall.

For an example with made up numbers, suppose the heat loss flux on the outer surface of the insulation is proportional to the temperature difference (not true for radiant heat loss, I don't know about convection). Suppose a wire 100 mils in diameter has insulation with a thermal conductivity of 200 BTUs-mil/in^2/sec/degree, the surface heat loss is 1 BTU/in^2/sec/degree, and the wire is generating 10 BTUs/in^2/sec at its outer surface.

If the insulation is 10 mils thick, then its thermal conductivity is 20 BTUs/in^2/sec/degree, so to move 10 BTUs/in^2/sec through it requires a temperature difference of only 0.5 degrees. The outer diameter of the insulation is 120 mils, versus 100 mils for the wire, so the surface area of the insulation is 1.2 times as great. So we need to dissipate 10/1.2 = 8.3 BTUs/in^2/sec from the surface of the insulation, requiring a temperature difference of 8.3 degrees. Thus the inner surface of the insulation is 8.8 degrees warmer than ambient.

If instead the insulation is 50 mils thick, then its thermal conductivity is 4 BTUs/in^2/sec/degree, so to move 10 BTUs/in^2/sec through it requires a temperature difference of 2.5 degrees. The outer diameter of the insulation is 200 mils, versus 100 mils for the wire, so the surface area of the insulation is 2 times as great as the wire. So we need to dissipate 10/2 = 5 BTUs/in^2/sec from the surface of the insulation, requiring a temperature difference of 5 degrees. Thus the inner surface of the insulation is 7.5 degrees warmer than ambient.

In practice, I have no idea what the relative magnitudes involved are. I'm also unclear on whether I calculated the conductive heat flow through the cylindrical shell of the insulation properly, as at that step I didn't account for the increasing shell area with increasing diameter.

Cheers, Wayne

 

[mbrooke]

Doesnt all insulation have a bell curve? At one point more insulation reverses the trend?

 

[FionaZuppa]

Doesnt all insulation have a bell curve? At one point more insulation reverses the trend?

For general scientific use, thermal conductance is the quantity of heat that passes in unit time through a plate of particular area and thickness when its opposite faces differ in temperature by one kelvin

the thermal conductance will approach zero as thickness increases to infinity, there is no curve shape change.
when thickness L gets very large the conductance becomes insignificant and specific heat capacity becomes significant in the equations (this can be approximated by say putting a big hunk of AL into a styrofoam cube that has 6" thick walls and the heat source (#14 wire) is held firm to one of the AL block surfaces. the heatsink flux from wire into the AL will have max flux at t=0, heat flux will decrease until equilibrium is reach, mostly governed by specific heat capacity of the AL. there is no heat loss flux of the system as we approximated the 6" thick walls to be lossless for the experiment, etc.

but in this context we are looking at insulation thickness that is required to be proper dielectric for the AWG wire sizes, thus relatively very small. anyone know if any of the UL tests are swift enough to measure the OD temp of the wire (metal) with varying currents in STP free air environment? i think as a basis it would be good to start off by knowing what amps in the wire brings the OD temp of the wire to 60/75/90C.

and i was not aware this was how OCD's were made, thus wire can see 200% OCD rating for 2min (for 20A OCD, i assume this trip is under STP conditions).

Actually, a 20 amp breaker must trip at
a sustained current of 27 amperes (135 percent) at less than one hour, and at
40 amperes (200 percent of wire rating) in less than 120 seconds—far differ-
ent from what the cited text implies. These two trip points (135 percent and 200 percent) are defined in NEMA Standard AB-1, MCCBs and Molded Case Switch-es[2]. TABLE 1 lists the 200 percent allowable trip times for different size (amperage) circuit breakers.

 

[Tony S]

Mr. Brooke, there is a practical test you can run but it’s on another channel.

 

[mbrooke]

Mr. Brooke, there is a practical test you can run but it’s on another channel.

:lol: You are not far off.

All I need is an amp clamp, thermo-couple, variac and a high current transformer. A few tests in air and a few in fiber glass. :thumbsup:

 

[Tony S]

If I was still working it would give me something to do on weekend afternoons. It wouldn’t have been difficult with all the old junk we had about the place.

The 600/1000V grade T+E breakdown voltage test was an I’m bored Sunday afternoon project. It would be interesting to compare T+E to Romex

 

[Tony S]

In case anyone wonders what we’re on about

600/1000V grade 1.5mm² T+E (flat twin and earth)
(600/1000V grade is rated 600V to earth and 1000V between Ph and neutral. You could get slightly cheaper 300/500V grade, I wouldn’t use it 300V is to close to the 250V to earth we use)

A 1 metre length of 1.5mm² T+E withstood 16KV for 10 minutes, it broke down at 17.5KV