NEC Changes For #14 Ampacity

[mbrooke]

after 30min cooking NM at 14.45A the temp is at 74.7F. wow, this is very close to the free air test done of the one insulated wire in free air. remember, cooking NM is two CCC's, so thats 2x the heat at any given point.

i suspect the NM is itself acting as a good heatsink, there's much more surface area to dump the heat coming from the wires inside. this might lead me to make a statement about some of the other free air NM tests that have been documented, and that is, measuring the temp of the NM on the sheath may not be an accurate description of actual wire temps inside.

And the EGC, which is acting as a heatsink while providing spacing.

that said, having NM in a rigid foam sandwich should allow all of the components to reach same equilibrium temp, and as such, i am expecting both TC's to have same readings. this was of some debate back in this thread, where exactly do you measure temps in a multi-layered cable like NM? the most important area has to be the copper itself, from the OD of the copper outwards there has to be a temp gradient due to the R values of the things between, etc, the gradient is very small when in a foam sandwich = "NEC worse case scenario".

I think you have it down. One on the outside one on the copper. Keep both near the center.

 

[FionaZuppa]

the rate of heat dissipation is linear, the rate is directly proportional to delta-T

And the EGC, which is acting as a heatsink while providing spacing.
I think you have it down. One on the outside one on the copper. Keep both near the center.

the bare copper will come to equilibrium. i nipped off the egc where it exits the foam board, i will put a smear of silicone over that area once the sandwich is weighted, this will keep air movement to a minimum, the egc should come into equilibrium offering little in the way of heat sink ability. the R10 foam board is a major factor in the "heatsink", etc.

 

[GoldDigger]

the rate of heat dissipation is linear, the rate is directly proportional to delta-T

Which proves mathematically that as long as physical parameters such as thermal conductivity are independent of temperature the 10-for-10 rule applies
Note that the independence from temperature does not apply to radiative heat transfer. But that is not a significant part of the calculation for these cases and temperatures.

 

[FionaZuppa]

and just to note, since i channeled only one side of the foam board (about 7/32" deep) the R value one side is 10, the R value on the other side is ~8.91

the board has R value of 0.15625 per 1/32"


thermal conductivity (main mode convection) should be constant for the heatsink (thermal source, the wire) in open air.

 

[mbrooke]

the rate of heat dissipation is linear, the rate is directly proportional to delta-T


the bare copper will come to equilibrium. i nipped off the egc where it exits the foam board, i will put a smear of silicone over that area once the sandwich is weighted, this will keep air movement to a minimum, the egc should come into equilibrium offering little in the way of heat sink ability. the R10 foam board is a major factor in the "heatsink", etc.

Got it. I think...

 

[FionaZuppa]

hmmm, 1hr cooking the NM and i got a drop in temp. its very possible the metal TC is to blame as it was inside at ~74F prior to its use.

14.45A @ 74.7F @ 30min (~60F ambient)
14.40A @ 72.1F @ 1hr (~60F ambient)

if i normalize back to ~30C ambient the temps are ~100F, still well below 60C. and btw, 140F is more like "warm" vs "hot"

i up'd the amps to 25, will let it cook there for 1hr.

 

[mbrooke]

hmmm, 1hr cooking the NM and i got a drop in temp. its very possible the metal TC is to blame as it was inside at ~74F prior to its use.


14.40A @ 72.1F @ 1hr (~60F ambient)

i up'd the amps to 25, will let it cook there for 1hr.

I bet it wont even go to 55*C let alone 90*C.

 

[FionaZuppa]

14.45A @ 74.7F @ 30min (~60F ambient)
14.40A @ 72.1F @ 1hr (~60F ambient)
24.0A @ 91.4F @ 47min (~60F ambient)

24A, doesnt even feel warm. if i normalize that, its 121.4F (49.4C) @24A

i'll let the cooking continue @ 24A

 

[mbrooke]

14.45A @ 74.7F @ 30min (~60F ambient)
14.40A @ 72.1F @ 1hr (~60F ambient)
24.0A @ 91.4F @ 47min (~60F ambient)

24A, doesnt even feel warm. if i normalize that, its 121.4F (49.4C) @24A

i'll let the cooking continue @ 24A

I say proposal should go in the remove 240.4(D) for table 310.15 (B) 17.

 

[FionaZuppa]

I say proposal should go in the remove 240.4(D) for table 310.15 (B) 17.

and maybe allow NM under 75C column? so basically removing the two restrictions, cant really use wire under 90C because terminations are mostly, if not all, rated 75C.

but, i might also argue another hmmmmmmm question. should the 90C column (the whole table in general) be updated with more realistic #'s so that derating makes more sense?

 

[FionaZuppa]

14.45A @ 74.7F @ 30min (~60F ambient)
14.40A @ 72.1F @ 1hr (~60F ambient)
24.0A @ 91.4F @ 47min (~60F ambient)

23.75A @ 101.2F @ several hrs (~80F ambient, my garage changed due to outside warming up)
the wire simply feels warm (a light warm), less warm than my coffee cup in the AM just after it brews and makes it into my cup.

i'll let the cooking continue @ 30A

 

[FionaZuppa]

free air #14 Romex
14.45A @ 74.7F @ 30min (~60F ambient)
14.40A @ 72.1F @ 1hr (~60F ambient)
24.0A @ 91.4F @ 47min (~60F ambient)
23.75A @ 101.2F @ several hrs (~80F ambient)
29.1A @ 120.2F @ ~30min (~80F ambient)


at the ~30A mark the wire feels like my cup of hot coffee.

 

[wwhitney]

Nice work. A couple comments:

There's no point in measuring the wire temperature to 3 significant figures and measuring the ambient to only 1. Temperature rise is the relevant data, so you probably want to measure ambient more accurately.

I also wonder if 4 feet of insulation is the right order of magnitude to do a test like this, or if the heat conduction of the wire itself will allow a lot of the heat to be conducted to the free air portion of the wire and dissipate that way.

Cheers, Wayne

 

[mbrooke]

I also wonder if 4 feet of insulation is the right order of magnitude to do a test like this, or if the heat conduction of the wire itself will allow a lot of the heat to be conducted to the free air portion of the wire and dissipate that way.

Cheers, Wayne

I have the same concern. I would go 8 feet, the NEC actually lets you omit bundling if the wires a short length of the total run. I think is because they know heat will be conducted away.

 

[mbrooke]

and maybe allow NM under 75C column? so basically removing the two restrictions, cant really use wire under 90C because terminations are mostly, if not all, rated 75C.

For open air applications and NM run in walls with no insulation your test proves NM can be used at the 75*C column. For worse case scenario, we let the foam test speak.

but, i might also argue another hmmmmmmm question. should the 90C column (the whole table in general) be updated with more realistic #'s so that derating makes more sense?

That has been on my mind. But I think you would be further ahead of that then me

 

[FionaZuppa]

Nice work. A couple comments:

There's no point in measuring the wire temperature to 3 significant figures and measuring the ambient to only 1. Temperature rise is the relevant data, so you probably want to measure ambient more accurately.

I also wonder if 4 feet of insulation is the right order of magnitude to do a test like this, or if the heat conduction of the wire itself will allow a lot of the heat to be conducted to the free air portion of the wire and dissipate that way.

Cheers, Wayne

i am not a lab. ± a few degrees either way is not significant here.

4ft of insulation, what are you guys talking about? 4ft of NM? if you mean 4ft of wire in the foam board sandwich, the heatsink factor of the wire will also be insignificant, i am taking point temps in the middle of the wire. there's a temp gradient, and if you look at the heatsink side of the wire itself, the center of the wire will be T(max) and T(min) will be ambient that is 12" away. from a R perspective, the gradient is across the foam board @ approx R10 in the vertical (heat rises) and way more than that in the horizontal.

i can wrap the exposed NM to limit free air convection around the 12" section on each side of the foam board. is this what you are suggesting?

please provide more detail to help me understand what you mean.

 

[FionaZuppa]

For open air applications and NM run in walls with no insulation your test proves NM can be used at the 75*C column. For worse case scenario, we let the foam test speak.

i am not so sure it describes that. my 42" plasma TV in my bedroom warms the space at night, a 15x15x10'h room. a heat src in a non-infinite space can impact ambient in significant way.

20 BC's running down a non-glassed bay could cause ambient in that bay to rise quite a bit. this type of study needs to be done also.

 

[wwhitney]

i am not a lab. ± a few degrees either way is not significant here.

Oh I agree, it's just confusing to quote one number to the nearest 0.1 degree, and another number to the nearest 10 degrees, when the first thing you need to do with those numbers is subtract one from the other. So it would be typical to pick a level of accuracy, say the nearest degree, and record both the ambient and observed temperatures to that level of accuracy.

4ft of insulation, what are you guys talking about? 4ft of NM? if you mean 4ft of wire in the foam board sandwich, the heatsink factor of the wire will also be insignificant,

Yes, that is what we are talking about, whether with 4 feet of wire the heat sink factor will be insignificant.

You said the foam board is R10. If I looked up and converted from metric properly, the R value of copper per inch is 1/2700. So the R-value of 24 in of copper is about 1/100, i.e. the copper is on the order of 1000 to 10,000 times better thermal conductor than the foam.

#12 wire has a diameter of 0.081" or a cross-sectional area of 5.1 * 10^-3 in^2. 4 feet of the wire has a surface area of 12 in^2. So the ratio of surface area to cross-sectional area is 2,400, which is the same order as the ratio of thermal conductivities. At first glance, it appears heat sinking could be significant.

With a little more physics and a lot more math, one could model the distributed heat source accurately and solve the heat equations, but I'm not up for that.

If you are concerned, you could measure the effect: use the same configuration of wire in the foam test and the free air test, and measure the wire temperature at two locations during each test. One point would be in the middle of the foam when the foam is present, the other would be well in free air even when the foam is present. In the free air test, the two temperature rises should be the same (unless you have heat sink effects from your terminations). If the temperature rise of the free air portion during the foam test is greater than the temperature rise of the free air test, that tells you that the free air portion of the foam test is dissipating additional heat because it is heat sinking the section of wire in the foam.

Cheers, Wayne

 

[FionaZuppa]

here's a hypo, what happens when some university in the UK develops a new type of insulation for use in residential, one that is 2-3x better than glass, and IRbC adopts it as approved type. isnt this an issue for wires that are subject to coverage by building insulation? are we moving towards a rule that says all wires subject to coverage by insulation has to be in conduit?

you'll need to derive a proper R value for the internal wires up through the sheath to the ambient air. convection and conduction mode analysis would apply.

 

[GoldDigger]

A technique used in lab thermal conductivity measurements would be applicable here.
When measuring heat flow between two flat plates separated by insulation you have to deal with the edges of the sample and test plates. (The equipment holds a set temperature difference between the plates and measures heat flow.)
The technique known as guarding surrounds the active part of the slab with an additional hotvor cold plate section where the temperature is regulated but the heat transfer is not measured.
The result is that there is no temperature difference between the active plate section and the guard section and so no heat movement parallel to the plate to upset the measurement.
In this case the equivalent would be to actively heat the wire just outside the foam to a temperature just below the temperature measured at the center. This will completely eliminate heat sinking out of the test area and simulate an infinite length of wire running inside insulation.
I suspect that you will find the resulting temperature to be well beyond your wildest expectations with the rigid foam.