NEC Changes For #14 Ampacity

[FionaZuppa]

I used a slightly different value of resistance of copper there. The fact we are not correcting the copper resistance for temperature is a source of error in our calculations; your value is slightly higher than mine and so probably closer to reality.

Cheers, Wayne

so, yeah, i typed 0.00259 when i should have used 0.00253, not a issue, but it is 0.00253/ft

.....and correction factor, i do correct for it in actual measurement, via the more significant term, I^2, i take amps reading at t=0 and then t>>>0 (when temp gets to equilibrium) and the amps later is always a tad less than t=0. the math is real close too, thus it seems to be insignificant.

the dominating factor is the external thermal insulation.

 

[FionaZuppa]

so, after doing some more equation hunting i found that the heatsink affect seems to die off expoentially with distance, leading me to say the ends are insignificnat. however, i have no test data to back that, so here's a test (just one).

wire is 42" long, #14, insulated with THHN. 10" sticks out each end. the center section has a woven fiberglass sheath (i use this stuff around wires on another application i have to protect the wires from ~800F). the ends of the sheath are wrapped with some tape to close them off from air movement.

the TC is inside the sheath at center. cooking at 30A. will take temp at equilibrium, and then i will add ice water to the containers where the clamps are, lets see if cooling the ends has any affect on TC temp.

 

[FionaZuppa]

ok, absolutely weird.
i added 8 ice cubes and water to each container, clamps are both in about same depth of water.

here's the odd part:
1) amps went up, ok, makes sense, less ohms right on the clamped connection, not much (few tenths), so i adjusted amps back to pre-ice amps, etc.
2). within just ~1-2min the temp on TC went up, not down.

ambient has not changed at all.

i am letting it cook for another ~60min to see where the TC settles at.

 

[mbrooke]

ok, absolutely weird.
i added 8 ice cubes and water to each container, clamps are both in about same depth of water.

here's the odd part:
1) amps went up, ok, makes sense, less ohms right on the clamped connection, not much (few tenths), so i adjusted amps back to pre-ice amps, etc.
2). within just ~1-2min the temp on TC went up, not down.

ambient has not changed at all.

i am letting it cook for another ~60min to see where the TC settles at.

After you adjusted the amps, did the current go up afterwards? Was there water around the ice cubes? Ice take much more BTUs per unit to undergo phase change then it takes water or ice to go up or down 1 degree.

 

[FionaZuppa]

After you adjusted the amps, did the current go up afterwards? Was there water around the ice cubes? Ice take much more BTUs per unit to undergo phase change then it takes water or ice to go up or down 1 degree.

i wasnt measuring temp of the ice water. if the water/ice bath is pulling joules out of the copper faster than before then we expect the TC temp to go down.

i adjusted amps after i noticed temp went up. the few tenths of amps is not enough to account for the +4F i saw on TC. have to let it cook for some more time, then i get TC temp again. migration of heat along a conductor is cumulative.

 

[FionaZuppa]

ok, so there must be something else going on.

here's the sequence

• no ice water - 27.8A @102.7F
• add ice water - 28.23A - adjust amps back to ~27.8 , was able to get it to 27.92
• let cook into equilibrium
• measure - 27.92A @105.4F

this seems odd to me. i made a small hole in the fiberglass sheath right in the center for TC probe, the TC probe length is ~2.5"

i have now removed the ice water and will let the clamps come back into ambient equilibrium, lets see where the TC temp goes.

my hypothesis is this. the clamps allow heat to move axially. this movement has to be cumulative. the rate is governed by T1-T0, then i made T0 much colder which means the rate axially is now greater. does this account for temp rise for T1? dunno, i need some math to back my observations. the rise in T1 would be ~2.5" away from center, the TC probe has the end of the TC junction at its tip.

so, as of now, a tad confused by the observations.

 

[FionaZuppa]

so maybe the initial TC temp was not correct.
with the ice water removed and cooking there for some time until a see temp not move, it settled in at 107.5F, thats +2.1F. this indicates a 2% diff, but duly noted, a 2% diff when the clamps were in ice water, real test everything is at ambient.

i am ok with adding 1% to measured temps of the sandwich setup i have due to the clamps at the ends of the wire.

but still, for the purposes of this study, 1% is insignificant in my book. this seems to coincide with my post #540.

notice what i underlined above. there are some other unaccounted-for things, which are not significant, so i am ok in adding another 1% for these other heat losses taking it to +2% of measured temps. that's just +3F on 150F measured. not a diff to worry about for the purpose of this study.

 

[FionaZuppa]

ok, Omega TC connectors showed up today. tomorrow i start the sandwich test. instead of vice-grip clamps i will use #8 machine screw and a couple of small fender washers to clamp the big lugs to the smaller #14 eyelets, thus reducing some of that heatsink affect.

i think increments of 5A starting at 10A is a good stepping #. should be well into equilibrium at the 1hr mark for each step. a 6hr test taking the wire up to 30A.

the sandwich will be laying flat on a piece of 3/4ply which is on top of a plastic rolling cart.

anyone have any last min concerns before this test tomorrow?

if the wire survives w/o damage (which i suspect it will) i will turn the sandwich up on edge @30A to see if the wire temps change at all.

 

[mbrooke]

ok, Omega TC connectors showed up today. tomorrow i start the sandwich test. instead of vice-grip clamps i will use #8 machine screw and a couple of small fender washers to clamp the big lugs to the smaller #14 eyelets, thus reducing some of that heatsink affect.

i think increments of 5A starting at 10A is a good stepping #. should be well into equilibrium at the 1hr mark for each step. a 6hr test taking the wire up to 30A.

the sandwich will be laying flat on a piece of 3/4ply which is on top of a plastic rolling cart.

anyone have any last min concerns before this test tomorrow?

if the wire survives w/o damage (which i suspect it will) i will turn the sandwich up on edge @30A to see if the wire temps change at all.

Sounds good, but I would try 25 amps for 6 hours since this is the maximum real world load #14 could see on a 20amp OCPD (125%x20=25).

 

[FionaZuppa]

why 6hrs at a single amp #?
i expect temps to be in equilibrium for more than 10min during each 1hr soak. 25A is in the steps so we can get temps there, i would like to go past, perhaps even to 35A so we can see what the graph of temp vs amps looks like.

what does going ~+500% in equilibrium time get us in terms of data?

 

[wwhitney]

with the ice water removed and cooking there for some time until a see temp not move, it settled in at 107.5F, thats +2.1F.

So what was ambient? If ambient was 70F, then changing the ambient on the clamps from 70F to 32F changed your heat rise from 37.5 to 35.4. That's a 5.6% decrease.

If the heat sinking effect is linear in temperature difference, which we think it is, then changing the clamps from 70F to 108F would likewise change the the heat rise 2.1 degree F the other way, i.e. to 39.6 degrees F, a 5.6% increase. So it looks like the heat sinking effect for the conditions of this test is between 5% and 6%.

BTW, do you have an R-value for the woven fiberglass sheath?

this indicates a 2% diff, but duly noted, a 2% diff when the clamps were in ice water, real test everything is at ambient.

It's not particularly useful to take a temperature difference (the 2.1F) as a percentage of a particular temperature on a Fahrenheit scale. That temperature scale is not an absolute scale, the choice of 0F was an arbitrary one.

Cheers, Wayne

 

[romex jockey]

anyone have any last min concerns before this test tomorrow?

I'd say you've considered every possible variable FZ

bravo!

~RJ~

 

[FionaZuppa]

wayne,
i am not following your analysis.
with the ice water the heatsink is really just a pool of cold water. the clamps exposed to air are no longer exothermic but rather endothermic as the heat from ambient air is now using the clamps to put heat into the cold water. the exothermic copper wire now simply has a conductive path directly to the cold water, and with fluids the main mode is typically convection.

w/ ice water 105.4F
w/o ice water 107.5F
that's a diff of 2.1F, which is ~2% more than 105.4F

but, lets work in K
105.4F = 313.93K
107.5F = 315.09K
1.16K diff = 0.37% diff
thats much less than the 2% factor i said earlier. with the temp scale change and new way to fix the ends, i am ok with the following adjustments

[(0.5% for clamped ends + 0.5% for other losses)*measured(^oK)]+measured(^oK)
so basically +1%^oK, that would be +5.6F on a measured 100F. i am ok with this.

 

[FionaZuppa]

So what was ambient?

BTW, do you have an R-value for the woven fiberglass sheath?

ambient was ~55F
R value, dunno, but it was same as this http://www.delcity.net/store/Tru!Fit-Fiberglass,-High-Temp-Wire-Loom/p_806824
(i have my product specs somewhere, just cant find them now)

 

[mbrooke]

why 6hrs at a single amp #?
i expect temps to be in equilibrium for more than 10min during each 1hr soak. 25A is in the steps so we can get temps there, i would like to go past, perhaps even to 35A so we can see what the graph of temp vs amps looks like.

what does going ~+500% in equilibrium time get us in terms of data?

I thought you were doing this at 30amp, so I was thinking about using a real world value. Though in any case I like your idea of going higher.

As for the temp rise you saw earlier Im still trying to wrap my mind around it, however I think Wayne might be correct. (I think)

 

[mbrooke]

So what was ambient? If ambient was 70F, then changing the ambient on the clamps from 70F to 32F changed your heat rise from 37.5 to 35.4. That's a 5.6% decrease.

If the heat sinking effect is linear in temperature difference, which we think it is, then changing the clamps from 70F to 108F would likewise change the the heat rise 2.1 degree F the other way, i.e. to 39.6 degrees F, a 5.6% increase. So it looks like the heat sinking effect for the conditions of this test is between 5% and 6%.

BTW, do you have an R-value for the woven fiberglass sheath?


It's not particularly useful to take a temperature difference (the 2.1F) as a percentage of a particular temperature on a Fahrenheit scale. That temperature scale is not an absolute scale, the choice of 0F was an arbitrary one.

Cheers, Wayne

But doesnt water have a much higher thermal conductivity? If you look at a water to refrigerant heat exchanger it is much smaller then an air to refrigerant heat exchanger of the same btu.

 

[FionaZuppa]

But doesnt water have a much higher thermal conductivity? If you look at a water to refrigerant heat exchanger it is much smaller then an air to refrigerant heat exchanger of the same btu.

air 0.024 W/m-K
Cu 401 W/m-K
water 0.58 W/m-K

so, i guess it depends on what you are comparing to. yes, much better than air, and no, much worse than copper.

the ambient really doesnt matter with the ends sitting in ice water because the main mode of the "heatsink" at this point is conductive Cu to convective Water. the whole time the heat from ambient air was entering the ice water via the vice-grips and water-to-air interfaces.

 

[mbrooke]

air 0.024 W/m-K
Cu 401 W/m-K
water 0.58 W/m-K

so, i guess it depends on what you are comparing to. yes, much better than air, and no, much worse than copper.

the ambient really doesnt matter with the ends sitting in ice water because the main mode of the "heatsink" at this point is conductive Cu to convective Water. the whole time the heat from ambient air was entering the ice water via the vice-grips and water-to-air interfaces.

IMO, I think it shows the heat sink does play a role. Personally I would go with as you mentioned screws rather then clamps and maybe another few feet of foam board.

 

[wwhitney]

OK, my initial analysis for the clamps in ice water was flawed because I was comparing convection through the air versus conduction through the water. But here is what I am trying to say:

First, the output data of your experiment is the temperature rise. Everything should be expressed in terms of change in temperature. So if you compare a change in temperature of 2 degrees F to a change of 40F, say, that's a 5% effect. That computation is independent of the temperature scale.

Second, as Golddigger suggested a long time ago, what you would like to measure is the temperature rise at the middle of your conductor when there is no axial heat flow. He suggested the typical way to do that is to use heaters on the boundary of the test area to keep the boundary at the same temperature as the test probe, so that there will be no temperature difference between the two, and no heat flow.

That sounds complicated, so I'm suggesting that with your clever idea of using ice water you can instead fix the boundary temperature at two different values and measure the respective temperatures at the test probe. Then you can extrapolate what the temperature at the test probe would be if there were no axial heat flow to get a heat-sink corrected temperature rise.

For example, suppose ambient is a constant 62F and you place the clamps in a large body of water at ambient. Hopefully large enough that its temperature will remain constant during the experiment. Then for a particular set of test conditions you measure a conductor temperature of 90F. Now you repeat the experiment with the clamps in ice water at 32F. The temperature at the conductor becomes 88F. So lowering the boundary temperature 30F changed the conductor temperature 2F.

From this we can extrapolate that raising the boundary temperature from ambient to 92F would raise the conductor temperature by 2F from 90F to 92F. Since the conductor temperatures and boundary temperatures would then be the same, we conclude there would be no axial heat flow in this situation. So we use 92F to calculate the heat-sink corrected temperature rise.

I picked the numbers so the math would be easy, but in general it's just a little algebra for the extrapolation.

Cheers, Wayne

 

[FionaZuppa]

IMO, I think it shows the heat sink does play a role. Personally I would go with as you mentioned screws rather then clamps and maybe another few feet of foam board.

the foam board setup has already been completed. will collect the measurements and we can pick at the results after.

OK, my initial analysis for the clamps in ice water was flawed because I was comparing convection through the air versus conduction through the water. But here is what I am trying to say:

First, the output data of your experiment is the temperature rise. Everything should be expressed in terms of change in temperature. So if you compare a change in temperature of 2 degrees F to a change of 40F, say, that's a 5% effect. That computation is independent of the temperature scale.

Second, as Golddigger suggested a long time ago, what you would like to measure is the temperature rise at the middle of your conductor when there is no axial heat flow. He suggested the typical way to do that is to use heaters on the boundary of the test area to keep the boundary at the same temperature as the test probe, so that there will be no temperature difference between the two, and no heat flow.

That sounds complicated, so I'm suggesting that with your clever idea of using ice water you can instead fix the boundary temperature at two different values and measure the respective temperatures at the test probe. Then you can extrapolate what the temperature at the test probe would be if there were no axial heat flow to get a heat-sink corrected temperature rise.

For example, suppose ambient is a constant 62F and you place the clamps in a large body of water at ambient. Hopefully large enough that its temperature will remain constant during the experiment. Then for a particular set of test conditions you measure a conductor temperature of 90F. Now you repeat the experiment with the clamps in ice water at 32F. The temperature at the conductor becomes 88F. So lowering the boundary temperature 30F changed the conductor temperature 2F.

From this we can extrapolate that raising the boundary temperature from ambient to 92F would raise the conductor temperature by 2F from 90F to 92F. Since the conductor temperatures and boundary temperatures would then be the same, we conclude there would be no axial heat flow in this situation. So we use 92F to calculate the heat-sink corrected temperature rise.

I picked the numbers so the math would be easy, but in general it's just a little algebra for the extrapolation.

Cheers, Wayne

well, i dont think its that simple. if we look at just the clamps in water for -30F and +30F water temp around ambient, then perhaps we see same magnitude for increase and decrease in wire temp. but, thats not really what we were looking for, we wanted to see if the heatsink mode was significant compared to the radial heat loss path. but for the most part, with the main mode being conductive/convective into ice water, ambient doesnt play a significant role there as the ice water is always at ~32F.