High current & warm wires

[G._S._Ohm]

110305-1511 EST

G._S._Ohm:

One way to measure temperature rise of wire is to use the change of resistance of copper with respect to temperature. This technique is also useful to measure average internal temperature rise in a transformer.

For your #14 Romex experiment a 250 ft length will provide enough resistance to allow moderately accurate resistance measurements. Use both wires in the Romex to generate heat, but use only one for the resistance measurement. Make the one wire a 4 terminal resistor to get accurate resistance measurements. #14 copper is about 2.525 ohms per 1000 ft. So 250 ft is about 0.631 ohms. For annealed copper the coefficient at 20 deg C is 0.00393 and 0.00382 for hard drawn copper. For a 10% change in resistance the temperature rise is about 0.1/0.00393 = 25.4 deg C. This is a wire temperature of 114 deg F. for room ambient of 20 deg C.

I seriously doubt that you only had a 2 deg C wire temperature rise in your experiment.

.

My way fed the wire with a more or less constant current.

Try it. I can always use experimental corroboration - or lack of it. I used a precision thermistor to measure the temp. under the jacket.

Another way is to use one of the Weller soldering guns with the two posts. They put out about 150 A into the tip @ 1 or 2 vac so with a few feet of Romex you can drop the current down to 30 A or so.

The next thing I want to do is put a few amps of DC into cable armor. Along with the AC impedance this should give the inductance per foot of this coiled shield.

 

[gar]

110306-1513 EST

G._S._Ohm:

You need to run the experiment I suggested to see how far off your result is.

I ran a quick experiment as follows:
A 250 ft roll of #14 copper Romex. The precise length is of no importance. I did not uncoil the roll so the rise will be higher than if the Romex was run between joists, and not bundled.
A Fluke 27 for the voltage drop meter.
A Beckman 4410 used to measure the voltage across a 10 A 50 mV Simpson shunt.
The load was a nominal 1500 W cube heater.
Room temperature was 68 to 70 deg F.
Line voltage around 123 V and moderately stable, within +/- 2 V.

Results:
0.636 ohms at T = 0 -------- 6.74 V and 52.95 mV, 0% change
0.653 ohms at T = 5 min ---- 6.91 V and 52.93 mV, 2.7% change
0.688 ohms at T = 13 min -- 7.24 V and 52.64 mV, 8.2% change
0.709 ohms at T = 20 min -- 7.44 V and 52.45 mV, 11.5% change

Looks close to a linear rise in this time frame. The experiment really needs to be done with a straight piece, and 25 feet might be adequate. This should eliminate end effects, and still provide sufficient voltage drop for useful measurements. But if the Romex is not in free air, but is in conduit, bundled, or other insulation is around it, then the test setup needs to be different.

Your thermistor measurement is not close enough to the actual wire. The change of resistance of the wire is a much more useful technique.

.

 

[G._S._Ohm]

110306-1513 EST

G._S._Ohm:

You need to run the experiment I suggested to see how far off your result is.

I ran a quick experiment as follows:
A 250 ft roll of #14 copper Romex. The precise length is of no importance. I did not uncoil the roll so the rise will be higher than if the Romex was run between joists, and not bundled.
A Fluke 27 for the voltage drop meter.
A Beckman 4410 used to measure the voltage across a 10 A 50 mV Simpson shunt.
The load was a nominal 1500 W cube heater.
Room temperature was 68 to 70 deg F.
Line voltage around 123 V and moderately stable, within +/- 2 V.

Results:
0.636 ohms at T = 0 -------- 6.74 V and 52.95 mV, 0% change
0.653 ohms at T = 5 min ---- 6.91 V and 52.93 mV, 2.7% change
0.688 ohms at T = 13 min -- 7.24 V and 52.64 mV, 8.2% change
0.709 ohms at T = 20 min -- 7.44 V and 52.45 mV, 11.5% change

Looks close to a linear rise in this time frame. The experiment really needs to be done with a straight piece, and 25 feet might be adequate. This should eliminate end effects, and still provide sufficient voltage drop for useful measurements. But if the Romex is not in free air, but is in conduit, bundled, or other insulation is around it, then the test setup needs to be different.

Your thermistor measurement is not close enough to the actual wire. The change of resistance of the wire is a much more useful technique.

.

Your results probably mean I screwed up way back when I measured it.

And, your temp rise when plotted is linear even after 20 minutes rather than leveling off like the [1-{e^(-T)}] curve would predict, which means the wire temp would have stabilized at an even higher temp. My temp seemed to level off after 15 minutes but I only had a 1' length so it had less thermal inertia. Wire in place is not coiled so I'd try it with a straight piece. Even better, for worst case, use a straight piece embedded in a few inches of insulation.

I'll bookmark this thread and use your results to see if I can reverse engineer some of table 310-16. Thanks for your help.

 

[gar]

110307-1006 EST

G._S._Ohm:

Your thermistor technique does not get close enough to the copper wire itself to give a good measure. The 1 foot length also may have had a noticeable end effect, but probably not too great.

I would have tried a 25 ft straight piece but I did not want to cut my roll. My temperature is definitely too high because of the coil and my guess would be it probably levels off at about a 20% to 25% change. Outside surface was warm to the touch. Means maybe 105 deg F.

.

 

[G._S._Ohm]

I used Excel to check whether your 4 readings correspond to the exponential equation.
Within a few percent it does but it's not as close as I'd hoped. Possibly it takes two exponentials for a better fit.

 

[gar]

110308-1742 EST

G._S._ Ohm:

My measurements are not accurate enough other than to illustrate that there is substantial temperature rise. I did not want to spend the time to plot rise as it approaches equilibrium.

.