ECN Forum Forums Code Related Discussion NEC & other Code issues Torque Specs

Note these comments refer to mechanical wire terminations and not buss bar connections or connections between a wire connector and a buss bar. The materials involved in these connections are harder and do not behave in the same way as wire.

All torqued connections relax over time. This relaxation is accounted for in the original torque spec.

If the connection is too tight, the wire has more "cold flow" and is extruded out of the connection. This is made worse with heavy cycling loads where the heat causes the wire to expand and produces even more cold flow. This will in some cases eventually result in a high resistance connection and failure.

In my opinion the same effect happens when wire connectors are retorqued as when they are over torqued.

Under torque also will result in a high resistance connection over time. With an under torqued connection, excess heat is produced at the connection and the wire expands causing cold flow and as this cycle repeats the connection get worse and at some point it will fail.

I was taught never to retorque a connection. If the connection is showing early signs of trouble, the wire should be removed from the connector, cut off, stripped and installed in a new connector with the correct amount of torque.

Don

Thanks Don, that was what I needed... Good ol' physics explained in laymens terms... Right up my alley! And thanks for clarifying about the busbar thing too... How do they differ? (If ya don't mind 'splainin' some more...)

Thanks again...
-Virgil

Wait... I just realized what you're saying...

The guys who over torque are finding their connections to be loose later because they are over torquing them in the first place! And my properly torqued lugs won't loosen as badly as theirs, and they think that mine have a "head start" on being loose...

Wow... An epiphany!

Think of it this way........ The threads are 'basically' a triangle observed from the side, they are designed (by what type of metal they're made from, and what their Rockwell hardness # is) for a certain stress.
Understress, it's not tight enough, in other words, it's already loose, and the problem ain't getting any better with the cycling Don mentioned.

Overstress, and the shape is malformed and no longer mates with the lug, leaving voids and eccentrics, THAT usually loosens quicker than an undertorque.

I've got some literature on WHY it is improper to retorque after proper torque has been reached, but I may have to wait until Monday, and I'll talk to our testers to see what they've got. This stuff is a lot more "official" than just putting it here, so if I can find it (I'm pretty sure I can) I'll pass it along.

One term for an overtorquing effect is "coldflow;" or the tendency for the conductor to "ooze" over time. Copper is not exempt from this characteristic, but it less sensitive to it.

For aluminum, it's important to call for "AA-8000" conductor alloy, [99NEC§310-14] which is compounded with a small amount of iron in the recipe. Be aware that not all aluminum insulated conductors are automatically 8000 series. Type USE cable is permitted to still be formulated of “EC-1350” alloy, as may also be utility-furnished 600V URD cables.

IMO and experience, nonreversible hydraulic-compressed terminals are the only sure bet. They just take a little more planning.

Out of a complete fluke, I was inadvertantly delivered two in-lbs torque wrenches, Sears Craftsman. I was charged for both and the local Sears cannot take it back, I have to send it back to Chicago via $$. They'd probably give me a refund for the postage too, but I figured I'd print this thread out and try to sell the extra one to my partner, Joe Musser.

We'll see...

Fate?

Interestingly, I called Leviton and their Rep told me that all of their standard devices with screw terminals are listed at 12 in-lbs...

I said: "Man that's light!" And he said, "Yes sir, 12 in-lbs."

(Oh, and Joe kept the second torque wrench without any prodding at all!)

the 'machinist's handbook has some interesting stuff on threads thier applicability,
apparently not all are created equal....

i don't know how all this fits in with those new seimens speed-screw terminations .....?

From an engineering standpoint torque is a poor measurement.

If we wrap a wire 360 degrees around a screw and tighten to a specified torque, we have 1/2 of the load (#/sqin) on the wire then if we wrap the wire 180 degrees around the screw and tighten to the same torque. Thus we reach the elastic limit (cold flow) at different torques based on the uniformity of our wrapping.

In the case of plate clamps (backwired devices, large breakers) and ground/neutral bars there is more uniformity in the clamping area and there the torque is more reliable.

-----

Don't confuse the torque numbers used in testing with torque specifications for instalation.

Have fun with your toys.

If torque is not specified, UL has defined a "default" torque for various fastener sizes. Maybe someone can come up with a table?