Back tie, also known as tieback, is a method for applying tension on an anchor in order to reduce or eliminate an undesirable leverage (moment) on that anchor. I was recently asked by Reed Thorne of Ropes that Rescue to test a particular back tie method to determine the tensile strength when using the 5.9mm PowerCord. After a brief discussion with Reed and other interested parties it was determined we would test a configuration like this:
The important feature is the tie off at the left carabiner. We wanted to make sure the tie off was snugged up against the carabiner and was around all of the strands.
After kicking this around with the other nerds in the engineering office we came to the conclusion we would see relatively low breaks. We figured there would not be good load equalization due to the tie off method. Tying off around all 3 strands, as is pictured, would cause the strands to bind and not move freely.
Moving the discussion over to the tensile tester I performed 5 pulls.
And here’s the results:
Notice the variation. This resulted in a relatively large standard deviation, which nets a very poor 3-Sigma value compared to the average of the 5 samples.
With the nerd collective hypothesis in mind, I decided to do 5 more samples. This time, however, I would only tie off a single strand instead of all 3.
Here are those results:
The important value here is the StdDev (standard deviation). In this series of breaks the standard deviation is only 2.7642 as compared to the 6.6176 of the previous series. Remember the 3-Sigma value is 3 times the standard deviation subtracted from the average. Click here to learn more about what 3-Sigma is and how to calculate, but the short story is that it is a statistical analysis method used commonly throughout the work at height industry.
Even though tying off all 3 strands gave us the single highest break of both series it also gave us the lowest. To me this indicates the 3 strand tie off method has unpredictable results. The 1 strand tie off method, with its significantly lower standard deviation, appears to be more consistent and predictable.