Fig. 1. Expansion curves for some common metals

Question

We are using a heating rate of approximately 25-30°F per minute in our atmosphere retort-type brazing furnace for low-carbon steel assemblies we are brazing. Each assembly consists of a fairly large tubular part that needs to be brazed along its entire length to another tubular part (same alloy) next to it with a somewhat smaller diameter. We projection weld a small clip onto the two tubes to keep them aligned with each other and include several load-thermocouples (load-TCs) in the load. We find that we have to include at least two built-in “holds” during the heating phase (about 20-minutes “hold” each time) in order to allow the load to equalize in temperature before we can continue heating once again. Even so, we still get a high percentage of the tubular brazements “pulling apart” somewhere during the cycle, i.e., the smaller-diameter tubing pulls away from the larger-diameter tubing, even snapping the welded clips off one of the tubes, so that they are not brazed together along their length. What’s happening, and how can we fix this problem?

This is really a two-part question. First, let’s look at the heating rates you are using. Why are you using 25-30°F/minute heating rate for these parts? Unfortunately, it is fairly common for companies to use a “standard” heating rate for assemblies in their brazing furnaces, and then build in the number of “holds” necessary along the way up to temp in order to keep all the load TCs within a certain allowed temperature spread. I challenge this thinking and ask you to do the same.

When a “standard heating rate” is used, it induces thermal stresses not only throughout the entire furnace load but also within each assembly, as thinner sections heat up faster, and thicker parts come up to temp much slower. These thermal stresses are powerful and can result in distortion and cracking (which we’ll look at next week). Using “standard heating rates” is often just a habit that gets carried over from one generation to another, so to speak.

Try to break the heating-rate paradigm by trying the following: use a much slower heating rate and also eliminate the built-in holds you’ve scheduled in along the way. Find out the heating rate that will allow all load-TCs to stay with the allowable tolerance band without the need for any built-in holds. A few test cycles may be needed to determine this, using dummy loads of the same mass you are currently using.

You may be very surprised to see that not only will the parts appreciate this greatly (from a stress point of view), but by eliminating the holds, your slower heating rate may actually result in a shortened brazing cycle overall.

Now Let’s Look at the Effect of the Mass of Each Part on Heating Rates and Distortion/Fracture of Parts

Growth of two different mass tubes of 1018 steel during furnace heating

Fig. 2. Growth of two different mass tubes of 1018 steel during furnace heating

When a “standard heating rate” is used, it induces thermal stresses not only throughout the entire furnace load but also within each assembly as thinner sections heat up faster, and thicker parts come up to temperature much slower. These thermal stresses are powerful, and can result in distortion and cracking of assemblies during a brazing cycle. Look at the graph in Fig. 1, which shows the expansion of the metals as they are being heated over time. Notice that as low-carbon steel (1018 steel on the graph) is heated, it reaches a temperature where it actually starts to shrink and then begins to expand again at a higher temperature. Under laboratory conditions, this takes place at one temperature, but under the rapid heating of production loads, this shrinkage can occur over a temperature range.

This so-called “hiccup” in the curve for 1018 steel occurs because this steel is an example of a polymorphic material, i.e. it changes its atomic orientation at certain temperatures to minimize internal stresses, and this “re-alignment” of atoms actually results in an atomic spacing that is smaller than the spacing of atoms at room temperature. Thus, the metal actually gets smaller upon heating through that temperature range. Once all the atoms have been rearranged, the metal can expand once again. The reverse happens upon cooling.

Fig. 2 shows the effects of this expansion/contraction on the assembly as it is being heated up.

Note that the expansion of the smaller-mass tubing (with its thinner wall, and slightly smaller diameter) occurs more rapidly than the larger diameter, heavier-mass tubing. During heating, the smaller tubing will begin to shrink (contract, as it transforms from a body-centered alignment of atoms to a face-centered alignment) while the heavier-walled tubing is still slowly expanding with heat. Soon the heavy-walled tubing has reached the temperature where it begins to shrink as well. The damage will already have been done during the region on the chart (Fig. 2) where the thinner tubing is shrinking while the heavier-walled tubing is still growing. The combination of a shrinking smaller tube along with an expanding heavier-walled tube can be enough to literally snap the weld of the clip that was trying to keep the two tubes in close alignment.

Conclusions

Slow down the heating rates of the furnace when the transformation temperature zone is being reached so that both the thin-walled and heavier-walled tubing will transform together at the same time. This will prevent the tearing stresses from occurring. Repeat this on the way down from temperature as well. By doing this, a lot of scrap and rework can be eliminated. Fig. 1. Expansion curves for some common metals.

Problem solve and improve with our technical experts.