Fig. 1 – Typical furnace chart showing two (2) built-in temperature-stabilization holds, one at about 1250°F (675°C) for 45-minutes, and another at about 1750F (950C) for 30-minutes, with a hold at brazing temperature for an additional 30-minutes at about 1975°F (1080°C). Brazing is with AMS 4777/BNi-2 (solidus temp = 1780°F/970°C, liquidus = 1830°F/1000°C).
For a number of years I have been encouraging people to re-think the ramp-rates they use for their vacuum brazing cycles. Many brazing shops using rather high ramp-rates during heating claim that “this is the way we’ve always done it”. Perhaps it’s time to re-think this. From a metallurgical point of view, too-rapid a heating rate can lead to stresses and strains in the metal assemblies being heated, which can often lead to distortion of parts during their heat-up, and can even lead to parts-failure (I’ve seen this too many times).
I have recommended to a number of brazing shops that they slow down their heating ramp rates (and I’ve seen excellent results), using the following guideline:
Heat the parts at the fastest rate that will allow you to bring all the parts (assemblies) up to brazing temperature without the need for any holds (for temperature-equalization) on the way up.
Let’s look at an example of how this might be done. Figure 1 represents a temperature chart for a typical vacuum brazing cycle, such as might be used when brazing at a temperature of approximately 1975°F/1080°C with a high-temperature nickel-based brazing filler metal (BFM) like AMS 4777 (AWS A5.8, Class BNi-2). In the absence of any information to the contrary, I always recommend brazing at a temperature approximately 100F/50C above the published liquidus of the BFM being used. This will insure that the BFM will indeed become fully molten, and will have enough extra thermal energy to flow well and spread out by capillary action.
The chart shows two (2) temperature-stabilization holds built into that cycle, and represents the temperature being measured by the main furnace thermocouple (TC) which extends slightly into the top of the furnace hot-zone.
Total cycle time in the example shown in Fig. 1 is about 3.5 hours, more than a third of which (36%) is merely the time spent in two (2) temperature-stabilization holds during the heating portion of the furnace cycle. To hopefully eliminate those two holds, a straight line is drawn on the chart from the start of the heating cycle up to the end of the “time-at-brazing-temperature” line, to determine what heating-rate would hopefully get you up to brazing temp without the need for any stabilization holds. In this case, that straight line represents a heating-rate of approximately 12°F (6°C) per minute, and connects with the end-point of the “time-at-temp” brazing line, i.e., at a point where cooling normally begins, since the slower heating-rate of only 12°F/6°C per minute all the way up to brazing temp should be able to eliminate the need for any “hold at brazing temp”. Here’s why:
Time at Brazing Temperature
I am concerned primarily with the total time the brazed-assemblies are above the solidus-temp of the brazing filler metal (BFM), because the BFM will start to melt and flow by capillary-action as soon as the BFM is heated to a temp above its solidus temperature.
Important: It is important that this temp be measured within the coldest section of the coldest assembly being brazed in the load. It should be observed that the chart in Fig. 1 does NOT show any secondary temperature curves, i.e., there are no load-TC measurements shown on the chart. This was done intentionally, so that the new recommended furnace-heating ramp-rate with no holds could clearly be seen in comparison to the previously used rapid heating rate that have the built-in holds. Normally there would be additional temperature lines on the chart representing the temperatures recorded by other TC’s placed throughout the furnace load, but they have been excluded from the chart for clarity. Obviously, placing numerous TC’s throughout the load itself would tell the brazing shop personnel what is happening throughout the actual load, since that cannot be determined merely by looking at the furnace-TC readings. I always strongly recommend the use of at least three (3) load TC’s or more in every braze load.
Assuming proper thermocoupling of the furnace (including the load), the TOTAL amount of time above the solidus temp of the BFM, i.e., from the time the TC within the coldest part crosses that solidus on the way up to brazing temp, until it again crosses that solidus temp on the way down from brazing temp, should represent the “total capillary flow time for the BFM”. And, considering the slow heating ramp-rate being recommended in this example, the total time above the BFM’s solidus temp may actually be quite long. Therefore, as soon as the load-TC in the coldest assembly in the furnace-load reaches the brazing set-point temp (1975F/1080C), slow cooling can immediately commence (no extended hold is necessary) — unless there are particular metallurgical reasons for extending the hold (such as special boron diffusion requirement for isothermal solidification, etc.).
Important Reminder: Therefore, to effectively accomplish what I am recommending in this article, i.e., to develop an effective brazing cycle that can eliminate the need for stabilization holds on the way up to brazing temperature, it is essential that TC’s be well placed throughout the furnace load during the development/test phases of this work. TC’s must be connected to the thinnest (most easily heated) section of ONE part (assembly) as well as within the thickest (slowest heating) section of the same assembly, so that the temperature-differential (i.e., delta-T, or Δ-T) within that ONE assembly can be determined, because distortion of parts is the result of the Δ-T within a single component, not just the Δ-T of the total load (when many assemblies are being brazed at the same time). Thus, for proper brazing cycles, TC’s should be throughout the entire furnace, in the hottest and coolest section of the total load, as well as having at least one of the assemblies in the coolest part of the load thermocoupled to show the temps of the hottest and coldest portions of that part/assembly.