In everyday practice we can encounter problems connected with brazing imperfections – often manifested as leaking. In order to identify the failure root causes, usually a detailed examination needs to be perform. Optical metallography is one of the most important tools for such an analysis.
This section describes the necessary steps and control procedures to ensure a properly brazed joint.
1. Clean the Components
The joint area must be cleaned free of cutting and machining lubricants. Aqueous cleaning, solvent dipping or wiping are acceptable procedures.
2. Assemble the Components
The components are assembled with the filler alloy ring in place. There must be intimate contact between the two components to be joined and the alloy ring.
3. Apply the Flux
The flux is then applied with a small brush around the circumference of the joint at a loading of about 25 to 30 g/m2.
4. Dry the Flux
The flux should be allowed to dry before the application of intense heat to begin brazing. This can be done by allowing the joint to air dry or alternatively by gently heating the surrounding joint area with the flame, which will heat the metal and dry the flux. Intense heat should be avoided before the flux has dried, otherwise splattering and flux fall-off will occur.
Once the flux has dried, more intense heat to begin the actual braze sequence can be applied. The braze flame should not be allowed to impinge on any one area very long to avoid overheating and burnthrough. The component with the higher thermal mass should be heated more. The flame should not be allowed to rest on the flux or preform ring to avoid premature melting before the joint area is uniformly heated to braze temperature. The flame should be kept moving at all times, moving back and forth between the components of different mass in such a way as to bring the entire joint to temperature uniformly.
There are three temperature indications in NOCOLOK ® flux flame brazing. The first is the appearance of a yellow flame at the Al surface. This indicates that the surface is starting to overheat/burn, since the aluminum skin always runs hotter than the component center.
The flame must visit the area less frequently to avoid burning. The second indicator is the first sign of flux melting, that is the fluxed area turns from white to clear. This indicates that the joint temperature is about 565 °C.
At this point the flames can be played directly on the joint and filler metal ring. Very shortly after flux melting, the filler metal ring begins to loose shape (third temperature indicator) and begins to melt at 577 °C. The molten filler metal is quickly drawn into the joint by capillary action. As soon as the full preform ring is molten, the flame should be quickly removed and the brazed joint allowed to cool.
6. Post Braze Treatment
After cooling, no further treatment is required. The flux, although visible is non-hygroscopic and in standard applications non-corrosive. With the brazing conditions optimized, meaning minimal flux residue, the surfaces can be painted with relatively good paint adhesion over the flux residue.
If absolutely desired, the flux residue can be removed, but only by mechanical means such as wire brushing and grit blasting. Removing the flux residue is recommended only when joint cleanliness is absolutely imperative.
Recommended equipment for flame brazing
Since the principles of flame brazing can be explained using the most basic equipment, only the equipment necessary for manual flame brazing is described. From the basic principles, all other equipment is only a matter of the degree of automation the end user wishes to achieve.
It is critical that the joint area is brought up to temperature uniformly. For this reason a dual headed torch capable of heating the joint from 2 sides is recommended.
A multi-orifice tip generates a broader flame at the exit of the tip. This feature enhances component temperature uniformity during heat up. Pin-point flames should be avoided as burn-through can easily occur.
Most commercial gas mixtures are acceptable for flame brazing Al:
- oxygen – propane
- oxygen – methane
- oxygen – natural gas
- oxygen – acetylene (oxyacetylene)
- Oxyacetylene combination produces the hottest flame and may be used, but with extreme care to avoid overheating and burn-through.
The filler metal alloy most commonly used for flame brazing Al is AA4047 which contains 11 to 13 % Si. The Al-Si phase diagram shows the eutectic at 577 °C with 12.6 % Si. AA4047 filler alloy therefore has the lowest melting temperature with the highest fluidity, ideal properties for flame brazing Al.
The filler metal is available in a variety of shapes and forms including wire, rings, foil and powder. When used as a powder, it is usually mixed with flux and a carrier to form a paste (more on pastes below). The filler metal wire is also available commercially either cored or coated with flux, precluding the application of flux.
Commercially available brazing pastes consist of the flux, powdered filler metal and a binder/carrier to keep everything in uniform suspension. This paste is all inclusive, there is no need to supply flux or filler metal to the joint separately. Brazing pastes can also be applied with automatic dispensers, with syringes or by brush application.
This is very similar to brazing pastes except that there is no powdered filler metal present, meaning that flux pastes requires filler metal in one form or another to be added to the joint separately. The advantage of using a flux paste is that the end user does not have to prepare his own paste.
In-House Paste Preparation
The least expensive and most common is the in-house preparation of flux pastes. The flux is mixed with either water or alcohol and/or water at 40 % to 60 % solids. Using some alcohol in the paste formulation allows for quicker drying. Using pastes prepared in-house of course requires that the filler metal be supplied to the joint separately.
These pastes are not easily dispensable automatically and are most often applied with a brush.
For brazing a tube-to-tube joint, the table below summarizes the complexity level in applying the flux and filler metal in their various forms:
|Flux Cored or Coated Wire||Brazing Pastes||Flux Pastes||In-House Paste Preparation|
|1. Preplace ring at the joint||1. Apply or dispense paste
at the joint
|1. Preplace ring at the joint||1. Prepare paste|
|2. Braze||2. Dry||2. Apply or dispense paste
at the joint
|2. Preplace ring at the joint|
|3. Braze||3. Dry||3. Apply paste at the joint|
|4. Braze||4. Dry|
Flame brazing of aluminum is not new. In fact the very first brazed aluminum assemblies were produced using a chloride based flux and a flame as the heat source. What has changed over the years is the sophistication of the types of fluxes available and to a certain extent the alloy selection.
However, even if one returns to the absolute basics of a flame, filler metal and flux, there remains a great deal to be learned about the fundamentals of flame brazing of aluminum. This becomes especially evident when the brazing engineer applies his techniques and equipment to NOCOLOK ® Flux flame brazing and years of learned practice seem to fail. This is largely due to the fact that the years of acquired knowledge of flame brazing aluminum has come from corrosive chloride-based flux brazing. Unfortunately, the same techniques can not be directly applied to NOCOLOK ® Flux flame brazing. It is therefore the intention of this article to re-familiarize the brazing engineer with the fundamentals of flame brazing aluminum and use those fundamentals to realize all the advantages of NOCOLOK ® Flux brazing.
What is Flame Brazing?
According to the American Welding Society, brazing is the joining of metals using a molten filler metal, which on cooling forms a joint. The filler metal melting temperature is above 450 °C, but below the melting point of the metals.
Flame brazing then implies the use of a flame as the heat source to accomplish what is described above.
Flame brazing lends itself well to joining components with simple configurations such as tube-to-tube, tube-to-fitting and joints having large thermal mass differences. Since much faster heating rates are possible than in furnace brazing, flame brazing is versatile and as will be explained in more detail later, can braze some Mg containing alloys.
What is NOCOLOK ® Flux?
NOCOLOK ® flux is a white powder consisting of a mixture of potassium fluroaluminate salts of the general formula K1-3 Al‑F4-6. The flux has a defined melting point range of 565 °C to 572 °C, below the melting point of the Al-Si brazing alloy. The flux is non-corrosive and non-hygroscopic and is only very slightly soluble in water (0.2 % to 0.4 %). The shelf and pot life of the flux is therefore indefinite. The flux does not react with Al at room temperature or at brazing temperature and only becomes reactive when molten.
Role of the Flux
Once molten the flux works by dissolving the oxide film on the Al surfaces to be joined and prevents further oxidation. The flux wets the Al surfaces and allows the filler metal to flow freely into the joints by capillary action. Upon cooling, the flux solidifies and remains on the surfaces as a thin, tightly adherent film, which need not be removed.
The recommended gap tolerances for flame brazing range from 0.1 mm to 0.15 mm. Larger gap clearances can be tolerated, but capillary action is reduced, gravity activity is increased and more filler metal may be required. Friction fits should also be avoided as this will restrict filler metal flow and result in discontinuities in the brazed joint area.
Powder Fluidity Indicator
Definition of the Variables and Calculation of the Results
The spraying factor Rm is a relative value for the evaluation of powders used for dry fluxing – especially when the material transport in the used equipment depends on the fluidization property of the powder.
Expansion factor [cm/cm] = Hfluid [cm] / H0 [cm]
For the calculation of the expansion factor, the mean values for Hfluid and H0 are used. The data for the mean values results from measurements of the powder height at 5 points.
Hfluid: powder height in fluidized condition
H0: powder height not fluidized and vibrator shut down
Hfluid = (Hfluid1+ Hfluid2+ Hfluid3+ Hfluid4+ Hfluid5) / 5
H0 = (H01 + H02 + H03+ H04+ H05) / 5
Powder flow (m) [g/ 0,5 min]
The mass (weight) of powder flowing out through the calibrated hole in 0.5 minutes calculated as median from 10 measurements.
Calculation of the median:
Median = m9+m2 / 2 for 10 single measures of m and m5< m3< m1< m7< m9< m2< m4< m8< m10< m6
Spray factor (Rm)
Rm [g/ 0.5 min] = m [g/0.5min]* expansion factor
The Spray factor results from the median of powder flow multiplied with the calculated expansion factor.
End of the article.
Flux Transfer Systems in Electrostatic NOCOLOK® Flux Application
For these experiments, a fluxing booth from Nordson designed for semi-continuous production was used. This unit’s (216 cm height, 143 cm width, 270 cm depth) key components are a hopper, a spray gun, two filter cartridges and the necessary control units.
The work piece is placed on a grating, which can be manually moved back and forth. The spray gun automatically progresses from left to right and back in intervals of approximately 21 seconds (21 seconds for 65 cm; 3.1 cm/s).
Responding to recent market developments, a second flux transfer system was installed in this fluxer. An ITW/Gema hopper including spray gun and control unit were added to the booth.
The distance between the spray nozzles and the grating is 34 cm
Principles of Flux Transfer Systems in Electrostatic NOCOLOK® Flux Application
The Nordson hopper utilizes the principle of powder fluidization to convey the flux via a Venturi pump and a feed hose to the spray gun. An agitator in the hopper supports flux fluidization.
The ITW/Gema system has a hopper with a helix screw conveyor to mechanically transfer the powder into a funnel. From there, a Venturi pump transports the flux through a hose to the spray gun.
Both systems are equipped with vibrators in some positions to reduce flux buildup. The spray guns are operated with 100 kV to charge the powder.
The design of the Venturi pump and the electrical spray gun of the two systems are very different from each other. However, in view of the experiments described here, this was only of minor influence. The focus is on trends of flux behavior when samples with fine and coarse particle size distributions are compared. Using the technology type rather than the manufacturer’s name is even more in line with the objectives.
Experiments with Flux Transfer Systems:
Trials to determine the consistency of flux flow and deposition on radiators were performed, using sample 1 and sample 2 in the powder fluidization (Nordson) and the mechanical transfer (ITW/Gema) equipment.
As the first step, the control units (for flow air and/or helix speed) needed to be adjusted for each test to a flow rate that provided a flux loading of approximately 5 g/m². The experiment was then continued for 30 minutes without changing the equipment settings. In intervals of two to four minutes, radiators were placed on the grating for coating, and then weighed to determine flux loading. Each test series included ten or eleven units.
The results are summarized in the table ”Flux Deposition on a Heat Exchanger”. Material of sample 1 (”coarse” material) showed relatively consistent behavior in both systems. Variations are in a range of 0.7 g/m² with powder fluidization and 0.5 g/m² with mechanical transfer. For sample 2 (”fine” material), the findings indicate more significant fluctuation with powder fluidization. The range is 1.4 g/m². In the mechanical transfer system, the variations of sample 2 are lower, 0.5 g/m².
Due to the influence (and statistical variances) of the experimental conditions with different Venturi pump designs and different spray guns in both systems, the results only provide information on trends. As could be expected, the equipment relying on powder fluidization showed lower consistency with fine material. It is more difficult to fluidize fine material, and consequently it is more difficult to transfer flux powder by fluidization.
Sample 1, which has better fluidization properties, has lower fluctuations in both systems. This indicates a more beneficial performance of a flux powder with good fluidizing capabilities.
Mechanical Transfer System – Sample 1
Mechanical Transfer System – Sample 2
Table 3: Flux Deposition on a Heat Exchanger (30 Minutes Test)
Powder Fluidization System – Sample 1
Powder Fluidization System – Sample 2
These results were confirmed by other observations made during additional trials. Material with spray factors of approximately 45 g/0.5 min and higher (with good fluidization) flowed through the equipment pipes and hoses with less material buildup and created lower amounts of residue buildup on the spray nozzles. Large particles also partially compensate for the Faraday Effect, which makes it difficult for the electrostatically applied flux to penetrate the fin package (center of heat exchanger with tubes and fins).
The experimental work for this paper identified and evaluated essential performance characteristics for electrostatic flux application:
- Powder fluidization
- Powder adhesion
The flux particle size distribution and the relative ratio of fine particles in the flux powder are key factors in dry fluxing.
A specific proportion of fine material in the flux is important for adequate adhesion.
Larger particles contribute to proper fluidization.
Equipment parameters for electrostatic fluxing must be adjusted to suit the specific flux properties.
Fluxes utilized for electrostatic application need improved fluidization characteristics, but not at the expense of adherence performance.
To be continued …
Flux Powder Adhesion:
Dust formation and flux fall off are of general interest to the brazing industry. Regardless of the application method, dust generation (particularly airborne fines) must be avoided or kept to a minimum. If dust formation cannot be prevented, local exhaust ventilation and meticulous housekeeping are recommended.
The inhalation of flux dust in high concentrations over a long period of time constitutes a health hazard to exposed personnel. Due to the abrasion caused by flux dust, unprotected equipment surfaces of moving parts can show premature deterioration if not regularly maintained.
As mentioned above, flux adhesion in dry application is lower than in wet application. Forced convection heating zones are one possible area in the process where flux losses may occur. Other factors might be manual transfer of units or vibrations during mechanical transport. Some users improve adhesion by applying the powder on surfaces still lubricated with residual evaporative oils.
When excess flux dust is generated in the drying oven or the furnace, it can get into the exhaust steam and create difficulties with the exhaust treatment (i.e., quickly overload the filter or scrubber). If the exhaust is treated with thermal or catalytic processes (i.e., incineration of evaporative lubricants), separation of solid and gaseous components can become necessary.
Excess flux dust in the brazing furnace can also settle on the conveyor belt or on the furnace muffle. The conveyor belt can take this powder through the brazing zones, where it eventually melts. This may contribute to accelerated corrosion.
Users of dry fluxing technology are aware of the reduced flux adhesion. At most of the operations we were allowed to visit, dust formation due to flux fall off is kept to minimum levels by appropriate technical installations.
Experiments for Flux Powder Adhesion:
We researched the ability of flux powders to adhere to aluminum surfaces in electrostatic application. A very simple test was used to determine adhesion tendencies. This experimental arrangement is not simulating real production conditions. Nevertheless, it provides very useful information.
A plain square aluminum plate (0.5 m x 0.5 m) is electrostatically coated on one side with flux powder. The total flux weight is determined to calculate flux loading. The plate is then dropped (in vertical position) from 5 cm height to the ground and the flux loss is registered as percentage of original flux weight.
Attached is a diagram with the results for flux sample 1 and sample 2. For each material, ten measurements were performed. There is a certain variation of the individual figures; nevertheless, the trend is obvious. Sample 1 (”coarse” material) shows an average loss of approximately 33% compared to approximately 3% of sample 2 (”fine” material). This general tendency of powder with larger particle size distribution to adhere less than fine powder was also confirmed by additional experiments we made. The flux fall off in wet flux application under these test conditions is approximately 1%.
Dry Flux Application on an Aluminum Plate (0.25 m2) Flux Loss for a Fall from 5 cm Height
Flux powder is electrically charged in the gun. Usually, adhesion in electrostatic application is dependent on electrical forces. The flux loses the charge when it hits the grounded heat exchanger. Gravitational forces are now competing with relatively weak Van-der-Waals forces. This explains why fine particles adhere better.
Large flux grains are affected more by gravity, and consequently fall off more easily. We were able to synthesize a flux with very large average particle size distribution which fluidized perfectly (spray factor 143 g/0.5 min; i.e., twice as ”good” as sample 1). However, when this material was used for electrostatic application, the air flow from the spray gun blew away a lot the flux just deposited on the surface.
To be continued …
Experiments for Flux Powder Fluidization:
To illustrate the relationship between flux properties and fluidization, a series of tests was carried out using Sample 1 and Sample 2. Attached are print-outs of the particle size distribution analysis (Sympatec Helios H0851; dry powder analysis with laser) of both materials.
Table 1: Particle Size Distribution
Sample 1 shows coarser grain structure than sample 2. There are considerably more fine particles in sample 2, and most of that material has a size of below 5 μm. The spray factor of sample 1 (”coarse” material) is 71.88 g/0.5 min. This correlates with very good fluidization properties which was confirmed during tests in the electrostatic spray booth (see below).
For sample 2 (”fine” material), a spraying factor of 7.35 g/0.5 min was found. This reflects extremely poor fluidization properties, also confirmed by tests in the spray booth. The above indicates that there are at least three material characteristics connected to particle size affecting fluidization:
- Average particle size
- Quantity of fine particles
- Maximum particle size
To further identify the effect of these factors, we tested mixtures of the two samples. In increments of 10%, sample 1 and sample 2 were blended. Then the spray factors of the mixtures were determined.
Table 2: Spray Factors in [g/0.5 min] of Sample Mixtures
As illustrated in the graph (see attachment), the relationship of spray factor and sample mixture ratio is not linear. Instead, it shows a rapid decline once the content of fine material is approximately 20 to 30%. We were able to specify the spray factor range of successfully performing flux powder to approximately 45 g/0.5 min in experiments with our dry fluxing booth, and from situations reported by our customers.
Spray Factor for Sample Mixtures
The ability of a powder to fluidize is very important for its performance in electrostatic application. However, it is not the only factor.
To be continued …