Premium Automotive Research and Development Project (PARDP) Cold Metal Transfer (CMT) Brazing of Galvanised Steel Sheets.

Transcription

1 Premium Automotive Research and Development Project (PARDP) Cold Metal Transfer (CMT) Brazing of Galvanised Steel Sheets. Project Summary Cold Metal Transfer (CMT) is a welding process developed by Fronius Welding, Austria. The process is based upon conventional Dip Transfer MIG welding where material deposition is initiated at the point of short circuit of the wire electrode into the molten weld pool. However, whereas traditional welding processes are controlled electrically CMT employs an innovative wire feed system, integrated to a high speed digital control which controls material transfer and the amount of thermal input to the work piece. CMT is characterised by an arcing period during which both the work surface and the wire electrode are heated. The molten electrode is then fed forward to make contact with the work surface. At the point of short circuit the welding current is cut practically extinguishing the welding arc and hence limiting the amount of heat transferred to the work piece. After a defined short circuit duration the electrode is mechanically retracted pinching the molten droplet from the end of the electrode. The arc is then re ignited and the process repeats. This hot cold, arc on arc off cycle is repeated up to 70 times per second depending on welding parameters employed. When compared to conventional MIG welding processes, CMT exhibits a high wire melting coefficient requiring in the region of 20 30% less thermal energy for welding. Although suited to a range of welding applications, notably thin aluminium, when considering arc brazing the process offers single sided access, with reduced part distortion - due to low thermal input, minimal weld spatter - greatly reducing the requirement for rework, and good gap bridging capabilities. This report details the results of PARDP trials conducted at WMG to evaluate the operating boundaries of this process (Process Window.) Trials were conducted based upon simple overlap joints constructed from DX52 galvanised steel sheets, brazed using a CuSi 3 filler wire in a pure Argon shielding environment. The weld power source used throughout these trials was a TPS3200 CMT unit integrated via Devicenet to an ABB IRB 2400L Robot. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 1 of 23 Last Updated 14/02/2007

2 Contents 1. Experimental Procedure 1.2. Applied Brazing Angles 2. Process Features 3. Comparative Analysis of Heat Input Between Brazing Processes. 4. CMTB Seam Braze Applications 4.1. Effect of Process Tolerance Mechanical Strength 5. CMTB Stitch Braze Applications 5.1. Examination of Braze Bead 5.2. Mechanical Properties Effect of Process Tolerance Mechanical Properties (Joints with 1mm and 2mm gaps) 5.5. Failure Mode. 6. Conclusion. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 2 of 23 Last Updated 14/02/2007

3 1. Experimental Procedure Experiments are conducted on simple lap joints constructed from 1mm thick material with a 20mm overlap and width of 100mm. Two separate sets of experiments are conducted. Seam Braze. Braze beads of 35mm length plan angle 0 0 as defined in figure 1. Stitch Braze. Short braze beads of 15mm length using plan angles of 0 0,20 0,45 0, 90 0 as defined in figure 2. For each experiment a total of 5-off samples are produced and an average of the results presented. Samples were fixtured in a purpose made jig to ensure repeatability with clamps situated 25mm either side of the joint line. In order that likely process condition can be replicated packing material is used to control joint gaps at 1mm and 2mm where defined. Top Sheet Bottom Sheet Figure 1: Braze Seam Sample Configuration. Braze direction arrowed. Top Sheet Figure 2: Stitch braze sample configuration using a variety of angles, length 15mm. Braze direction arrowed. Not to scale, shown on one joint for representation only. Macroscopic analysis is based upon a section being removed from the joint half way along the brazed bead in order that the influence of the start and end parameters can be eliminated. For stitch braze joints the critical bead geometry is measured, this being defined in figure 3. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 3 of 23 Last Updated 14/02/2007

4 Braze bead, starts and ends retained. Top Sheet Critical Bead Dimension Figure 3. Stitch Braze Critical Bead Mechanical testing is conducted on 35mm wide samples, accurately cut from the overlap joint. For seam brazed joints the start and ends are retained in order that process conditions can be replicated. For stitch brazed joints the critical bead dimension is centred in the middle of the sample. Seam and stitch braze tensile sample configurations are shown in figure 4. Each sample is deburred after cutting to eliminate the potential of crack propagation during testing. Samples are tested in tension on an Instron 5356 using a crosshead speed of 10mm min. Top Sheet Top Sheet 120mm A B 35mm Figure 4. Tensile sample configuration. A seam braze (starts and ends retained). B Stitch Braze. Not to scale. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 4 of 23 Last Updated 14/02/2007

5 1.1 Applied Brazing Torch Angles Bead geometry is dependent upon applied brazed torch angles. When considering overlap joints the available torch envelope is as much as 30 0 ( ) as defined in figure 5. However, consideration must be given to the effect on welding parameters when varying the applied torch angles. Due to the heat sink effect of the overlap joint, when inclining the torch into the overlap i.e. reducing the applied angle, greater thermal input is required in order that bead geometry can be maintained and adequate wetting of braze filler to the parent material can be achieved. This greater welding heat can in turn lead to greater part distortion and unacceptable zinc burn off. When considering applied torch angles selection must be based upon process conditions, in particular torch access. All documented brazing trials are conducted using a torch incline angle of 90 0 unless otherwise stated. Electrode stick out is maintained at 15mm For standard MIG brazing a torch travel angle of is typical. A recommendation from the system supplier is that the torch travel angle be maintained at 60 when CMT brazing, in order that arc heat can be dissipated ahead of the torch. Applying this travel angle also ensures that zinc burn off is initiated ahead of the arc and prior to deposition of the filler material, thus reducing the likelihood of the occurrence of porosity in the braze bead. Bead on plate (BOP) trials show that when maintaining constant welding parameters and applying an increased travel angle zinc burn off on the reverse joint sheet is evident. This suggests that the arc is more focussed on the parent material resulting in greater localised thermal input. If this angle is to be varied for example torch access, minor adjustments to the applied welding parameters are required Operating envelope Brazing Direction Stickout Length Figure 5. Torch Incline angle Figure 6. Torch travel Angle. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 5 of 23 Last Updated 14/02/2007

6 2. Process Features Simple bead on plate brazing trials are conducted to assess the operating principles of the CMT process. Trials are conducted during which the electrical transients of the process are captured and matched to high speed film images. A typical CMT braze cycle is shown in figure 7. This is defined as the time required to deposit filler material onto the work surface / joint and consists of two distinct phases of operation. 1. An arcing phase is evident represented by a constant arc voltage. This corresponds to an initial high pulse of current which ignites the welding arc. This heats both the work piece and the wire electrode. The current is then reduced in order that droplet detachment is not initiated but that a molten globule remains attached to the end of the electrode. The work surface is heated during this phase allowing vaporisation of the zinc coating prior to filler material deposition. 2. After a defined duration the electrode is fed into the weld pool initiating an electrical short circuit. This is represented by a reduction in arc voltage at this point. The corresponding welding current is reduced to a minimum extinguishing the welding arc and limiting the thermal energy transferred to the work piece. After a defined duration the electrode is retracted and the molten droplet is pinched from the end of the electrode and deposited on the work surface. The cycle then repeats. Reference to Figure 8 shows how the current is applied depending on the wire feed rate. At lower wire feed rates the current is applied in an initial pulse before being reduced to a minimum for the rest of the arcing period. This results in a low volume of electrode being melted (this being proportional to the wire feed rate), and low thermal input transferred to the work piece. As the wire feed rate is increased, greater current is required to melt a greater volume of electrode and transfer more heat to the work piece. This is achieved by increasing the duration of the initial arcing pulse of current which is then reduced to prevent material transfer before being further reduced at the point of short circuit. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 6 of 23 Last Updated 14/02/2007

7 Figure 7. Cold Metal Transfer operating Principles Figure 8. Cold Metal Transfer Current change as a function of wire feed rate.left 3m min, Right 6m min. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 7 of 23 Last Updated 14/02/2007

8 3. Comparative Analysis of Heat Input Between Brazing Processes BOP trials are conducted in which the heat input (HI) values of CMT brazing (CMTB), Dip Transfer brazing (MIGB) and Pulsed MIG brazing (PMIGB) are computed. Based on BS EN :1998 the HI values in J cm are calculated using the formula: HI = nui v Where n is the arc efficiency factor, U and I are the mean values for arc voltage and arc current respectively and v is the welding velocity in cm sec. The arc efficiency factor n is assumed constant as n=0.8, this is defined as the overall thermal efficiency of the MIGB process. Maintaining a constant welding velocity of 1m min -1 (1.666cm sec) wire feed is incrementally increased and the process values recorded. Figure 9 shows the relationship between HI and wf rates for the three brazing processes where CMTB exhibits substantially lower thermal input to the work piece than the other processes for the same amount of filler wire deposition Heat Input (j/cm) CMT Dip Pulse Wire Feed (Wf) Figure 9. Comparative Heat Input (HI) values of Brazing processes. Further trials are conducted examining the comparative average wire electrode melting coefficients of the three processes where using linear regression, the following relationship is derived where Im is the average measured welding current. CMTB wf = Im MIGB wf = Im PMIGB wf = Im Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 8 of 23 Last Updated 14/02/2007

9 The effect of the lower HI value and greater wire electrode coefficient of CMTB when compared to other processes is evident when considering parameter optimisation for constructing overlap joints. CMTB parameters are set using a wf rate that minimises zinc burn off but provides adequate joint wetting and filler wire deposition. This wf rate is kept constant and applied to the other processes. Process data is recorded and the HI values computed. The appearance of the weld bead is assessed visually and macroscopically as shown in figure 10 and table 1. The following observations are derived. CMTB HI = 215. Black soot like material deposited around the braze bead which is easily removed. Bead has a satisfactory appearance being narrow and uniform suggesting arc stability with no spatter. The reverse of the joint shows melting of the zinc layer which has re-solidified after brazing to maintain protective coverage. Cross section shows no evidence of pores or cracks with adequate wetting but with a high bead reinforcement height. MIGB HI = 462. Black soot like material evident as found in CMTB. Bead width is narrow and irregular with braze spatter. Reverse of the joint shows vaporisation of zinc layer which results in a hard whitish residue. Cross sectional analysis shows no evidence of pores or cracks with adequate wetting and high bead reinforcement height. PMIGB - HI = 610. Bead is regular although appears sunken with a greater width than the other two processes. Spatter is evident around bead. Reverse of joint shows vaporisation of protective zinc layer roughly corresponding to the bead width. Greater distortion than other processes. Cross sectional analysis shows a greater surface area of the joint is wetted with braze filler. Process stability/spatter Porosity HI Seam Appearance Wetting Zinc burn off CMTB MIGB - - o o PMIGB o very good, + good, o satisfactory, - adequate, - - unsatisfactory. Table 1. Comparative observations of brazing processes Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 9 of 23 Last Updated 14/02/2007

10 Figure 10. Comparative Braze Bead Profiles. Top CMTB, Middle MIGB, Bottom PMIGB Based upon these results the criteria for optimising arc brazed joints can be defined as employing brazing parameters which limit thermal input in order that damage to the zinc layer can be minimised whilst at the same time providing sufficient heat for melting of the electrode filler wire and providing adequate joint wetting. With CMT brazing the optimal criteria is met at the expense of reduced bead aesthetics particularly in a narrower bead with greater reinforcement height. Tensile tests conducted on simple overlap joints show failure of these joints is in the parent material with no disbonding in the joint interface suggesting that this optimisation criteria is adequate and the joint is stronger than the parent material. When considering both MIGB and PMIGB, the greater HI values and the lower electrode melting coefficients result in a situation where either zinc burn off is minimised at the expense of reduced filler deposition, or in greater filler wire deposition and improved wetting at the expense of unacceptable zinc burn off and greater part distortion. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 10 of 23 Last Updated 14/02/2007

11 4 CMTB Seam Brazed Applications Seam brazed joints can be applied in a similar manner to conventional MIG welding. Length of seam is dependent on process requirements although consideration must be given to part distortion i.e. a longer seam will produce greater heat in the joined component. Typically when considering MIG welding in automotive applications short seams are produced at regular intervals to stitch the component parts. This method has been applied for CMT brazing as shown in Figure 11 where a prototype roof joint is brazed using seams of length 35mm deposited at a pitch of 75mm. Figure 12 shows the same joint after repeated bending where although deformation of the component is evident, no visible damage to the braze seam results. Figure 11. Typical seam brazed application on prototype roof joint. Figure 12. Seam brazed prototype roof joint after repeated bending. 4.1 Effect of Process Tolerances The effect of changes in the part tolerance / joint fit up on the braze seam is examined based upon visual inspection of the braze bead. In particular, adequate joint wetting is identified. Figure 13 summarises the results where the maximum working envelope is defined as a joint with a 1mm gap and a 1mm sheet offset in the -ve X direction. Successful joining is however achieved at the expense of zinc burn off on the reverse of the joint as shown in figure 15, this being attributed to the diminishing heat sink effect as the joint sheet is moved in the ve X direction i.e. away from the wire electrode which has the effect of concentrating arc energy on the bottom joint sheet. Movement of the joint top sheet in the +ve X direction results in the braze bead being deposited predominantly on the top sheet with inadequate wetting onto the bottom sheet. As a result a maximum of 0.5mm movement in this direction can be tolerated, although this is at the expense of reduced bead aesthetic as shown in figure 14. In particular, a high reinforcement height and a reduced bead width results. Although further trials show that greater joint gaps and greater sheet offsets can be joined the repeatability of the seam becomes questionable. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 11 of 23 Last Updated 14/02/2007

12 +Z +2mm Working Envelope -X +X -2mm -1mm +0.5mm +2mm Figure 13. CMT Brazing Process Window. Figure 14. Braze bead with +0.5mm sheet offset showing high reinforcement height. Figure 15. Zinc Burn off, reverse of joint sheet with offset in -ve X direction Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 12 of 23 Last Updated 14/02/2007

13 4.2 Mechanical Strength Based upon the previously defined process window mechanical testing is performed. Brazed joints are tested in tension and compared to a control dog bone sample constructed from the parent material. Figure 16 shows the results where similar strength values are evident for all samples with a similar small spread of values. Notably the values are similar to those for the control sample LOAD KN material control 0ptimised "- x "1mm 1mmGap "- x "1mm/1mm Gap Figure 16. Mechanical Strength Defined Process Window Examination of the failure mode shown in figure 17 shows failure is in the parent material with no disboding of the braze filler. The samples exhibit deformation of the parent material and elongation in the region of 15mm prior to failure. This suggests that the parameter optimisation criteria is adequate and that the joint is stronger than the parent material. Figure 17. Failure Mode Process Window Figure 18 shows the tensile results for samples constructed outside of the defined process window. Although joining was successful, inadequate joint wetting with braze filler results. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 13 of 23 Last Updated 14/02/2007

14 The measured values are lower than those shown for the optimised samples with a greater spread of values indicating that the joint integrity is questionable Load KN ptimised "- x "1.5mm 1.5mmGap "- x "1.5mm/1.5mm Gap Figure 18. Mechanical Strength Outside of Process Window. Examination of the failure mode for these samples shown in figure 19 highlights failure is predominantly in the braze bead. Due to inadequate joint wetting the parent material has torn out of the filler with disbonding evident. Figure 19. Failure Mode Outside Process Window Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 14 of 23 Last Updated 14/02/2007

15 5. Stitch Braze Applications Stitch brazing is examined as a possible alternative application to seam brazing. In contrast to seam brazing, stitch brazing is virtually immune to surface movement in the joint sheets. However as the braze is deposited over the joint step/joint interface, two different thickness of material are brazed. As a result greater thermal energy is required than for seam brazing with the result that zinc burn off on the reverse of the joint sheet is unavoidable. Additionally this greater thermal input implies greater part distortion. However this can be countered by varying both the pitch and the sequence of joining. Trials are conducted examining likely process conditions with respect to joint gaps only. When applying stitch brazing selection of braze plan angle must be based upon torch access into the joint whilst maintaining previously defined torch brazing angles. Likewise, length of braze bead can be varied depending upon likely process conditions i.e. torch access, aesthetics of the joint is joint discrete / hidden, amount of part variation etc. When compared to traditional dip transfer MIG brazing, CMT brazed stitch joints appear virtually spatter free, despite brazing over the joint step. This is shown and compared in figure 20 and 21. Figure 20. CMT Stitch braze with virtually no spatter. Figure 21. MIG Stitch braze showing spatter around the joint. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 15 of 23 Last Updated 14/02/2007

16 5.1 Examination of Braze Bead Braze beads are deposited on over lap joints using the defined plan angles shown in figure 2. Stitch Brazed beads produced using the previously defined welding angles are shown in figure 23 to 26 with corresponding zinc burn off on the reverse of the joint. Visually the beads are virtually spatter free, any spatter evident is local to the bead being deposited within approximately 5mm of the seam. Similar to the results for Seam brazed joints, a black soot like deposit is evident on the parent material which is easily wiped off. Likewise residual zinc burn off is easily removed from the rear of the joint. Measurement of the critical bead dimension is conducted using the previously defined criteria shown in figure 3. Figure 22 illustrates how bead dimension is influenced by the braze plan angle. When employing shallower brazing plan angles ( ) a greater volume of braze filler is deposited into the joint i.e. a greater critical bead dimension is evident. Critical Bead Dimension (mm) Seam Angle Bead Dim 0 0 Seam 20 0 Seam 45 0 Seam 90 0 Seam (mm) Maximum Mean Minimum Figure 22. Measured critical bead dimension Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 16 of 23 Last Updated 14/02/2007

17 Figure Stitch weld x 15mm with corresponding zinc burn off Figure Stitch weld x 15mm with corresponding zinc burn off. Figure Stitch weld x 15mm with corresponding zinc burn off. Figure Stitch weld x 15mm with corresponding zinc burn off. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 17 of 23 Last Updated 14/02/2007

18 5.2 Mechanical Properties Tensile test results are summarised in figure 27. Greater mean strength is evident for samples brazed with 0 0 and 20 0 degree seams with the mean values diminishing for the greater angles. This can be attributed to the larger critical bead dimension of the 0 0 and 20 0 samples Load (KN) Seam Angle Load KN 0 0 Seam 20 0 Seam 45 0 Seam 90 0 Seam Maximum Mean Minimum Figure 27. Mechanical properties stitch braze joints Typical failure modes are shown in figures 28 & 29. For braze beads produced using angles of 0 0 and 20 0 failure is in the parent material adjacent to the braze bead. Beads produced using angles of 45 0 and 90 0 exhibit failure in the braze bead. This corresponds to the tensile data where brazed samples using angles of 0 0 and 20 0 exhibit greater mechanical strength due to the greater critical bead dimension. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 18 of 23 Last Updated 14/02/2007

19 Figure 28. Beads produced using 0 0 and 20 0 brazing angles showing failure in the parent Figure 29. Braze beads produced using 45 0 and 90 0 welding angles showing failure in the braze bead. 5.3 Effect of Process Tolerance Brazed joints are produced with gaps of 1mm and 2mm as previously defined. Introducing gaps into the joint has only a minor visual influence of the braze bead geometry. The seam is continuous and regular with only a slight depression in the bead in the area of the gap/joint. As defined for the braze seam trials, joints produced using 0 0 angle i.e. shortened weld seams have poor resistance to joint gaps. When employing gaps greater than 1mm these joints become unrepeatable with inadequate wetting of the joint sheets with braze filler. This is shown in figure 32. Visually greater spatter is evident for joints constructed with gaps between sheets. This appears to increase as the joint gap is increased. Spatter generation is however minimal and occasional and hence cannot be easily quantified. Reference to figure 30 shows minor spatter on the sample when welded at 45 0 with a 2mm joint gap. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 19 of 23 Last Updated 14/02/2007

20 When employing joint gaps greater than 2mm for all of the bead angles employed, melting / damage to the parent material occurs as defined in figure 31. Although adequate joining is achieved this situation is unacceptable from a quality point of view. Figure 30. Stitch braze (45 0 ) with 2mm joint gap showing minor spatter generation (arrowed) Figure 31. Stitch Braze (45 0 ) with 2.5mm joint gap showing damage/melting to parent material. (arrowed) Figure 32. Stitch braze 0 0 with 2mm joint gap showing inadequate wetting around joint line.(arrowed) Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 20 of 23 Last Updated 14/02/2007

21 5.4 Mechanical Properties (Joints with 1mm & 2mm gaps) Mechanical properties of stitch brazed joints constructed with gaps of 1mm and 2mm are summarised in figures 33 and 34 respectively. Stitch joints using a plan angle of 0 0 are omitted from the 2mm gap trials due to the joints being unrepeatable as previously mentioned. For both sets of results adequate joint strength is evident being similar to the previous trials conducted without joint gaps. This indicates that that in terms of mechanical strength, joint gaps have very little influence on the mechanical performance of stitch brazed joints LOAD KN Load KN 0 0 Seam 20 0 Seam 45 0 Seam 90 0 Seam Maximum N/A Mean N/A Minimum N/A Angle of Seam Figure 33. Mechanical Properties Stitch braze with 1mm gap. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 21 of 23 Last Updated 14/02/2007

22 Load KN Angle of Seam Figure 34. Mechanical Properties Stitch brazed joints with 2mm gap. Failure Mode In contrast to the joints brazed without gaps, all of the tested samples brazed with 1mm gap and 2mm gaps fail in the parent material of the joint.this is typically local to the braze bead being in the HAZ. This suggests that wetting of the joint gap with braze filler improves joint performance. Figure 35 show typical failure modes. Figure 35. Typical failure modes for stitch brazed joints with gaps Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 22 of 23 Last Updated 14/02/2007

23 6. Conclusions Process Benefits. Single sided joining process. Low heat input and higher wire melting coefficient than conventional dip transfer and pulsed MIG brazing. Low spatter generation reducing the requirement for rework. High strength joints. Under optimal brazing conditions joint strength is as good as the parent material. Stitch brazing has a large process window with respect to part fit max joint gap 2mm, Wide process window with respect to applied torch angles and electrode stick out. As the arc length is controlled mechanically the process is insensitive to changes in the joint surface under normal operating conditions e.g. Part distortion. Joint strength enhanced with small joint gaps. Low thermal input and lower part distortion joint gaps better maintained. Defective brazed seams can be effectively reworked using either TIG or Oxy Acetylene welding. Process Limitations. Seam brazing is sensitive to production / part tolerances. Control of spatter generation is dependent on brazing speed i.e. greater speed = greater spatter generation. Zinc burn off on reverse of joint may require rework in addition to being a possible corrosion issue. Owner: PARDP (C.Pickin) / AJT (mshergo1) Page 23 of 23 Last Updated 14/02/2007