I have been working through mechanical and material properties of various silver bronze filler alloys since seeing the same recommendation of Castolin 38230 for Columbus Cento tubes. Columbus also recommend Castolin 38230 for all their other steel tube materials. 38230 has an ISO designation Ag 230, composition (Ag 30%, CU 38%, Zn 32%). The exception is for Columbus stainless steel XCR they recommend T99 (Ag 56% Cu 22% - Zn 17%). T99 is simply a 56% silver braze filler used for lug joints like Harris Safety Silv 56. Note, T99 and Safety Silv 56 are silver braze fillers with 5% Sn. If you have a look at bikes from Saffron Frameworks London, "33% silver" is designated for fillet brazed frame using XCR tubing. Saffron state that they use Fillet Pro for silver brazing. I have used Fillet Pro but have never got an element composition. Wade Barcosi did confirm that Fillet Pro has an ultimate tensile strength (UTS) of 430N/mm (430MPa), a solidus of 1200 F (649 C) and a liquidus of 1330 F (721 C). Solidus and liquidus temperatures that align with Fillet Pro are silver alloys with 38% Ag such as Harris Safety Silv 38T (Ag 38%, CU 32%, Zn 28%, Sn 2%). Safety Silv 38T has an ISO designation Ag 138. Ag 138 from Johnson Matthey states a UTS of 430 MPa and a fluidity of 2. By comparison JM states a fluidity of 3 for Ag 230 and 1 for Ag 156 (56% Ag). 1 = Free flowing filler metal when molten, 2 = Medium flowing filler metal when molten, 3 = Sluggish flowing filler metal when molten. The Harris 'fluidity' scale is similar; Ag 156 = 8, Ag 138 = 7, Ag 230 = 6, a lower number is less fluid.

The Johnson Matthey silver braze catalogue is a great reference list where you can see the effects of filler composition on material properties such as wetting, flow and UTS. Unfortunately the UTS for fillers without tin like 38230 (Ag 230) is not listed in the JM catalogue, however, Stella from Italy has this information.

Unfortunately as a new member I cannot insert URLs or attachments.

You will see that the addition of tin is to lower the melting temperature, however, it does not necessarily make the filler more fluid, e.g. JM Silver-flo 33 (Ag 33%, Cu 33.5%, Zn 33.5%, Sn 0%) has a fluidity of 1 with an upper liquid temp of 740 C compared to Silver-flo 34 (Ag 34%, Cu 36%, Zn 27.5%, Sn 2.5%) with a fluidity of 3 and an upper liquid temp of 730 C. Silver-flo 33 UTS 540 MPa, Silver-flo 34 UTS 470 MPa, Ag 230 UTS 490 MPa. Flo 33 has a narrow 40 C liquid range, Flo 34 has a wide 100 C liquid range, Ag 230 has a wide 95 C liquid range.

Also of importance are the braze filler properties such as joint bonding strength, resistance to corrosion, ductility (elongation) and fatigue resistance. So you need to use a material that works and is matched to the tube material. In this regard I would trust the Columbus recommendation.

What is my experience? I have manufactured 6 frames since 2010, 3 using nickel bronze (nickel silver in the USA), and 3 using silver (Ag 38%, CU 32%, Zn 28%, Sn 2%). The 3rd frame I was set up to use Fillet Pro but could not get results I was happy with (unable to weld up slope and down slope like nickel bronze, and some porosity). I did use Fillet Pro to finish the rear triangle on the 3rd frame. This included breezer rear dropouts (first time). Following this I sought advice from the Fillet Pro agent in Australia about the best welding technique. His advice was pretty well was the same as from a post on this forum in 2009. It's tricky. Very fluid weld pool, a real technique to build a wider fillet, unable to weld in a position other than very slightly up hill. On a whim after finishing this 3rd frame I did a trial weld using a 38% silver filler rod from the UK sold as a lower cost rod for lug brazing (SIF No. 39) and found it was easier to build a fillet and porosity was much less. But, still no where near as 'weldable' as nickel bronze (SIF No. 2) with the purity of weld. I have never had any porosity using nickel bronze and I can weld at high angles up slope and down slope. It's just that the high temp of nickel bronze creates distortion of head tubes, bottom brackets and seat tube / seat stay junctions that require post reaming and re tapping / facing of BB threads. Hence, why trialling Fillet Pro and using SIF No. 39, but no more. Either I find a silver filler that is weldable or I go back to nickel bronze and weld smaller size fillets similar to the size that Dave Anderson uses. The importance of filler size was drummed into me by Jim Cook from Nimbus Cycles. According to Jim the fillet throat size is calculated by the ratio of tube UTS versus filler UTS plus 100% safety factor. e.g. Columbus Niobium (Spirit / HSS / Life) UTS 1250 MPa, Fillet Pro 430 MPa, Max 35 dia bi-oval 0.8mm joint end = 1250/430 x 0.8 x 2 = 5.8mm. Using nickel bronze at 540 MPa UTS reduces this fillet throat size to 4.6mm. Of course if a Spirit 35 dia down tube was used the throat size would reduce by the ratio of 0.65 / 0.80. I have also tried to contact Vincenzo Forgione in Italy to discuss silver filler materials. Vincenzo exclusively uses silver for his fillet brazed frames. He has been using the same technique for over 25 years making everything from mountain to fold up bikes. Vincenzo's fillet sizes are small, similar Saffron.

Nickel bronze was my initial filler material of choice after many discussions with Jim Cook of Nimbus Cycles. Jim was a mechanical engineer and metallurgist in the UK nuclear power generation before he started Nimbus Cycles. He made many bikes for TdF riders as well as doing research for time trial rider positioning for maximum output in the early days of sports science medicine with the University of Leicester. Jim was 76 back in 2010 when I did this research for making frames. He was a big believer in nickel bronze due to its higher strength (UTS 540MPa) and very low porosity. He also used to post shape the fillet welds using the torch rather than shaping with a file. This was a technique he learnt from a welder that used to weld engine subframes in Spitfire WWII aircraft. I have yet to try the post weld torch shaping technique.

One thing to consider is that with silver where the position (only slightly up hill) is required and building a fillet is difficult, it therefore takes longer to weld a joint and hence more heat builds up in the job. Hopefully Castolin 38230 (Ag 230) takes away these downsides. Any opinions and information is welcome.