Fillet Size Follow-up

[Major Tom]

Inspired by the very helpful question and answer in "the Path" about fillet size (see: http://www.velocipedesalon.com/forum...ize-43619.html) I'd like to add some follow-up questions:

My experiences with first fillet brazing with Fontargen A210 are similar to what I've learned at school, but still are not clear to me:
I can melt and add the brass in a "fillet shape" line around the tube-intersection. I also can (after some training) control and influence fillet size and steel temperature quite well and create bigger and smaller fillets without separation of copper.

BUT: the inner fillet is always very small when I do the brazing in one step. This is always the case, no matter how fast or slow I work, no matter how big the outer fillet is. (Never looks like the pic shown by Ben in the initial thread:

But usually like this:


If I work with a lot of heat, I can create a very flat outer fillet and some significant inner fillet.
Also, if I start with a small outer fillet, I can melt it a second time to make it flow to the inside. In both cases, an additional workstep gets necessary to create a smooth fillet.

1.) Is it possible and useful to create a significant inner fillet (>0.5mm) and a properly sized outer fillet at the same time (without totally overheating the tubes?
2.) Is a beatiful outer fillet and some penetration into the gap all you can ask for?
3.) I can create beatiful inner and outer fillets by heating up the seam three times - but I think, that cannot be the right way...

I thought, I had sorted the fillet brazing out at a base level, but now I'm confused... :f:

Clarification highly appreciated: For how much of an inner fillet do you aim?
And to what degree do you sacrifice "steel overheat prevention" for brass penetration

[Mark Kelly]

I couldn't answer the last time. The thing that is missing in the discussions you reference is an understanding of plastic constraint.

In simple terms plastic constraint is a strengthening effect which applies to joints where there is a thin layer of relatively weak* material sandwiched between between two parts of relatively strong* material and the width of the joint is substantially greater than the thickness of the material (eg it has a high aspect ratio). In this case the stress on the joint will largely be absorbed by the stronger material because it prevents the movement of the weaker material. In engineering parlance it constrains the weaker material from entering plastic deformation, hence the name plastic constraint.

For more information on this, Google “Plastic Constraint Brazed Joint” and read the extract of the book Advances in Brazing Science Dušan P Sekulić, Elsevier, 2013.

The upshot of this is that simply comparing the tensile strengths of the two materials and saying that you’ll need four times as much of the one that’s a quarter as strong ignores what’s actually happening in the joint**.

If you doubt that plastic constraint applies to a fillet brazed joint, look at Dazza's cutaway here:

24838714940_430464f03e_b.jpg

The strip of brazing alloy in the actual joint between the two tube members has an aspect ratio of about 16 : 1 by my rough measurement, more than enough for plastic restraint to apply. Yes, I know you're not Dazza and neither am I but we should all be aiming for mitre fit up this good: it's not just aesthetics, it's good engineering.




* Weak = lower stiffness and lower strength, strong = higher stiffness and higher strength.

** and it also completely ignores the stress path and the difference between extensional strength and shear strength, but that's a different discussion.

[Mark Kelly]

I realised I didn't directly answer your questions, but my take on them flows from the answer I gave above:

1. I don't think a large inner fillet achieves anything.

2. Aim for complete penetration into the narrowest possible gap, the rest is cosmetics.

3. I don't think it's the right way either. From 1 and 2, it isn't necessary.

[devlin]

But also note Mark he pulls a bunch of silver through with heat to form a good size fillet on the opposite of the join on the inside of the tube. While an accurate miter is essential, having enough meat in the fillet itself giving larger fillet to tube surface area is essential as well. I'm probably already telling you what you already know though.

25108014786_792027ea81_b.jpg

[Mark Kelly]

, having enough meat in the fillet itself giving larger fillet to tube surface area is essential as well. I'm probably already telling you what you already know though.

While I know that that's how it's normally done, I don't know that it is essential.

I haven't seen any experimental data to decide one way or the other, I would be interested in anything anyone has done in this area: in theory I can see very little benefit to a large fillet but experiment trumps theory every time.

I currently fillet braze the stainless chainstays to the BB shell for frames that have SS chainstays and I lay much, much smaller fillets than those shown and in silver to boot. On the other hand these joints have a composite structural overlay so it's a bit moot.

BTW I was mostly taking aim at the simplistic and ignorant thinking that went into the "rule" stated in the insert in the OP.

[shrodinger]

The Welding Institute TWI and Sheffield university covered this some years ago and made public information that showed the surfaces between joints made impacts greater than fillet size

Aesthetically I'm not sure how you could photo of a fillet and claim them as beautiful if only 3-5mm radius. The part of fillet brazing is also the flow.

Torben

[devlin]

I went onto the TWI website and registered but am struggling to find the relevant information. I know in the frame building world we call it fillet brazing but I am sure the wider welding community calls it something else. Any help here?

[Mark Kelly]

The Welding Institute TWI and Sheffield university covered this some years ago and made public information that showed the surfaces between joints made impacts greater than fillet size

Aesthetically I'm not sure how you could photo of a fillet and claim them as beautiful if only 3-5mm radius. The part of fillet brazing is also the flow.

Torben

I'd love to see this if you can provide a link.

Personnaly I like the look of 5mm fillets but I'm sure they don't inspire confidence in the average observer.

[Major Tom]

I couldn't answer the last time. The thing that is missing in the discussions you reference is an understanding of plastic constraint.

Thanks so much for your answer, your thoughts and the discussion started right now. There is a lot to learn from that for me...
Maybe there is some misunderstanding - maybe due to my weak English also...

I am fully aware, that a steel to steel junction will take way more stress (particularly per mm²) than the brass - but the "gap effect" was not in my attention. Thanks a lot for this link!
Also I am aware, that brass in greater distance may distribute the stress but with less influence farther away from the junction itself.
But based on my understanding of mechanics, in the junction of two tubes main stress will be "tension" and some shear.

Based on that, there might be an optimum for
a) overall fillet intersection size (that might no be in the range of "huge" to avoid destroying the elasticity of the materials)
b) inner/reverse fillet size
c) outer / primary fillet size
I would expect - based on my mechanical experience- something in the 4x wall thickness inner and outer to be the mechanical optimum. Although due to the brazing prozess this may not be a useful goal. But the provided link to "Advances in Brazing" seems to show, that stiffness/rigidity inside the joint areas are is pretty much the same as base material values. So maybe "our" fillet size is somehow over the top anyways most of the time...
Also we may imagine - to reach the same stability in the brass joint as in the steel base material - we will have to add ony some rather small amount of brass to the joint (fillet) to make it evenly robust at the yield strngth limit of the steel tube.

What I understood so far:
Brass fillets will give a very good amount of stability at rather small fillet size.
Smooth fillet into tube transition (geometric) may be of higher priority.
Outer fillet plus proper joint penetraition seems to be enough to build stable frames
Even and small "gap" at the tube-to-tube-joint is crucial to improve joint stability

Conclusio: Aiming for a joint as shown in your picture should be a very good starting point :D:

[Ben Hudson]

I couldn't answer the last time. The thing that is missing in the discussions you reference is an understanding of plastic constraint.

In simple terms plastic constraint is a strengthening effect which applies to joints where there is a thin layer of relatively weak* material sandwiched between between two parts of relatively strong* material and the width of the joint is substantially greater than the thickness of the material (eg it has a high aspect ratio). In this case the stress on the joint will largely be absorbed by the stronger material because it prevents the movement of the weaker material. In engineering parlance it constrains the weaker material from entering plastic deformation, hence the name plastic constraint.

For more information on this, Google “Plastic Constraint Brazed Joint” and read the extract of the book Advances in Brazing Science Dušan P Sekulić, Elsevier, 2013.

The upshot of this is that simply comparing the tensile strengths of the two materials and saying that you’ll need four times as much of the one that’s a quarter as strong ignores what’s actually happening in the joint**.

If you doubt that plastic constraint applies to a fillet brazed joint, look at Dazza's cutaway here:

24838714940_430464f03e_b.jpg

The strip of brazing alloy in the actual joint between the two tube members has an aspect ratio of about 16 : 1 by my rough measurement, more than enough for plastic restraint to apply. Yes, I know you're not Dazza and neither am I but we should all be aiming for mitre fit up this good: it's not just aesthetics, it's good engineering.




* Weak = lower stiffness and lower strength, strong = higher stiffness and higher strength.

** and it also completely ignores the stress path and the difference between extensional strength and shear strength, but that's a different discussion.

Hi Mark - I can't find any reference to plastic constraint in the bits of AiBS that are on Google Books. Do you have a page reference?

[Mark Kelly]

There is no reference to plastic constraint per se but section 2.2.2 assumes plastic constraint in its analysis of the applicability of the Tresca and von Mises criteria to failure of brazed joints. For instance on page 34 he writes

"Even in such a ductile material as pure silver, the level of constraint is so high that hydrostatic stress is very close to the axial one, which means that the values of the principal stresses are very similar."

[Major Tom]

Today I made a fillet to show, what I mean:

The fillet on the right side was built up as usual, but heated up to melt into the gap in a second workstep.
Fillet on the left is just a usual fillet, made in one step.



Sorry for the raw filing.
But I believe you can see the r=1mm fillet on the inside left and the much bgger fillet on the right side.
I still van imagine, that right side version might addd some (additional, unneccessary) strength to the joint...

[ljm]

I'm reminded of Peugot's internally filleted frames, which if memory serves had minimal external fillet. I'm too young (or my memory is failing with age) to remember whether there were any issues with those frames, in (long-term) use. But that reverse approach offers an interesting counterpoint to this accepted approach.

[shand]

Today I made a fillet to show, what I mean:
I still van imagine, that right side version might addd some (additional, unneccessary) strength to the joint...

It may do but you'd have to question if the additional heat cycle makes it worthwhile. I'm still not 100% sure why you're unable to get that internal fillet in one pass?

[EricKeller]

maybe it's because the inside of his tube isn't properly cleaned?

[Gattonero]

I'm reminded of Peugot's internally filleted frames, which if memory serves had minimal external fillet. I'm too young (or my memory is failing with age) to remember whether there were any issues with those frames, in (long-term) use. But that reverse approach offers an interesting counterpoint to this accepted approach.

I'm no framebuilder but I can tell you that I've seen several of them frames crashed and none having issues with the joints.
Brompton does use the same principle of brazing, with zero failures.
Seems to save a lot of time in welding, when aesthetics are not paramount.

[ljm]

I'm no framebuilder but I can tell you that I've seen several of them frames crashed and none having issues with the joints.
Brompton does use the same principle of brazing, with zero failures.
Seems to save a lot of time in welding, when aesthetics are not paramount.

I believe Peugot's used, essentially, a high-tech hearth, quite an automated process. Then of course, there's no work to shape and finish the fillet (I know many aim for zero file work). So I imagine you're right; there would be time savings in production. I guess some of that time would be lost to QC; as there's very little to visually inspect, I suppose there was greater use of NTD.

As for aesthetics, I actually like the look of those joints.