Quantcast

Best gusset design?

Pslide

Turbo Monkey
Anyone know or know someone who has analyzed the most effective type of gusset for strengthening the headtube junctions?

i.e. which is better, this type:



or this type:




This would be for a steel dh bike. I have my own opinions, but would like to hear others...
 

ATOMICFIREBALL

DISARMED IN A BATTLE OF WITS
May 26, 2004
1,354
0
Tennessee
I'm no scientist or R&D guru but the second version would be my guess as the better one.Seems like the whole headtube area is strengthened & maybe less overall flex!
You could always add both & problem solved.
 

jonKranked

Detective Dookie
Nov 10, 2005
85,559
24,182
media blackout
Well I think stronger = more efficient.

The whole point is to eliminate the possibility of a failure, so stonger yes, and more efficient and dispersing forces, yes. :)
stronger doesn't always = more efficient. I think what you are trying to accomplish is maximizing your strength to weight ratio (which is typically a reasonable measure of efficiency)
 

Tame Ape

BUY HOPE!!!!!!!
Mar 4, 2003
2,284
1
NYC
Anyone know or know someone who has analyzed the most effective type of gusset for strengthening the headtube junctions?

i.e. which is better, this type:



or this type:




This would be for a steel dh bike. I have my own opinions, but would like to hear others...
My suggestion would be to look at the leading products out there and copy that. Then, find a real engineer and have a loooong talk with them.
 

buildyourown

Turbo Monkey
Feb 9, 2004
4,832
0
South Seattle
Option A, BUT:

They should be tapered to a parabolic shape at the end. Not the U with point as in the picture.

They should be affixed with silver. When they are welded on, it leaves the ends open and they will rust from the inside out. Brass creates too big of a HAZ.

edit - and be short enough that they end in the thicker butt of the tube if you are using butted tubing.
 
Last edited:

Pslide

Turbo Monkey
Option A, BUT:

They should be tapered to a parabolic shape at the end. Not the U with point as in the picture.

They should be affixed with silver. When they are welded on, it leaves the ends open and they will rust from the inside out. Brass creates too big of a HAZ.
I wasn't going to braze the frame, but you can TIG the frame and then braze the gusset?

Thanks!
 

Steve M

Turbo Monkey
Mar 3, 2007
1,991
45
Whistler
You gotta look at the failure mode you're expecting. If you're expecting buckling of the tube due to compression, type A offers significant advantages in terms of thickening/stiffening/strengthening the tube, but it doesn't necessarily do much for the weld strength in tension since you've got two welds directly over one another, and as BYO said, there is plenty of opportunity to inadvertently create stress risers at the ends of them - heaps of the old Planet X and On One frames used to crack at the ends of those style of gussets, for example. If you're running relatively thick walls in the tubes there however, I'd suggest other styles of gussets.

Type B provides significant stiffness that helps protect the welds on the top of the downtube and bottom of the top tube. Personally, due to the relatively small 2nd moment of area of steel headtube junctions, I'm inclined to say one of these gussets would be pretty beneficial. Seen a couple of frames (aluminium) break at the inside of the headtube junction there recently, so I'm convinced it's a real concern.

And there ends my armchair expertise :)
 
Last edited:

mrbigisbudgood

Strangely intrigued by Echo
Oct 30, 2001
1,380
3
Charlotte, NC
Every situation is unique. Every suspension system will load a frame in a different way. Different size frames will load junctions in different ways. There is no off the shelf best solution. There is no simple answer. If there was a simple, everybody would be an engineer.

Are you designing a frame in CADD? Try running some FEA simulations!! FEA done properly is really good at predicting failures. I use it every day at work with incredibely accurate results.

Build Prototypes. Do destructive testing. Find your failure modes, identify the root cause and correct the problem.

I guess I would be one of those engineers you should have a long discussion with.
 

Pslide

Turbo Monkey
You gotta look at the failure mode you're expecting. If you're expecting buckling of the tube due to compression, type A offers significant advantages in terms of thickening/stiffening/strengthening the tube, but it doesn't necessarily do much for the weld strength in tension since you've got two welds directly over one another, and as BYO said, there is plenty of opportunity to inadvertently create stress risers at the ends of them - heaps of the old Planet X and On One frames used to crack at the ends of those style of gussets, for example. If you're running relatively thick walls in the tubes there however, I'd suggest other styles of gussets.

Type B provides significant stiffness that helps protect the welds on the top of the downtube and bottom of the top tube. Personally, due to the relatively small 2nd moment of area of steel headtube junctions, I'm inclined to say one of these gussets would be pretty beneficial. Seen a couple of frames (aluminium) break at the inside of the headtube junction there recently, so I'm convinced it's a real concern.

And there ends my armchair expertise :)
Thanks socket. I kinda was seeing things the opposite of you, so it's good to see a different perspective.

Most of the failures I've seen are at the weld at the top of the top tube where it meets the head tube. It seems to me that Option B wouldn't really reduce the tension there, but Option A would reinforce that area and spread the load over two welded surfaces rather than just one.

I agree that Option B would be a lot stiffer, but might that also increase the localized forces in critical areas?

Of course, I could use both gussets at the same time, but I think aesthetically that looks overkill.
 

Pslide

Turbo Monkey
Every situation is unique. Every suspension system will load a frame in a different way. Different size frames will load junctions in different ways. There is no off the shelf best solution. There is no simple answer. If there was a simple, everybody would be an engineer.

Are you designing a frame in CADD? Try running some FEA simulations!! FEA done properly is really good at predicting failures. I use it every day at work with incredibely accurate results.

Build Prototypes. Do destructive testing. Find your failure modes, identify the root cause and correct the problem.

I guess I would be one of those engineers you should have a long discussion with.
Thanks for the response man, appreciate it. I don't have FEA capability, except what is in my head, which is pretty limited! It seems like a great study for someone still in school...

I also don't have the ability to build multiple prototypes or do destructive testing, but that would be ideal.

I'm looking at having a one off custom steel dh frame built of my own design. I've always thought gussets can be a unique signature for a frame aesthetically, but above all they must do their job. It's really the only part of my design that I've not worked out, so bouncing ideas off you guys helps.

What I can say is that it seems that Option A is much more common than Option B, and Option A is not necessarily easier than Option B - if anything it's harder. So there must be a reason other builders have chosen A over B.
 
Feb 14, 2011
18
0
Hello,

Without access to calculation or testing I would try and think of the gusset design like this.

Its purpose is to avoid stress concentrations by making the rate of change of stiffness gradual across the load path and to reduce the peak stress at the weld (welds will propergate cracks at lower stress cycles compared to parent metal). To describe in a different way you are trying to spread a peak high stress at the join over a large area along the gusset.

Design one is pretty good, it can be improved as someone has already said by a nice curve tapering out at the ends to spread the change in stiffness and avoid a stress concentration.

http://www.cotic.co.uk/product/BFe <- like this.

In design two the large, consistant, depth of the gusset will result in a big change in stiffness and a stress concentration at the end of the gusset where it is welded to the tube (look up second moment of area - its depth cubed). This could result in fatigue crack propergation in this area. I would try and improve this gusset by making it taper away at the ends where it meets the tubes with a nice curve, sort of like a swallows tail. I would also weld the length of it (although this could be difficult to manufacture in terms of alignment) as fatigue design codes tend to give a higher allowable for stresses parrallel to welds rather than perpendicular to welds (i.e. don't weld across the principal stress direction or have avoidable weld ends in areas of high stress). However Curtis use a similar design and they are very tough.

One thing both designs do well is to weld only to the middle of the tube which is the neutral axis in a vertical loading. This means that the welds will reduce the fatigue life in an area of low stress. Welding the first design along the back to seal it would be disasterous as it would put a worse feature (weld versus parent metal) in a more highly stressed location as well as being perpendicular to the principal stress.

Hope this helps, on reading it back it sounds confusing, but then most fatigue design codes run to 150+ pages! Basically nice gradual curves, make it the right sort of weight and you'll be fine :thumb:. Cheers.
 

Pslide

Turbo Monkey
Thanks guys, really liking this feedback. (Why I hang around RM.)

So I just did a little sketch combining all of your ideas, and include both gussets for maximum strength, it would look something like this:



Apart from watching for stress risers in the welding, anything further I can do to optimize?
 

zdubyadubya

Turbo Monkey
Apr 13, 2008
1,273
96
Ellicott City, MD
Apart from watching for stress risers in the welding, anything further I can do to optimize?
1.5 or 1.125 headtube?

tubing diameter?

I think your analysis/optimization would be different whether you are using similar diameter tubing or mating a slender downtube/top tube to a bigger diameter headtube.
 

Pslide

Turbo Monkey
1.5 or 1.125 headtube?

tubing diameter?

I think your analysis/optimization would be different whether you are using similar diameter tubing or mating a slender downtube/top tube to a bigger diameter headtube.
Because it will be steel, I think I'm limited to 1.125, and I don't think I want to pay the weight penalty of 1.5 even if it is available.

Top and downtube diameters are also relatively small, but I don't have the specs in front of me. I am speccing True Temper, Supertherm if available. Will use the largest diameter downtube they make that is suitable.

NOTE: Don't expect to see pictures any time soon, this is a long term project.
 

Slater

Monkey
Oct 10, 2007
378
0
I'd be careful with True Temper and Supertherm. In my experience and that of some other frame builders I've spoken with at length, many of the super high KSI air hardened steels are arguably too brittle for DH use. I was told to find the strongest tubing I could that had an elongation not less than about 12-15%. These slightly more elastic steels while having less yield strength, are far less prone to catastrophic failures. Also, with a steel frame, you often have to cold set the tubes if something pulls out of alignment while cooling. With these extremely strong/brittle tubes, cold setting is not possible, as it can often crack while doing so.

I've used the butted tubing from Nova with great results. It is stronger than straight normalized tube, has great elasticity/elongation, and is very reasonably priced. I have never once felt while riding any of the bikes built with these tubes that it lacked stiffness and would have been better with 185 KSI tubing. Strength and stiffness is more in the design than anything else.

Food for thought.
 

buildyourown

Turbo Monkey
Feb 9, 2004
4,832
0
South Seattle
Supertherm is also limited to shorter lengths (last time I checked)
It's marketed as BMX tubing so building a low BB frame makes the DT too long.

Versus HT is a very good compromise. Easy to work with too.


Keep in mind that every welded structure flexes. Your goal is not to eliminate flex, as an overly stiff structure just concentrates load to the weakest area.
Steel has incredible fatigue life, and unlike aluminum, can cycle a lot and still be safe. I prefer option A as it doesn't concentrate loads but doubls the wall thickness on the DT. When a DT fails, it first buckles.

You're second sketch is backwards of my preferred gusset shape. They should look like the end of your thumb.
 

Pslide

Turbo Monkey
Supertherm is also limited to shorter lengths (last time I checked)
It's marketed as BMX tubing so building a low BB frame makes the DT too long.

Versus HT is a very good compromise. Easy to work with too.


Keep in mind that every welded structure flexes. Your goal is not to eliminate flex, as an overly stiff structure just concentrates load to the weakest area.
Steel has incredible fatigue life, and unlike aluminum, can cycle a lot and still be safe. I prefer option A as it doesn't concentrate loads but doubls the wall thickness on the DT. When a DT fails, it first buckles.

You're second sketch is backwards of my preferred gusset shape. They should look like the end of your thumb.
Yeah, I've not done all the tube research at length, but I thought TT was coming out with some longer tubes now. I'll have to look again and check out Nova as well. Trying to keep things light.

I thought that what you meant in your original post regarding the gusset shape. You really don't see too many like that though, it seems most are inverted like shown.

I agree with your concerns with an overly stiff structure, and had this in mind when I posed my original questions.

I guess there are varying opinions on both counts.
 

Slater

Monkey
Oct 10, 2007
378
0
I have to agree on not making the gussets have sharp points as well. The HAZ of the welds on CrMo has a fatigue strength that is about 30-40% less than that of the rest of the tube. Hence why it is beneficial to disperse any loads in the HAZ as evenly as possible. Sharp points like that generally concentrate loads.
 

4130biker

PM me about Tantrum Cycles!
May 24, 2007
3,882
447
Just curious here (since I think doing the gussets with silver is a cool idea). Could using silver on the entirity of gusset option A create a problem similiar to what is mentioned by Chequebookracer?:

"Welding the first design along the back to seal it would be disasterous as it would put a worse feature (weld versus parent metal) in a more highly stressed location as well as being perpendicular to the principal stress."

If so, I'm assuming it would be possible to silver solder only the sides of design A? I guess I don't completely understand how closing the end in a "thumbnail" shape is better than being open.

Thanks for any explanations...
 

4130biker

PM me about Tantrum Cycles!
May 24, 2007
3,882
447
I have to agree on not making the gussets have sharp points as well. The HAZ of the welds on CrMo has a fatigue strength that is about 30-40% less than that of the rest of the tube. Hence why it is beneficial to disperse any loads in the HAZ as evenly as possible. Sharp points like that generally concentrate loads.
I think that answers my question about the thumbnail shape :)
 
Feb 14, 2011
18
0
Hmm, how to tactfully disagree with people on the internet...;)

The purpose of the pointy bit on the gussets is to prevent stress hot spots. As the material tapers away to a thin point gradually it means the change in stiffness is gradual and the load is distributed over a large area. Think of it like this, put a thin strip of wood in a vice and pull on the end. It snaps at the vice (head tube). Get a long strip of wood and a short strip (the gusset). Put it in the vice together and pull on the long bit. The long bit snaps at the end of the gusset/short bit. Now sand the short strip so it goes from thick to thin. Put the thick end in the vice next to another long strip. Pull on the strip and they bend together nicely and take more force to break. Hope that makes sense, basically thin gusset end bends with the tube, wide gusset end makes the tube bend around it.

AFAIK HAZs are used when referring to heat treated metals. They refer to how the extra strength gained from the heat treatment can be lost when an uncontrolled heating process (such as welding) is used. This mainly affects the yield strength of the material.

The fatigue is more complicated. Fatigue design codes assign experimentally determined allowable stresses against different features (such as welds, parent metal or holes). The lower allowable for say welds against parent metal is the result of the local geometry and imperfections caused by the welding process.

I don't really know anything about pushbike tubes so could all be wrong! I know mainly about linear static analysis of steel and aluminium (including heat treated aluminium) structures against BS 7608/BS 8118/BS EN1993/BS EN 1999.
 
Last edited:

Slater

Monkey
Oct 10, 2007
378
0
AFAIK HAZs are used when referring to heat treated Aluminium. They refer to how the extra strength gained from the heat treatment can be lost when an uncontrolled heating process (such as welding) is used. This mainly affects the yield strength of the material and not the fatigue.
Every weld has a HAZ, regardless of material. Any part of the material that exceeded the temperature at which the metal's grain structure is free to change, is the HAZ. In Steel, the HAZ on a CrMo tube junction approximately .8 to .9mm thick with appropriate heat will extend about 10-15mm from the weld bead.
 

Pslide

Turbo Monkey
Alright, in the spirit of design by committee, here is Revision 1 to my sketch.

I've changed to thumbnail (copyright buildyourown) gussets, although not all may agree. I've reinforced the tube where forces from the end of the thumbnail gussets will be greater with the plate gusset, which should counter the effect of having the gusset cross over highest stress parts of the tube (top and bottom). As for the plate itself, I have elminated the sharp edges and added longer section at the ends so that the loads at the ends are distributed over a larger area.

 
Feb 14, 2011
18
0
Just curious here (since I think doing the gussets with silver is a cool idea). Could using silver on the entirity of gusset option A create a problem similiar to what is mentioned by Chequebookracer?:
I don't think it would as soldering is like gluing so does not effect the fatigue classification. Unless the heat in soldering changes the material properties in some way. I haven't ever done analysis of soldering so this is a guess.
 
Feb 14, 2011
18
0
Every weld has a HAZ, regardless of material. Any part of the material that exceeded the temperature at which the metal's grain structure is free to change, is the HAZ. In Steel, the HAZ on a CrMo tube junction approximately .8 to .9mm thick with appropriate heat will extend about 10-15mm from the weld bead.
You beat my edit - so a quick backtrack from me! I don't believe that HAZs are important when looking at materials that are not heat treated as the properties stay the same. I certainly don't ever use it when looking at normal steel like S355. We don't use any fancy steels in my industry and the design codes I use don't discuss it so can't comment on its effect on fatigue so maybe I should butt out and leave it to the pushbike experts! For heat treated aluminum the fatigue allowables are no different to non-heat treated aluminium.

Final edit before dinner is late and the wife shouts at me. What I'm trying to say is that the proof/resistance to extreme loads side of things (heat affected zones etc) is separate from the crack growth/fatigue cyclical loading side of things and require separate analysis.

It doesn't really matter, I've owned bikes that have clearly been designed by FEA (Sunday) and ones that are clearly designed by testing and common sense (any Specialized or Orange) both are valid solutions. I think Brant Richards intimated in a article about the new Nukeproof that he'd just made it look sensible and weigh about the right weight without including any poor features, bet its fine with his experience of making bikes.
 
Last edited:

Pslide

Turbo Monkey
It's interesting that the Superco Silencer is probably the closest existing frame in terms of materials used and abuse expected to what I would be building, and they've chosen an entirely different gusset arrangement.



The gusset under the top tube seems a bit odd to me, but I think I have seen frames cracked there.
 

w00dy

In heaven there is no beer
Jun 18, 2004
3,417
51
that's why we drink it here
It's interesting that the Superco Silencer is probably the closest existing frame in terms of materials used and abuse expected to what I would be building, and they've chosen an entirely different gusset arrangement.



The gusset under the top tube seems a bit odd to me, but I think I have seen frames cracked there.
That's a new one on me too. I own a bike he welded. It has the standard top and bottom welds.