4130 Steel, Aluminium, and as welded strength and fatigue

I’m not trying to start an “aluminum vs steel” conversation. This isn’t about comfort or weight. Instead, I’m curious about as welded strength, heat treating, and fatigue. I’ve been doing a deep dive into materials and have a couple of questions for people more experienced than I.

Specifically, lets talk about welding.

When you weld either steel or aluminum, you have a heat affected zone (HAZ). From what I can find, the best estimate of the strength of the heat affected zone is to use the yield strength of the annealed condition of that alloy.

In the case of 6061, this means a T6 tempered part drops from 40ksi (276 MPa) to around 8ksi (55 MPa) for yield strength (MatWeb - The Online Materials Information Resource) in the HAZ. You can re-heat and strengthen 6061, but the process involves impractical DIY heating and quenching (for a discussion of that in detail, see thread at Aluminum Thread - #6 by ben.land101)) and creates distortion. The only alternative, as best I can tell, is to design the frame under the assumption that it only has the strength of “as-welded” aluminum, or around 8ksi, which essentially means using lots of aluminum and making the frame heavy to decrease forces on each part. Couple this with the fact that there is no fatigue strength or infinite life for aluminum, and you have to design significantly under the yield strength to avoid short-life cracking.

Now let’s turn to 4130 steel. It is stronger and heavier than aluminum. It gets affected by HAZ, but best I can tell even annealed 4130 has ~52+ ksi (360 MPa). (I’m unclear on whether you need to post heat-treat or not, but for thin tubes air cooling may be sufficient, or just annealing with a torch around the location of the weld.) Add to that the fact that 4130 has potentially “infinite” fatigue life below its fatigue strength limit, and it gets much easier to consider working with steel. Altogether, if you ignore the weight, it is more durable (doesn’t fatigue), repairable, and approachable (no heat treating) for the DIYer, albeit more expensive and harder to machine.

With that background, two categories of questions:

  1. Aluminum: What do you all do with your aluminum frames? Do you not worry about fatigue cracks? Do you just check your frame over now and again? Do you use heavy aluminum? Do you go for 7005 or 7020 Al that home oven ages to 40ksi (if so, what’s your supplier?)

  2. Steel: Do you use steel because of fatigue life concerns? Simply because it is stronger? Because it is easier to work with? Where do you find good pricing?

I appreciate your time and expertise and helping me find the way through the maze,


These are great questions, and I look forward to the discussion! I know that steel always feels like the “safer” option. Aluminum feels like steel did 15 years ago where it almost seems like a “dark art” where it’s hard to readily find all of the info needed to do it yourself as a DIY person without overbuilding things. Perhaps these photos I got from the aluminum frame builders FB group can contribute to the discussion:

1 Like

It’s a great question. Most framebuilders who work with aluminum use 7005 aluminum because it does not need heat treatment:

7005 frames are just to be air-cooled and aligned right after welding; solution heat-treatment is not necessary. For full suggested strength (T6: UTS 380 MPa), artificial age 7005 frames to 6 hours at 200°F (+/-10°F) plus 4 hours at 320°F (+/-10°F). www.fairing.com

The required strength of a frame is also subjective. For mountain bikes, many custom builders use 9-6-9 butts without concern. In Asia, they use 1.4-.7-1 butts to pass testing. It also depends on how hard the bike is going to be ridden.

None of the 90’s road and mountain bikes would pass today’s testing. But none of those bikes can be ridden as hard as a modern bike.

Annealing is a very controlled process that raises the metal to a specific temperature and brings it back to room temp slowly over time (on the scale of hours). Using a torch will make the steel microstructures more angry.

Steel is always going to have a HAZ. This is why frames almost always crack 2-3mm after the weld.

Steel, Titanium, and Aluminum will all crack over time to fatigue. If you overbuild the frame to avoid this, it will ride like trash and weigh a ton.

A huge advantage of steel is the availability of diameters and butting profiles. This allows you to adapt the stiffness and strength of a frame to the individual rider. Aluminum and titanium have much less selection.

The bill of materials (BOM) for a steel frame is pretty cheap, $300-500. The majority of the cost of framebuilding comes from the shop space and tooling. Paint jobs can double the BOM.


Thanks for the responses.

Does anyone have a supplier source for 7005 for a DIYer? If not, it seems like aluminum is out of reach without overbuilding since you can’t heat treat after welding?

Why does steel still crack 2-3mm after the weld? Is there not a way to relieve the HAZ stress? Any way to make cracking there less of an issue? I’d heard of aluminum cracking, but not steel.

Bicycle Fabrication Supply has 7005

Fatigue life is described by an SN curve:


SN curves are derived from perfect machined specimens. The results are a statistical representation because the samples are imperfect. On a bike frame, the HAZ, the joint, and the bead will create a stress riser, which will eventually cause a failure.

I asked some Taiwanese manufacturers why they don’t heat-treat steel frames after welding, and it came down to two things: 1) the frames warped too much after heat treatment 2) it was too expensive.

As a engineering analysis guy, I used to get hung up on the strength and lifecycles of a frame, but these two things help me get over it:

  1. A frame will be too noodly to ride before it is unsafe (70’s steel road bikes are happily rolling around with tiny 25mm tubes)
  2. Everything eventually breaks. If so, either repair it, or retire it knowing it had a good life!

I’m less worried about failure than catastrophic or sudden failure. In other words I’d prefer the frame fail in a way that gives some warning so you can repair or retire not you-had-a-front-fork-and-now-you-don’t.

But yes, easy to do a deep dive into analysis when oftentimes it’s probably fine.

No, that’s too pessimistic an estimate. I’m much less familiar with alu but even I know that 6061 will eventually return to T4 with no heat-treat, only natural aging.

Steel at the softest point in the HAZ is not annealed, it is tempered. Tempered in the common parlance (non-engineers) means strengthened but it’s really the opposite, it means weakening. Often intentionally, such as with tool steel that was quenched from red heat to harden, leaving it too brittle. The tempering to a controlled degree gets back some ductility and toughness. You could say the weak point in the HAZ is “over tempered”. It has lost some of the strength the tube gained from cold-working and whatever heat-treat the tube was delivered with. Even non-heat-treated tube is delivered in (usually) the Normalized state, which is still quite a bit stronger than annealed. The tempered zone will be weaker than Normalized, but not as weak as annealed.

You can minimize the strength loss in the tempered zone by reducing the total heat input, and crucially, reducing the time-at-temperature. The processes that weaken the HAZ do not happen instantaneously, they proceed at some rate, and keeping the time to a minimum reduces these effects.

The reason the weakest spot is not at the edge of the weld but rather a few mm away, is the steel that gets above the transformation temperature cools rapidly enough (mostly from self-quench) to end up Normalized. The weak spot is the steel that never got that hot, that only made it into the tempering range of temperature. Note: silver brazing, if not overheated, never gets above the transformation, so all the steel in the joint is tempered, right up to the edge of the braze, not a few mm away. You can see the difference in crashed frames from the location of the buckling.

Mario Emiliani did a good article about this in the old Bike Tech magazine in 1982.

About me: though I have training in mechanical engineering, I never worked as an engineer or metallurgist or welding scientist. My professional experience is just from building bike frames. I hope the real engineers on the forum will correct me if any of the above is wrong.


Thanks. I didn’t know 6061-O turns back to T4 from room temp age hardening. Assuming that is the case and looking up the values from matweb:

6061-T6 (pre-welded state) has a yield strength of 40ksi
6061-O (annealed) has a yield strength of 8ksi
6061-T4 has a yield strength of 21ksi

So, perhaps the most pessimistic method is to design to the strength of 6061-T4 assuming that age hardening happens in a reasonable time at room temp (i.e. days not months or years).

Interestingly enough, from matweb above, while the yield strength of 6061-T4 is only 21ksi vs 40ksi for 6061-T6, the fatigue strength is listed as 14ksi in both cases at 5e8 completely reversing cycles (6061-O has a fatigue strength of only 9ksi). I don’t have full fatigue S-N curves to see how it changes with fewer than 5e8 cycles at stresses greater than 14ksi, but that at least suggests that if you design for stresses to never be greater than 14ksi the bike in a T4 state (without heat treat) should have much the same fatigue life as a bike with 6061 in a T6 state (correct me if I’m wrong).


And how do you propose one do that ? Which tool allows you to do that calculation locally ? My guess is that there is no such tool.

I plan to use a CAD model run through a finite element analysis (FEA) solver. There are quite a few FEA software options, including in Fusion 360 and SolidWorks. Most are expensive. At the moment I’m experimenting with PrePoMax a freeware FEA solver.

How will you simulate the variation in material properties as they transition from base metal to the various region changes in the heat affected zone to the weld metal while incorporating the geometric changes in the joint with this FEA tool ? All of these factors will have a significant effect on meaningful results.

In the case of aluminum, the material itself has the same elasticity across the tempers so they don’t behave differently as far as springing back, they just fail at different amounts of stress. As such, all you have to do is keep the force below 14ksi (6061-T4 and -T6 fatigue strength) in all regions. The FEA output shows a heat-map of sorts to highlight areas receiving lots of stress so you have an idea of where to reinforce as well as regions with very little stress where material can be removed to save weight.

It is true, the added weld material is a different alloy, and it’s strength would have to be found from a datasheet. A quick search suggests ER4043 is probably the TIG rod to use for 6061 and the datasheet gives a yield strength of 10-27.5ksi. I assume the range is because of some aging, but whether that is due to heat treating or natural aging would take more digging as would any figuring out the fatigue strength of the material rather than just the yield strength.

I think in practice, a much thicker weld bead might be used relative to the underlying material thickness which might make up for lesser strength and spread out the stress, but that’s more a guess than engineering…

You might find this helpful: (PDF) The HAZ in Aluminum Welding Revisited (researchgate.net)