Normalizes 4130 Frame?

[Joshkb]

Hi all,

I'm Josh, from Boston University Racing.

How many of you treat your frame after welding?

We are using 4130 ChroMoly with ER70-S2. Plenty of sources say that the tubing we use (=<0.095") does not need any post treatment, especially with a filler like ER70-S2. Yet plenty of formula guys, BAJA guys and frame builders recommend an Oxy-Acetylene post treatment into the gamma phase (800-900 c) to prevent brittleness and possible fracturing at the welds. So what do you do, and why? I've heard both sides of the spectrum from a seemingly equal number of people...

Best,
Josh

[jd74914]

We have normalized since 2012. It really doesn't take much time to do, and in our experience really does make a difference on more highly-loaded joints. David Finch of Raetech fame told me once that if you aren't normalizing your 4130 you might as well use 1020 because the brittleness outweighs any ultimate strength gains. I do have 2 anecdotes in support of annealing.

Before 2012 we used 4130 (ER70S2 filler) and did not normalize. In 2012 we started annealing after the car developed stress cracks at HAZ in a few highly loaded areas (namely near damper mounts). We actually tore an 0.065" tube at the HAZ where the right rear damper mount was welded after a good amount of driving. We welded it back up and then proceeded to crack at the left rear damper mount the next time out. How's that for consistency in fatigue life? We used a nearly identical damper mount for years after that and even driving the cars for 100s of hours more have not had any issues after annealing.

As similar trend was noticed with the 4130 control arms. In 2011 or 2012 (don't remember now...it's been too long) we also weren't annealing control arms and hit a curb with a car which resulted in weld cracking. Later arms were always annealed and one poor design even was severely plasticly deformed at the outer ball joint side and did not crack. I always attributed that to being annealed and not as brittle.

In any case, neither of those examples give true scientific evidence to support the need for annealing, but do give it some anecdotal credit. I do feel like it gives you some better margin of safety-especially if you have new welders who tend to overhead metal or weld too fast and hot. As a note, we can anneal all of the nodes in a chassis and suspension in about a night so it's not a huge time investment.

[Joshkb]

Thanks Jim,

Can you share your treatment process? Temp (color), torch, time held at temp, cooling methods etc?

Thanks,
Josh

[Adam Farabaugh]

Not saying this is the way you should do it, but we have never normalized. In fact before last year all of our frames were MIG welded, in house by students who never welded before. Needless to say the welds were awful! We have literally been told by design judges that our welds look like they are going to fall apart... . ... but they don't.

The only time we have ever had a weld/frame structural failure was around a bellcrank tab, brittle failure in the HAZ, but this was because the car was beached onto a curb and then run into a concrete light post base. If you have a light car, I really wouldn't worry about it. Better to get the chassis (and therefore the car) done a few days earlier!

[Joshkb]

I certainly would not call our car light. Not only is it our first car (ever), it is fully electric with very heavy batteries. I can't give you a number because it isn't built, but we are certainly on the heavy side.

I suppose your MIG welds are probably fairly large, which gave you some extra area - either that, or the welds are plenty strong to start with and won't be likely to fair.

In any case, I am at lease interested in what the process would involve, and what people in FSAE do for that.

Josh

[jd74914]

Can you share your treatment process? Temp (color), torch, time held at temp, cooling methods etc?


I'm going off memory here so you'll want to check my hard numbers, but this is our rough process. Hopefully it helps!

There is actually a pretty good instruction set in Carroll Smith's Engineer to Win to normalize.

For equipment, we typically use an oxy-acetylene torch with a rosebud tip. Our steps are as follow:
1) Turn up the room thermostat as high as it will go (to slow cooling to ambient as much as possible).
2) Heat the weld and surrounding HAZ to a very dull red. Its better to go colder than hotter. The dull red corresponds to around 1200F. You can also use temperature pens to get a rough temperature value.
3) Let the weld cool off as slowly as possible. To do this we minimize air movement (no opening doors, etc.) and don't move the frame. Alternatively, you can also insulate your part with mineral wool to really slow down the cooling process. We've never done this, but I don't think it's a bad idea for small welded parts.
4) Repeat for all the welds on your chassis, etc.

[jd74914]

Sorry...one more point.

My terminology above is technically incorrect. You are not normalizing the metal (this takes ~1600F), but actually stress-relieving it. Normalizing requires a furnace, etc. and would need to be done by a heat treater.

Also, here is the excerpt from Engineer to Win:

[Z]

MANY WAYS OF LEARNING WHETHER TO NORMALIZE, OR NOT?
================================================

1. OPINION - Ask or read many peoples' opinions. Follow the majority, or else follow those who are most convincing. ... Or those who shout loudest.

2. VOODOO - Disembowel a chicken, spread its entrails on the table, and study carefully. Consult with voodoo witchdoctor on how best to interpret the "signs". (Note: Wisest witchdoctor = one with the most, or biggest, bones through nose.)

3. ASTROLOGY - Consult the stars. Fortunately, most of this work is already done for you in womens' magazines. Pay special attention to Mars and Jupiter, because they do the blokey stuff, like racecars.

4. WORKSHOP TEST - Use "offcuts" to make up many test pieces of tube welded to tube, or sheet-metal-bracket welded to tube. Leave half of these "as welded" and "normalise/stress-relieve/anneal" the other half. Clamp one side of each test piece in really big vice, and bash the bejeezus out of other side with really big hammer. (DO NOT allow academics in workshop while you do this - they are afraid of hammers!) Note that clever student swinging hammer will soon find the weak point in the design, and so learns how to make a stronger joint.

5. LABORATORY TEST - Make up many similar test pieces to above. Mount test pieces in your school's expensive "Universal Tension Tester", and test to destruction (ie. pull until bits tear or snap). Show fancy Load/Deflection curves to academics. Note that "area under the curve" equates to energy absorbed during test (ie. more area = better, and roughly equates to number of hammer swings in #4 above). Also, look closely at the actual failed bits, and try to decide if it is a "brittle" failure, or else a "plastic" failure of the softest, most annealed part of the HAZ? ;)
~o0o~

In the big scheme of things, the quality of execution of the weld is more important than any post-weld heat-treatment.

But "quality" does NOT just mean "pretty"! A beautiful little ~1.6 mm wide weld on your 1.6 mm thick base metal is likely useless, because not enough penetration! Most welds on typical wall-thickness FSAE frames should appear at least 3 mm wide on the surface. Welds on the Roll-Hoops (~2.5 mm thick) can reasonably appear 6+ mm wide, or more.

Bottom line, get all your welders to do lots of #4 tests above (edit: with successful test = failure AWAY from the weld). Especially so, if they are welding critical parts of the car.

Z

[jd74914]

In the big scheme of things, the quality of execution of the weld is more important than any post-weld heat-treatment.

But "quality" does NOT just mean "pretty"! A beautiful little ~1.6 mm wide weld on your 1.6 mm thick base metal is likely useless, because not enough penetration! Most welds on typical wall-thickness FSAE frames should appear at least 3 mm wide on the surface. Welds on the Roll-Hoops (~2.5 mm thick) can reasonably appear 6+ mm wide, or more.


Good welding is important, but this is really not always realistic given time constraints and student training. It takes a long time to become a good welder (TIG especially), more time than a new team usually has. Additionally, the welds seen in a FSAE chassis are in some ways more difficult than most other tube frames or piping runs due to the frequent deep angles and thickness deltas.

IMO, the biggest reason to stress relieve your chassis is to help poor welds. An overly hot weld which is stress relieved will fare fatigue much better than it will if left in as-welded state. You can test this pretty easy with Z's hammer bend test.