I only did this once so don't consider this expert advice. Heat the part to
the point that it is no longer attracted to a magnet, then quench in oil
till cool. Then heat to a much lower temperature in the range of 350 to 600
Deg. F for long enough to soak the complete part. A lower temperature will
provide a harder but more brittle part, and of course if you heat it high
enough you anneal it and are back where you started. I was making a wrench
and used 400 Deg F in my wife's gas cook stove oven. Other will likely
provide a better description and more precise temperatures.
the trick is to get the metal to the right temp and have a accurate way
of measuring it. A thermocouple gauge is used in most ovens. an old
time way was to heat it until a magnet will not stick to it and then
quench it in warm oil. you must be able to completely submerge the
piece or you will not quench it properly and it will catch on fire. YOu
need at least a gallon of oil for every pound of metal to handle the
heat. also you should either agitate the piece in the oil or have a
circulator for the oil around the piece.
you now have to temper the piece (draw it back)... that is to remove
some of the hardness and make the metal tough. for this you need a chart
for the temperature to bring it up to for the hardness you want. then
you let it cool and if possible do a hardness test on it.
Be very careful _where_ you do that, of you do. One of the case studies
that was mentioned in on of my materials science courses was a brake or
wheel part which was hardness tested to verify heat-treat - they started
failing, and the failures were traced back to cracks starting in the
divot where the hardness tester had been used.
A sway bar is going to be a stinker to do in a home shop. 3'
long, several bends, possibly tube/bar combination. You need a
furnace big enough to hold the entire part at 1700 degrees or so,
then drop into an oil bath. Bar parts should be dropped into the
bath end first to minimize distortion and anomilies from one side
of the round to the other. So you will need something like a 30
gallon barrel on it's side for a furnace with a propane burner
plus another 30 gallon barrel with your quench medium.
Your heat treat specs are at http://www.matweb.com/index.asp?ckck=1
key in '4130' select the size and specs you need.
You might also want to look at 4140. Similar properties, a bit
more forgiving on the heat treat but typically wants an oil quench.
Be sure to design the bar with multiple holes in the end so you
can select a range of firmness without doing a new bar.
Other posters mentioned the safety issue: 4130 can be pushed to
some truely phenominal streght numbers. I've done tubing
assemblies over 200,000psi tensile with reasonable ductility. The
downside is that 4130 has a fairly narrow band between tensile
and yield. If the heat treat is not dead on, it tends to snap
with no warning. Not to mention the issues of surface finish,
stress risers, nicks and scratches, and warpage.
John Olson wrote:
Whether you are interested in racing cars or not, buy and read
(several times) Carroll Smith's "Engineer to Win". It contains the
best layman's introduction to metals and metallurgy that I have ever
read. When you are done you will know how to heat treat 4130, whether
you need to do so, and if so whether to do it yourself or to have it
done professionally- But you may also decide that you don't need to
use 4130 for your purposes.
Yeah, I was going to point out that 4130 may be a waste of money and time,
and heat treating it may waste more of each. I remember learning about sway
bar design and materials back when I was involved in amateur racing but,
unfortunately, I have a crappy memory. <g>
However, it would surprise me if a sway bar designed for a high-performance
car would even approach the elastic limit of untreated, cold-rolled steel.
It shouldn't deflect a great deal. Much more likely to need fancy steel and
heat treatment is an ordinary street sedan, in which the bar would be less
stiff, and therefore would need more elasticity.
In any case, the difference in elastic limit between the normalized
condition and the hardened condition is not that great for 4130. It's fairly
strong as it is, and it isn't tremendously strong even when heat treated.
The modulus of elasticity is the same whether it is heat treated
or not. ie for a force of x, it will deflect y. The question is:
does the deflection make it exceed the yield strength? Annealed
4130 is about half again stronger (higher yield strength) than
equivilent cold rolled mild steel. (80-90kpsi versus 50-60kpsi)
The heat treated version of 4130 can easily push past 200kpsi or
more than double its unheateated version.
I agree that a flat track car might get along with a non heat
treated bar, but an off road racing vehicle needs all it can get
from the sway bar.
The modulus is, but the elastic *limit* depends on its yield strength. In
fact, the elastic limit IS the yield strength.
This is a field in which terms always get misconscrewed. By more
"elasticity" I mean the ability to deflect farther without exceeding the
That's what I said. <g>
I don't think you want to make a spring -- or an anti-roll bar -- out of 200
kpsi 4130. If it's going to fail, you want ductile failure, and the
elongation at 200 kpsi strength is close to zilch.
The bigger point, though, is this: If you're loading 4130 to, say, 150 kpsi,
you're getting a lot of spring deflection at that load. In a
high-performance car you don't want a lot of spring in your anti-roll bar,
so you make it of a larger section or you use shorter arms. Of course, by
using shorter arms you also demand more twisting displacement from the bar
for a given amount of travel, so you could wind up loading the bar quite
heavily toward its ultimate strength if you compensate for a
smaller-diameter bar by using shorter torque arms.
The ideal is the longest practical arms. For the kind of deflection you want
in a race car or high-performance car, that means you want a bar of larger
diameter, which does less twisting in use. That keeps the specific torsional
load (the load per unit of metal) relatively low; low enough, IIRC, that
there should be no benefit whatsoever from using a high-strength bar.
Well that's a new one. Are we talking about
gas regulators here??
================================================= please reply to:
JRR(zero) at yktvmv (dot) vnet (dot) ibm (dot) com
Looks like deliberate wordsmithy to me. Misconstrue something and
you certainly can end up getting screwed in the process...
I like using "automagically" for something that normally isn't
automatically operated, but you rigged it up just for kicks and the
"Wow Factor". And sometimes it even went into production - The
original Ford Fairlane Retractable Hardtop from the 60's certainly
qualified... (When it worked.)
--<< Bruce >>--
Bruce L. Bergman, Woodland Hills (Los Angeles) CA - Desktop
Electrician for Westend Electric - CA726700
If you would substitute the word 'deflection' into your arguments
in place of elasticity, it would make more sense. If you do, I
As for the fatigue resistance, ductility arguments, take a look
at these frames:
4130 tube, heat treated to the 200kpsi range on tensile, 12 %
elongation in 2" (BETTER ductility than 6061-T6 !!!) These have
no springs, the driver just bounces along on a rough dirt track,
letting the frame flex. They have 'mixups' occasionally, not good
for horses, drivers, or frames. Typical sulky might get used for
1000 races (1250 miles) of race conditions. Never see any fatigue
issues. One got bent 90 degrees when a semi backed over it, still
didn't break. But the heat treating on these is precise. Pull
coupons attached to each batch.
I would agree that the need for a full high strength sway bar is
debatable in standard practice. But a couple things to consider:
heat treated parts tend to be much more uniform in
characteristics. I would much rather have a 4130 bar treated to a
low value than an equivilent CR bar if I wanted an accurate point
of stress beyond yield. The other thing is if the bar might need
to take a huge overload, say from a baja buggy landing on one
corner in a movie style launch. I have a picture sequence on a
buggy launching upward, 5' of air, slowly going nose down and
tilting off to one side, and coming down on one front tire. Next
pic in the sequence is the tire going a rather unusual direction.
Ed Huntress wrote:
Nice frames! Are you making them? Have you been at it for a while? If so, do
the names Rich Muntz or Charlie Kobash mean anything to you? They were
pretty well-known harness-race drivers on the East Coast. They were my
All good reasons to use 4130. Its impact strength, fatigue strength, and
general toughness are all excellent, and it can produce excellent properties
even when you're dealing with a lot of welded joints, thanks to the low
That's what it's for. Specifically, the alloy was developed in the 1920s for
aircraft applications that required those mechanical properties, plus easy,
reliable weldability. Airplanes were both gas-welded and stick-welded (yes,
stick) from 4130 tube through the 1930s.
Well, you're getting some interesting properties from your 4130. Getting
back to this application, the 12% elongation you're getting (are you
actually testing it for elongation?) is what SAE claims for 4130 hardened to
a yield strength of between 125 and 130 ksi, not 200 ksi. And the comparable
figure (13% elongation) for AISI 1060, for example, occurs at a heat-treated
yield strength of 112 ksi. Even 1040, which is a lot easier to get than
1060, comes close enough that it probably would do the job.
That's why I say it's unlikely you'd need or even notice the advantages of
4130, if you're going to heat treat it, in a race car or other
high-performance car. Ironically, the more demanding the racing, the less
strength you need in your anti-roll bar, because you're going to design it
for more than normal stiffness. Sizing it for greater stiffness results in
lower specific torsional loads on the bar. It's unlikely, say, in a road
racing application, that you'd even come close to challenging the steel in
that bar in terms of its yield strength. If you did, it would mean that it's
producing relatively low roll resistance.
Now, your point about off-road vehicles is a good one, and it's also true
that a racing anti-roll bar for a production car may put you in a bind for
space, and you have to use relatively short arms. In that case the bar has
to twist more for a given amount of suspension travel (actually, for a given
amount of differential displacement for the left- and right-side wheels, but
I hate talking like that <g>), and that could dictate the use of a
high-strength anti-roll bar.
It's like a lot of issues in racing: You're usually after stiffness rather
than strength, and, if you have adequate stiffness, you aren't even
beginning to load the part close to its yield strength. Tubular space frames
made for racing perform equally well if they're made from 1020 as from 4130.
Lotus, Porsche, and many other race cars of the 1960s proved the point
beyond any doubt. If the frame is properly designed, it never approaches the
yield strength of the material. It's all about stiffness, and that's
virtually the same for all grades of steel.
If it's a Baja buggy, then use what you have to. I don't even remember what
the subject was here <g>, so I don't know if that's the case or not.
But let's assume that the subject is what's in the header: Heat treating
4130 steel at home. My first response is, no way, Hose-A. Unless you have
some really terrific facilities, this is one steel you don't want to try
heat treating. It's right on the edge of air-hardening in terms of its
quench rate, and annealing, for example, requires ramping down the
temperature at 50 deg. F per hour. If you overquench 4130 from the
transformation temperature, you wind up with a disaster waiting to happen --
if it doesn't split lengthwise while you're looking at it.
It is almost never heat treated in normal transportation applications:
aircraft, cars, motorcycles. The bicycle guys sometimes order it heat
treated to greater hardness, but that's an exception. Then they silver-braze
or bronze-braze it into a frame, creating a structure of questionable
overall integrity. The tubes resist permanent bending better but the joints
can be really screwy.
Neither is it usually used in the annealed condition, as you suggested in
your last post. By far the most common heat-treat state of 4130 tube and bar
is normalized: somewhat stronger than annealed, but not as strong as
heat-hardened. In that state it's easy to work, it has close to 100 ksi
yield strength (in practice; official minumums are lower), and it has
terrific elongation -- over 25%. It's a very ductile material, easy to weld,
and it's reliable in a welded structure. It's actually tougher, with much
greater impact strength, than an identical structure made of annealed 1020.
The margin of strength between normalized 4130 and the same steel heat
treated to an acceptable level of ductility (elongation), IMO, is maybe 30%
or so. That may make a difference in the Baja buggy but it shouldn't make a
difference in most applications. As I said, you shouldn't be anywhere near
its yield strength in actual practice.
What you get by leaving it normalized is a rather large cost savings and a
lot less complication, unless you're buddies with someone who knows what
he's doing and who has a heat treating oven long enough to fit your bar. If
you try to heat treat it at home, I'll give you dollars to donuts that the
end result will be *worse* than if you just left the steel in its normalized
And if you really need heat treated 4130 for your anti-roll bar, next time
do a better design job. <g>
(remove "3" from email address for email reply)
The reason I heat-treated the sway bars I built for my car is that the
suspension travel is fairly great. Mild steel can yield in the application
(think 1960's F-1 cars). Modern cars have little suspension travel, so no
heat treat is required. The big down-side to heat-treating a long, skinny
bar on 4130 is that the damn stuff warps, and once it's hardened it's damn
hard to straighten again. The Carroll Smith edict to always heat treat your
bars was written when race cars had lots of wheel travel and the bars were
typically long U shaped things. May not be applicable now, except at the
pointy end of the racing hierarchy.
Polytechforum.com is a website by engineers for engineers. It is not affiliated with any of manufacturers or vendors discussed here.
All logos and trade names are the property of their respective owners.