[mc-motorsports]

I'm making a billet steel manifold for a diesel engine. It will be machined in 3 pieces, and welded together. It's expected to see temperatures of 2000*F which far exceeds what a normal gasoline engine would see, heat soaked at 2000*F would be a better explaination.

In order to prevent possible cracking, I was thinking that after welding, maybe a trip to the heat treater for annealing and or stress relieving? It's going to be a one off custom part, 1018 mild, TIG welded with 70-s2, final machining of the sealing and mating surfaces after welding and stress relieving.

Any advice is appreciated. I am a certified welder, no problems there, just wondering about the stress relieving, I'd hate to have this thing crack half way through the season.

TIA,
MC

[Geof]

It is not likely to crack but it will almost certainly distort.

I think normalizing is what you are looking for and if you want to be really picky, normalize, rough machine, normalize again and finish machine. But that is probably overkill.

Incidentally are you using hot finished steel?

[mc-motorsports]

Yes, normalizing is the term I'm looking for.

Planning on cold finish 1018, need to use 1.5 or 2" thick, 5" wide which I have in stock. Can't seem to get that size in hotroll, A36. Correct me if I'm wrong, hotroll seems to be under less stress than cold finish. I think I would prefer hotroll for that reason, although cold finish machines much nicer.

Thoughts?

[Geof]

Hot rolled almost no stress and practically no distortion after significant material removal.

It will machine almost the same as cold rolled but the technique is to take good cuts at high SFM and good feed.

Getting 2" x 5" would be difficult; I find hot rolled is not readily available in rectangular shapes and normally resort to chewing down round bar. You would need 6" diameter round bar and a lot of time.

[mc-motorsports]

Yep, I can only get hotroll up to 1" thick at my local supply. Cold roll it is, and normalizing for sure.

Thanks for the input!

[NC Cams]

Anneal makes it dead soft

Normalize offers some residual hardlness which makes post H/T machining easier due to the residual hardnes.

Anneal plus Normalize provides "double draw". THis is often recognized as optimum pre H/T process as the steel is left quite receptive of subsequent H/T's.

1018 is not noted for "hot strentght". This is where a more robust steel, IE 308 or 321 stainless or 409 for lower cost. Repeated heating and coolings will surely reslut on cracking. Long term running will simply tire out the 1018 steel as it has no alloy to provide high temp strength.

If you make the parts out of double drawn material. residual welding stress should be removed after weld and before bolt up by a 1200 deg high temp stress relieve. The part should be subject to heat for at least 1-2 hours. Even high temp cast iron "shrinks" over time. Same thing occurs with mild steel headers.

The Nascar guys don't use mild steel as EGT's there are easily 1700 plus. Mild steel flanges with stainless runner. IF you want it to live, use stainless, especially if it is a turbo engine.

[ImanCarrot]

Would it be possible to use ultrasonic stress releiving/ annealing for this application?

Just that I've heard of it but know very little about it. *curious*

[beone]

If this manifold will actually see 2000 F you don't need to anneal it, it will happen automatically. At that temp 1018 wont last long, its past cherry red.

[mc-motorsports]

SS will take too long to cut and will more than likely crack in a short period of time. It's for a sled pulling truck (turbocharged), probably one of the most abusive applications a manifold can see. It's single turbo so I'm assuming manifold pressure will be 20-30 psi.

Problem with SS is it will tend to move around too much going from a relativly cool temp, to 2200* EGT's in a matter of seconds. SS manifolds and headers don't last long on these apps. Guys that make custom SS headers of log manifolds do it because the SS retains more heat and they are fixable, vs a crack in cast iron can be a nightmare.

Fabricated SS would be cheaper, I'm looking for something that will offer years of racing without requiring repair. Rust won't be an issue, the manifold will be thick enough that it won't rot out any time within it's expected life, and ceramic coatings are out, EGT's are too high. That being said, a nice looking and flowing tubular SS header has it's place.

The only thing the manifold has to do is support a 35 lb turbo and seal off the exhaust between the cylinder head and turbo. I don't think that's asking for much out of 1018, the only thing is has to do is not crack.

The reason for normalizing is to relieve the stress before finish machining to hopefully reduce and distortion and or stress that I am expecting during it's normal cycle, brought to operating temps during a pull and back to ambient. Normalizing will hopefully reduce the stress in cold roll material, and definatly the heat effected zone of the weld, relaxing the metal so after final machining, it finds it's "happy home" and doesn't try to move around so much at high temps.

And steel starts to glow cherry red around 900*F, any turbo motor will make the header or manifold glow.

[HuFlungDung]

Would it be possible to design a manifold with expansion joints? Instead of one large weldment, smaller sections with relatively loose fitting telescopic joints might expand more readily without breaking bolts, etc. A person could even incorporate a small piston ring within the joints to provide a bit of a seal. Of course the thing might warp all to hell anyways, but it might be fun to try.

[Geof]

If you do use cold rolled my normalize, rough machine, weld, normalize, finish machine is probably not overkill.

NC Cams does have a point about C1018 cracking though because it tends to be hot short. Going to an annealed alloy steel, maybe 4340 or 4140, would not make the machining much different and you know about welding so that poses no problem.

[NC Cams]

For linear expansion, a slip lointed log would work. I saw onw that literally slipped together with round wire "c clip" o rings (2( that provideda an effective seal. Was used on BB chev P30 mothr homes and they'd literally melt the manifolds off the engine.

For non metalurgival types, "HOT SHORT means brittle when hot.

Yes, a lot of people make stainless manifolds, but FEW do it right. IF one doesn't PLAN for expansion, it will crack. SS expands MORE Than mild steell on an inch per inch basis. IT also has more HOT STRENGTH which is what you're essentially looking for when you go to the ballistic temps you're running at.

Remember, the stuff expands on an inch per inch basis. IE a 3 inch piece will expand less than a 9 inch piece thus a 3 inch branch might try to constrain a 9 inc runner from moving - this is what encites cracks.

Cost wise, 1018. durability, assuming you PLAN for and accomodate thermal expansion, SS.

CHeck out the basket of snakes headers on the Chapparal 2H. 30 some inches in a small box of snakes all made of "french weld" (pretty sure of the term) SS. No cooling as the thing was aero tight - header breakage was not a problem. As I recall french weld is essentially furnace brazing. THis seemed to be a lot more forgiving of thermal growth than mere tig weld.

BTW, most furnace braxe occurs at 1900-2000 deg in an atmosphere so the proces is slef annealing and "clean" too.

BTW if you don't anneal the welds, ESPECIALLY on chrome moly, you're ASKING for cracks. BTW chrome moly is only marginally better at hot strength than 1018. YOu have to get into martensitic (409) or austhentitic stainless (300 series) to get any appreciable hot strength.

Wjem ,alomg a "cherry red" manifolds, you don't just build for strength, you DESIGN for strength as well - depending on the thermal growth, you may not be able to prevent cracking no matter what you make the manifold out of. The early Jeep I6 tube "header" was ust sort of a creature. The front half expanded axially with growth confined rearward. TH rear half had some growth rearward and some frontward - It netted out zero.

Hence, the header tried to tear itself apart first thru the collector and then back up thru the log to the down branch runners from the port. NOting could fix it except a new bullet that ws doomed to crack as well.

OEM headers go thru this sort of stuff all the time. A lot of cut and try goes into making a "SS" OEM header that lives. Nascar simply makes them and fires new bullets at the car each race.

[mc-motorsports]

Actually SS wouldn't cost as much as I was thinking as far a material. Machine time would probably push the project into the over budjet catagory. I think for SS, tubular is the way to go in terms of cost effective (both time and material cost).

Normalize, machine, weld, normalize and finish machine sounds about right. Tack it together and pre-heat during welding .

As far as designing for expansion... I can only use some common sense and rely on past experience. First problem is the cylinder head flange is going to be the coolest part, it's bolted to a water cooled cylinder head so it gets "heat sinked", so to say. The hottest parts of the manifold will be where the runners turn or radius around to direct the exhaust gasses towards the turbo and the turbo collector/turbo entry point will be the hottest.

I wished there was a good way to possibly test to failure (hopefully the engine requires normal service first, i.e. rebuild) within a relatively short period of time. That's not in the budget either... That would require many hours of dyno time or sled pulling time (which you can rent a private sled), but when you figure all the normal wear and tear of the engine and drive-train, $$$ X $$$ = over budget.

Biggest problem is just to make a couple of these, I'll be working for pennies on the dollar when it's all said and done. For me, it's more about gaining experience to put towards future projects, although it may simply turn into a learning experience, as in "I learned not to do that".

It's really just something that I've wanted to do for about a year now. But the only application that would constitute a billet steel manifold would be a diesel racing app. I'm going to sponsor a local racer this year and in return, he's going to help with the "design", and I use that term loosley, more or less provide the platform and give this manifold hell.

I've fabricated log manifolds, I've fabricated tubular manifolds, now I'm going to machine one from billet! I hope it doesn't turn into a nightmare, but I'm well aware this project may only net another story to tell at coffee time.

[mc-motorsports]

Honestly, the whole reason I posted..., can you guys elaborate on annealing, normalizing, double draw please? I need to have a clue what i'm talking about when I goto the heat treater...:rainfro::withstupi

My total experience of heat treating is handing parts to my boss so he can take them to heat-treat back when I was working in the shops, and making parts for customers who supplied me with pre-hardened material.

Other than that, just the normal pre-heat, post-heat using temple sticks during normal welding opps which won't meet the requirements for this application.

Now that you guys know pretty much exactly what the manufacturing process will be, a quick education in normalizing and annealing would be great, just a push in the right direction. I just need a basic idea so when I get the parts back from heat-treat, I can hopefully get this right the first time.

[NC Cams]

anneal = anneal, see material spec for anneal temp

Normalized = normalize, see material spec for normalized temp and procedure - this can differ per material.

Double draw = anneal first cool to room temp then normalize.

THe above puts material in a highly relazed state and it is quite stable at this point, UNTIL YOU WELD OR MACHINE IT>

ONce you weld or machine, residual stresses are induced that can give you issues later on..

Normalize occurs at a lower temp than anneal and involves natural and some forced cooling via aoir .

After welding and/or machining, we do a "high temp stress relieve" this is at 1200-1250 for 1-2 hours which is BELOW normalize and it takes out the work related stresses. This makes it as relaxed and stable and should not remove the residual hardness/strength generated by normalizing.

You'd be a fool to cut the manifold flange out of billet for a header. 0.375 cold roll would be adequate, Have it laser cut in small sections. As it 2 cyls worth or maybe even one.

As you say, the head is gonna stay put as should anything bolted to it. So then, figure on the runners staying put at the head and moving about everywhere else. CHances are the manifold termination point at the turbo is gonna be the hottest point and therefore the most flexible/compliant. Essentially I'd support it and not let the runners and stuff in between the flange and "collector" not carry any weight but that of themseves.

If you're making a "header" style manifold, that is one thing, If your'e making a billet "log" that's something else. I"m sorta confused as to waht your'e making. I"m thinking/talking "header" style fabrication. THe Chapparal was all french brazed SS and it got HOT and didn't crack - the only place the things was tied together was at head (4 small flanges at port) and at collector - everything inbetween could move as it saw fit.

If you're making a LOG, and it is long, the damn thing is gonna grow with heat. FInd the coefficient for thermal expansion for the materail, figure size change for 1800 deg (typical HOT EGT for Exh vavles) and you now have the amount of expansion factor the manifld will have to endure from cold to hot, all thru the season. If you don't accont for/plan for this, the manifold will try to tear itself apart NO MATTER WHAT YOU MAKE IT OUT OF OR HOW YOU HEAT TREAT IT.

THe beauty of "french brazing" of SS is this. THe heat involved to make the bond is HIGHER Than the typical EGT you should ever see. THus the operating temp should have ninimal effect on the part thereafter. If it survives the weld cool down, it should breezed thru the operating cycle.

Note - check out the expansion coefficients of the materials you plan to use. Mild steel flanges - SS runners/tubing make for strange bedfellows when it comes to headers. Different thermal factors again result in the thing trying to tear itself apaprt due to
different expansion characteristics.

Common sense can be "calculated". Head is say 200 deg F. Manifold goes to 1900 F. This is a delta T of 1700. LEts pull a swag number of 0.000015" per inch/deg F for expansion. Youre actual number may vary. manifold is 24" long.

do the math and you get 0.000015x24x1700 or equals nearly 0.615" growth factor that it is trying to move. We actauly saw car manifolds break the end bolts off at the end ports doe to thermal issues. This is why header gaskets get chewed up, the head stays put an the manifold is moving all over hell's half acre.

A billet, wherein you impinge the exh gasses against walls at turns is going to get/be worse - you'll SEE the hot spots due to gas impingement and these spots will only try to grow WORSE.

The good part about stainless is that it has hot strength the take the heat and not fall appart under the duress. The bad part is that it has a higher coefficient of expansion which makse the thermal growth factor more to reason/deal with.

Lat Time I was at Detroit Diesel, they were making slip joint logs to deal with manifold cracking issue on LONG manifolds or extremely HOT ones. They had round ring slip joint "gaskets" to slide seal as manifold shrunk and grew with heat and cool. Leak? no problems, the diesel soot took care of that. I would not THINK of making a one piece billet for a manifldod that is gonna run "chery red" with EGT's at or above 1700. IMP, yo' realkind to do a lot of maintenance on something like that.
IF you're running the diesel the way I think you are, you'll have PLENTY of soot.

Hint look up and for (stainless steel bellows). These are a godsend in long, high expansion tubular headers.

[NC Cams]

What you're trying to do with billet, these guys already do with castings

http://www.wescast.com/en/products

These guys essentially make "cast headers" that outperform tube parts and don't have the cracking issue of steel weldments. Granted, you can't do something like this but ideas are free.

[mc-motorsports]

We're going to get together on Saturday to discuss it. I may just end up fabricating it. I plan on being very honest and expressing my optomistic thoughts and I'll also relay both your and Geoff's concerns and let the truck owner decide. The last thing I want is something to go wrong half way through a race... If he's up for it, I'm in! But I guess at this point the realization that it isn't really my decision is setting in. If it was my truck, I would just do it. Fingers crossed!

[NC Cams]

As teh "expert", you have more info that can help direct the decisions process.

You already know that HEAT is your major concern. Mild or even ally steel will not hold up to the exhaust heat on short or long term basis - point in favor of SS fab

Thermal expansion is gong to be high based on delta T - point against one piece billet.

Lenght of engine may/may not create more problems with regard to creation of manifold. Long manifold probably shoul have expansion joints will have to machine same for billet, can use "bellows" for tube fab. Spend money either way.

Lot of wasted materail for one billet fab (30-50percent), some wasted material for tube fab - point in favor of tube fab IMO.

If you want to make a limited production run of investement cast pieces, I"d go for the billet machining of the investment cast molds. I'd not be in favor of maknig a billet one off manifold. IT m ight behoove one to discuss thsi in person. YOu alveraly have my contact info.

[davereagan]

If you are reaching 2000 degrees, you are hundreds of degrees above annealing already. Annealing involves very slowly cooling down from 1600 degrees. If you exceed 1300, you've stress relieved it on the engine. The question is how fast you cool it and stress it out all over again. LOL. By the way, you put me out of business on the T3 downpipe flange on Ebay. How do you make them so cheaply? I was cutting them with a Z carb in one pass and for $8.00 after all the fees, they don't make sense.

http://en.wikipedia.org/wiki/Carbon_steel

[NC Cams]

The reason for stress relieving/annealing is simple. The einging will induce its own stresses based upon uneveh heating and cooling and impingment of EGT's upwards of 1700 F.

By annealing/stress relieveing BEFORE you put it on the engine, you DON"T COMPOUND MATTERS. Everytghing is nice and relaxed - theh you give it hell. The weld/fab process - but mostly weld - creates a lot of residual stress in/of/by itself. Add thermal stresses to this and you can generate stresses that can locally yield the already prestressed matl.

THis is a/the reason wehy 4130 airframe welds are annealed after welding.

By doing OFF the engine, you should enhance the chances of the thing NOT trying to tear itself appart when it heats up - and surely it will do so i n an unevern and thus more distressful manner. All in all, it may not make that much difference. BUt at least it gives you a running start.