forging die compression rings

[tbaran]

I am a ME doing die design for impact extrusion/cold forging tooling for aluminum. I am working on a secondary die for an impact extruded 1" square can that forms a shoulder on the closed end of the part, it is a square shoulder that requires approximately 32 tons to form. The die is a piece of D2 54-56RC 2.375" OD x 1.375" thick with a 1" square wire EDM'd inside of it with a shoulder milled into it that is 1.050" square and .250" deep. There is a .138R around the 1" square and a .070R around the 1.050" shoulder.

I am the 2nd engineer on this project and the die was initially compressed by .002" with an H13 40-45RC ring that only had a .437" wall and it cracked. I am redesigning this to have a larger compression ring with a compression of .003"/1" of diameter, which comes to about .007". The new ring will be H13 45-50RC and 6" OD. For a square die does this seem sufficient or should I be looking at multiple compression rings and if yes what materials and sizes for them? Actually any information on multiple compression rings would be appreciated.

 

[tbaran]

What I meant was that the D2 die with the square cut-out in it cracked not the original small compression ring.

 

[TVP]

tbaran,

I will take a look at this in more detail over the weekend, but if you want references for calculation methods and equations related to the use of multiple stress rings, then you should investigate the following:

ICFG Doc. 5/82 Calculation Methods for Cold Forging Tools

http://www.lft.uni-erlangen.de/SEITEN/ICFG/document.htm#jubilee

Quickcalc software

http://www.nationalmachinery.com/english/quikcalc.pdf

The ICFG document is 14 pages and thoroughly covers the calculations necessary for optimizing a multiple stress ring arrangement. It is basically an extension of Lame's theory for thick-walled hollow cylinders. The Quickcalc software is based on the same equations and allows the user to quickly review various designs.

Finite element simulation would probably be necessary to truly understand the stresses in this type of asymetric die, but I am guessing that you don't have access to this. You may try discussing this with a university professor like Joseph Domblesky (Marquette University) as part of a student project/case study.

 

[tripleZ]

What are your internal corner radii, and where did the die actually crack at? This could be a ring compression issue, geometry related, or could be caused by improper manufacturing of the die assembly (i.e. a crap bore job & a crap insert grind can result in a crap shrink fit and premature failure in shouldered dies).

If possible, can you post a picture of the failure area?

 

[Screwman]

We have always used 300M for our compression rings to reduce the number of cracked components. It seems to provide additional toughness and reduced notch sensativity. When aged to it's upper limits it also gives very good yield strength.
TripleZ has a great point; any kind of stess riser on the OD or ID will set up a crack site. Make sure the finish is flawless.

 

[Pressed]

Moving to a larger case OD was a step in the right direction, for sure. The calculations will tell you if it was enough.

Did the cracking occur at asembly or how many parts were formed? Do you know what the internal forging pressure is?



==========================================
Business Card

http://mech.e.tripod.com

__________________________________________
Cycle Heaven.......

http://www.tailofthedragon.com

 

[tbaran]

TVP, great info thanks for the links I will need to order those documents soon. We have done some finite element simulation using Deform software, but it was only to look for shear in material flow. Every fall Deform offers a Die Process Modeling Workshop at Marquette University in Milwaukee, Wisconsin, and I really need to attend it.

triple2, the corner radii are as I mentioned above, a .138R around the 1" square and a .070R around the 1.050" shoulder. This is where the cracking occurred, it was diagonally across the square and initiated at one of the corner radii. You're absolutely correct about the stress risers and when I pick up the new die Wed. AM I am really going to check the material finish in the corners and look for stress risers.

Pressed, the cracking occurred after about 100 parts were formed. I do not know the internal forging pressures. What I do know is that it takes the 32 tons across the 1" square area and the material is 6061 aluminum in the F temper.

The new D2 die is being EDM'd outside and it is coming in Wed then we will shrink fit it into the original .437" thin walled case with a .002-003" press fit, I need to do this because of the overall die configuration, and then the larger 6" diameter compression ring with a .009" compression will be shrunk fit over the whole thing. I tried to allow for the compression over the final sizing of D2 die I just hope the added compression doesn't crack it at the corners.

 

[TVP]

I am still working on some calculations (regular job just keeps getting in the way), but your mention of EDM leads me to another question: what happens to the die after it has been machined using EDM? There is a 10-15 micrometer (0.0004-0.0006 inches) white layer that forms during EDM that absolutely must be removed. This is usually done by polishing (manual or ultrasonic) the die. The fatigue life will be excessively diminished if the white layer is not removed.

 

[tbaran]

It is being polished by the EDM'er, but you can bet I'm going to check every square inch of this die when it comes in tomorrow morning. Thanks.

 

[TVP]

The surface roughness for a highly stressed die like this should be Ra < 0.1 micrometers, especially in the corner. This is difficult to do consistently, so you should definitely inspect all the corners with some type of magnification if possible.

 

[tripleZ]

Now the corner radii make a bit more sense. We just use different jargon. Looking at the basic design and what you've posted, here's a couple of things that you may want to run through your FEA software and/or look for:

1) Shrink method. Is this a blind shrink fit where the insert is shrunk into a pocket, a straight-thru sleeve (unlikely I'd guess), or an insert with a stepped OD and a stepped bore in the case? If you have the choice to use the stepped version vs. the blind version, go with the stepped design. The blind hole shrink fit will usually put more compression on the bottom, creating more tension at the top, and making it more likely to crack. This will not be as evident in a stepped OD. The stepped OD will also be effective for constraining the insert under press and during ejection.

2) Finish is important. You will want to look in 3 areas especially. The finish in the vertical choke corners (R.070), the finish in the thru corners (R.138), and the finish at the bottom of the EDM'ed land, primarily in the corner radius that exists there (you didn't mention one). If that transition is sharp, then you also have a likelihood of developing a crack. Use a loupe (7X mag will work) and a soft light. Throw the die on its side, put the light at the back end, and peer down the cavity. EDM should pop out(as will wire lines, etc.). Your EDM shop may rough polish the recast, but he also has to take out enough stock to remove the remnant pitting below the recast layer.

3) Too late now, but when making dies, I usually rough cut the insert out of the die, shrink fit the insert into the case, and then finish wirecut the cavity. My finish cuts will take around 0.030" out of the cavity in multiple passes. It's two setups, but I rarely crack inserts. Carbide occasionally does, but it's rare.

4) Insert material. I'm in a different industry, but we do deal with pressure cracking of material fairly often. Uddeholm make two steels you might consider trying: Vanadis 4 and Isodur K340. One is PM, the other is hot rolled I believe. The grades are very similar in performance, but the K340, due to its non-PM construction, seems to hold up to impact loads a bit better. You should also find that this steel wears better than the D2 you're using. It's a little more expensive, but the price will likely be negligible if measured in tens of thousands of parts.

5) To me, .009 seems a bit much. You'll probably find out in a bit.

 

[Pressed]

Your pressure is not extreme but that square shape is of a primary concern.

If your journey into the the suggestions don't solve the issue then you might consider sectionalizing the insert. Reducing the hardness of the insert is one cheaper/simpler test to try.



==========================================
Business Card

http://mech.e.tripod.com

__________________________________________
Cycle Heaven.......

http://www.tailofthedragon.com