Aluminum extrusion shape restrictions? 5

A portion of my job is the design of aluminum extrusions as framing for small signs, typically 6063-T6. What I learned early on is that the process a diemaker or extruder uses to determine the extrudability of a die isn't at all a matter of FEA or stress calculations or anything like that - someone there with years of experience would review the drawing and give it a thumbs-up or thumbs-down, and if I was lucky, some specific feedback as to what was wrong with it may have followed.
It's been a few short years now that I've been periodically working on these things, but some things have still got me wondering:

- The "tongue ratio", shown here, shown toward the bottom of the page. I've typically been given Area / (opening width)² as the guiding equation, and that it must be 3 or less, otherwise there's a chance that the steel "finger" on the die will snap off due to the forces exerted on it during the extrusion process.
However, I've got an extruded heatsink right here in front of me that has some very deep grooves as part of its extruded profile. The fins are 1.28" tall, and the opening averages about 0.090" wide. (It tapers, from 0.100" at the opening, down to 0.080".)
A/d² = 0.1152in² / (0.100in)²
= 11.52

Any of the extruders I've worked with call for this to be 3 or less, yet this heatsink extrusion here has an array of very deep grooves to form its fins.

- Thin-walled extrusions, either hollow or solid: Extruders like thick walls. Accounting and management would prefer that I extrude foil - more product, less material.
For example, I recently was tasked with making a stylized hollow tube. The outside dimensions are approximately 7x4". It took a few weeks, but I finally found an extruder that would build it with a wall of 0.090" thick. Most others wanted anything from 0.125" to 0.154".
The first run they tried jammed in the die, and produced a single piece that looked like a crumpled wet noodle.
They did some re-carving on the die, and ran it again. It worked rather well. One of the dimensions did end up being slightly out of spec, but the specific dimension and the application were not sensitive to this.

So here's what I'm asking then: Are these more a matter of the extruder not wanting to do these extrusions, or is it genuinely a matter of capability? Is it a matter of maintaining tolerances? Are the heatsink manufacturers using a special type of extruder machine, or die assembly, to get their high tongue ratios?



Thank you,
Jeff

 

[1gibson]

If they quoted you X dollars to make it .125" and you said you wanted it thinner so it would be cheaper, why would they bother quoting you a higher price to make it .090"?

Maybe your tolerances are too tight, maybe there are other risks, but they would need to cover risk of scrapping parts, redesigning tooling, lost production time on other jobs, etc. while they figure yours out.

If you really wanted a price I am sure they would give you one, seems like some miscommunication there. It would be higher than a "standard" geometry part that they were comfortable making, so they probably just didn't waste time putting a number to it.

 

[MikeHalloran]

I can answer part of your question(s).

Since the aperture of an extrusion press is limited in size by the tonnage and is typically round, large heat sink extrusions may actually be produced in a curled or 'rolled up' state, such that the major plane would actually be produced by a circle in the die, and the parallel sides of adjacent fins would actually be produced by wedge-shaped fingers.

I.e., it's not uncommon to produce an extrusion in a distorted shape, then uncurl it, straighten and flatten it in the soft state, then heat treat it.



Mike Halloran
Pembroke Pines, FL, USA

 

[GregLocock]

Sounds like you are on the path to learning some useful stuff. Like you my experience with extrusion guys has been of the yes/no variety, and that is a shame. Perhaps if you worked with the best of your current extruders and stopped putting everything out to quote (haha) you might be able to get more of a two way conversation.

Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[ajack1]

The thing with any tooling is it is not an exact science, be that extrusions, castings, mouldings or stampings and whatever simulation software or years of experience is used it is only a best guess.

Certain things ring alarm bells and most companies will shy away from them, even if they are actually possible to make but with a lot of maintenance work/ high scrap rate etc.

If you really want to know what is possible triple your tooling and part cost and work with a company rather than just hope the cheapest company comes good.

The example you give is a good one where most companies wanted thicker walls but one company went for what you wanted after extra work and cost they got something near to what you wanted. As it turned out it was out of limit but that did not matter to you, what would have happened if it did matter? Or you get a new QA guy in who just rejects the parts for being out of limit?

Most companies will quote what they are confident can be done not what the absolute limit of can be done is, unless you want to pay a massive premium.

 

[sp8472]

Maybe some day in the future they'll get the flow properties of aluminum figured out, and it'll be possible to feed a profile to some FEA software that'll then try to build the die from that.


"As it turned out it was out of limit but that did not matter to you, what would have happened if it did matter? Or you get a new QA guy in who just rejects the parts for being out of limit?"

If that dimension did matter, or if it was really off, I would have told them that the die had to be redone to meet the specified tolerances. They had agreed to take the job in the first place, and I don't think it's unreasonable to expect them to hold to the drawing. The same sort of thing's happened (though not often) at my own job: We quote and accept a job, but something shows up that wasn't expected, but we won't go back to the customer and say "This was harder than we thought. Can we have some more money?" We have to eat the cost, and then either learn not to take that kind of job again, or else figure out how to do it better so that we can get a decent profit next time.

It would have then been up to them if they would want to try again to fulfill the order, or else cut their losses, apologize, refund the initial die cost, and then we part ways.
Since that particular design's predecessor was in use for nearly 10 years, this extruder stood to get a steady stream of business from us, and we have thus far ordered more than 5 tons of this extrusion from them in about half a year, and there's no signs of slowing down.

 

[ajack1]

That is exactly my point the cost would be down to them, that is why the other companies probably didn’t touch the job, although it could be argued your limits were wrong if the part can be outside of them and still acceptable.

Asking what is the absolute maximum achievable is like asking what is the absolute maximum MPG you can get from your car? You can always squeeze a little bit more out but the hassle and the fact that everything needs to be perfect to achieve it means it is not worth the effort and risk unless the rewards are substantial.

 

[MintJulep]

Greg makes and excellent point.

Add to that, how much linear quantity are you ordering?

Extrusion suppliers want to minimize their risk.

If a RFQ comes in for a shape they know will be difficult, and it also asks for the minimum run they are likely to no-quote as the hassle of scrap in getting the die right and you rejecting stuff is not worth the small potential profit.

If you order 50 million meters, then expect a call "We want this order, but we'd like to discuss your shape, find out what's really important to you, and offer some suggestions."

 

[DesignerGuy16]

I used to design aluminum extrusion dies, and a lot is kind of industry "thumb in the wind" standards.

Typically tongue depth 4x width starts to bring the price up, there are cases for others that depending on the size of the shape to the size of the tool and how much support the die has, yada yada. Heat sinks were typically 100% non-warranty and in some cases 10-15x more expensive than a standard rectangular bar or round. A lot has to do with needing to protect the tongues from the direct force of the die. To do this they'll typically have to machine a one-off tool and EDM the hole and grooves under a protected area where the pocket will flow from the end of the tongue towards the tips.

The smaller you make it, the harder it will get typically. Your best bet will be to actually talk with the companies extruding what you're designing and get their feedback on best practices, shape restrictions, etc. The easy part honestly was designing for deflection and shrink, the hard part was the thumb in the wind this area needs this much "bearing length" to slow the extrusion down so all areas of the extrusion flow at the same rate.

It's more an artform like typical sand-casting foundries are.

James Spisich
Design Engineer, CSWP

 

[GregLocock]

Thanks James, that was an excellent post. Please post more!



Cheers

Greg Locock


New here? Try reading these, they might help FAQ731-376

http://eng-tips.com/market.cfm?

 

[sp8472]

That is exactly my point the cost would be down to them, that is why the other companies probably didn’t touch the job, although it could be argued your limits were wrong if the part can be outside of them and still acceptable.

The tolerances were in fact their limits. I will explicitly call out tolerances on critical dimensions, primarily if it's not symmetric: -0.002/+0.014, for example.

Besides that, the extruder will produce their own drawing, call out the tolerances they are willing/able to hold, and give the drawing to me for approval. If something is too far out, I will work with them to get it in range, or if needed, modify the design to produce the necessary tolerance.
The tolerance that the extruder gave on this dimension was already fairly wide. While we would certainly prefer that it be within the stated range, it's still a simple matter of cosmetics. Given how this profile will be used, no one is going to be looking at this part of it anyway.


James, thanks for that post.

Add to that, how much linear quantity are you ordering?

Typical batches consist of 80pcs of 240" lengths, which is over 4000lbs. Most that I've seen will do small runs, but there's a significant upcharge to cover the setup expenses; can't blame them for that. 2000-2500lbs is the usual minimum I've encountered

 

[SnTMan]

DesignerGuy16, yes excellent. I find it reassuring that even in an age where computer simulations are often the first resort to simple problems, that these "black arts" still exist.

That's what prevents chimps from becoming engineers

Regards,

Mike

 

[sp8472]

SnTMan: Someone's got to design the hyper-intelligent FEA systems of the future, at least until they're smart enough to design their own successors.

 

[SnTMan]

Jeff7, yeah, well, let's don't get started on the software designers, OK?

 

[DesignerGuy16]

There are some things that can be done, but honestly on things like heat sinks I remember hearing of customers spending $15k on a die, only to have it produce only a few hundred pounds before one of the tongues broke.

The same kind of rules with castings kind of apply to extrusions though. You're better off if you can keep as close to a common thickness throughout. If you took something the size of say a 1" x 6" bar and added a 1" tall flange at the ends that was only .03" thick I can guarantee that that flange will tear away from the base because it will be going so much slower than the main section that there isn't enough transition to slow the base material down enough to match it.

Keep it simple stupid should really be the mantra with this stuff.

James Spisich
Design Engineer, CSWP