Overcoming residual stress in Aluminum

Came across this interesting paper on the subject:

https://prod-ng.sandia.gov/techlib-noauth/access-control.cgi/2007/076811.pdf

One take-away for me was: Use 6061-T651 rather than T6 and, for most applications, this should take care of it.

In looking-up details on T651 I learned it is mechanically stretched by 1% to 3% in order to reduce (or was it equalize?) stress.

Wondering if anyone might be able to share experience comparing T6 vs. T651 and other horror stories about parts curling-up due to internal stress.

As a side note for anyone not familiar with how stress is introduced into aluminum...

During manufacturing the hot aluminum is quenched (I think the two options are oil or water). On a thick profile or plate, the outer surface cools first while the inside remains warm and gooey. The outer surface will compress a bit due to the change in temperature and squeeze the core. As the core cools it will try to compress as well. However, the outer layers are no longer gooey, which means they just say "nope" and don't give. This sets up stresses within the resulting product. When you machine aluminum with this built in stress the part can deform post-machining due to, for example, having faced-off the top "skin" that was stressed one way while the core is wanting to do something else. This paper discusses various strategies for stress removal. My take was that the simplest approach if you are worried about this is to use de-stressed T651 and life is good (or better).

Trying to understand this because I need to machine some rails for a linear motion application and it would be great if they stayed straight after machining.

I have put parts is a self cleaning oven to stress relieve before cutting in two

Disclaimer: I'm not a metallurgist, but I've specified heat treatment of some intermediate (3-5") section 6061 forgings and plate products on several occasions.

Page 29 and 30 of the following book, Aluminum and Aluminum Alloys Handbook, gives a good description of the various heat treat states and what the numbers mean: Aluminum and Aluminum Alloys - Google Books

I've never heard of heat treating 6061 to a T6 with an oil quench. The only quenches I've heard of are dunking in water or a spray quench (this is how bike frames are done to reduce quench-induced distortion). I have vague memory that there are also some glycol quenches for 6061. I would be surprised if an oil quench achieved fast enough cooling rates.

T651, T6511 and/or T652 can be difficult to find in really heavy sections because there just aren't many presses around that can stretch or compress the really heavy sections. But I'm talking about things like 12" thick plate, which when I ordered I was only able to find in T6 for that reason.

Sorry I can't help with offering guidance on internal stresses in machined 6061, but in my experience most lighter extruded 6061 products are already T651 or T6511. It's rare for me to see 2" extruded bar for example that isn't at least T651 if not T6511.

If you need high stability and low internal stress, 6061-T6XXX may not be the right material for your application. There are products like MIC-6 which are designed explicitly to have low internal stresses, presumably at the cost of reduced mechanical properties (UTS, ductility, etc). For most linear motion applications, mechanical strength is far secondary to precision and stability, so MIC-6 is a superior choice to a solution treated and aged alloy like 6061-T6.

I'd rather use 2024 or 7075 than crap 6061...never have issues like that....unless it needs to be welded.