Small Fasteners in Aluminum Threads

[Synesthesia]

I often go through the following design process and have been looking for ways to improve the performance of my designs. Perhaps you folks might have some insight!

I have two aluminum parts that are mated together through the use of screws and tapped holes. I am trying to make a design that is as small/light as possible while also withstanding high shear loads. Often I also have geometric constraints that limit the size of fasteners. So the problem I wind up with is how many 2-56 screws do I need to take this load? I usually put a comfortable factor of safety on the screws failing in shear and go ahead and hand tighten the screws. Often this works but sometimes I break screws while tightening and sometimes my design is over-built.

So I thought ah I should pull out my Machinery's Handbook and finally build a nice chart for estimating shear friction and tightening torque. I did this and it looks great but the problem that I'm running into is that tightening torques are very small and very sensitive to friction coefficients. A) Torque wrenches seem to be a poor (or very expensive) choice for these low torques and B) if I put what I think is a reasonable factor of safety on my torque due to friction uncertainty, it looks like I will have to over-build everything!

Anyway, any suggestions on how I can improve my process here? Thank you.

 

[MintJulep]

Step 1 is to quantify the problem to be solved.

How small is small?

How light is light?

How many pounds or Newtons is a "high shear load"?

 

[KENAT]

Using torgue methods with small fasteners is problematic because of the small scale. The difference between under torqued and over torqued (sheared head/deformed drive) is very small - especially given the variability of friction and the limited fidelity of most torque wrenches etc. I've previously worked on projects where we couldn't even get it to work on M4's which are bigger than you're talking about.

I'll be interested to see what others have to say, we too use a lot of small fasteners though mostly without much loading so 'finger tight' with some thread lock if required seems to work.

One thing, for low profile where csk can't be easily used (due to fixed-fixed tolerance issues etc.) then button heads offer the next lowest profile option. Sadly, the hex drives are also small and tend to get stripped on smaller fasteners but hey - you can't have everything. I can't remember if you can get torx or similar instead or how much better they might be.

Posting guidelines faq731-376

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

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[Synesthesia]

Well I'm intentionally trying to keep the problem general. The bottom line is that the designs I have usually are small/light enough that I'm using #2 screws and the loads are high enough that the screws are in the realm of failing due to shear stress/overtightening.

I think the problem is significantly easier when the screws in question are larger. Its generally easier to avoid over-tightening screws and less costly from a mass/size standpoint to overbuild things a bit.

 

[Synesthesia]

Thank you for the input KENAT, these are the sort of issues I'm getting at. I don't run into screw head issues too often, usually I have room for a socket hex head and usually I break the screw before the hex head strips.

Maybe I should give up on the idea that I can utilize the shear friction but I too am hopeful about what others have to say.

 

[MintJulep]

Well, you've summed up essence of the problem in your first post.

The fastening process is dominated by variables that you have very little control over.

The classic way to reduce scatter in the torque vs. preload characteristic of a joint is to lubricate the threads and underhead areas. That may or may not be tolerable for your general part, and doesn't address the inability to accurately measure very small values of torque.

If it has to be able to come apart I would think about a mechanical key to handle the shear loads.

If it doesn't need to come apart, adhesives or interference fit.

 

[KENAT]

Synesthesia, that's why I mentioned about button heads. Perhaps you can substitute #4 button heads for #2 cap screws and so on if it's a 'head height' issue on the clearance. Still too small to use torque but might help in other ways. If the head height isn't a driving factor then it's obviously no help.

Mint makes a good point about separating the shear loading from the fastening (and potentially location control).

Posting guidelines faq731-376

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

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[MikeHalloran]

Maybe you're using the wrong wrenches.
Get a 'torque watch' of appropriate range and try using that.

Or eliminate the screws, by subbing a v-band clamp, or a peripheral crimp, or a ductile shear ring.



Mike Halloran
Pembroke Pines, FL, USA

 

[drawoh]

Synesthesia,

Are your screws actually failing. We use a lot of small screws where I work. They seem to work.

We use stainless steel screws, mostly. These are ductile, and withstand quite a lot of over-torquing. My screw spreadsheets are based mostly on 90% yield stress.

An alternative is to specify torque angle. I have questioned this in past threads, but I have since analysed it. You can install a 2-56UNC screw finger tight, then turn it an additional 90[°]. The screw will be plastically deformed, but otherwise, intact and clamping. The limitation is that you cannot install the screw an infinite number of times. Probably, you should discard each screw you remove. So far, nobody I work with likes this idea.

JHG

 

[KENAT]

We sometimes use the "finger tight +1/4 turn" approach but as 'finger tight' itself isn't a very robust definition I'm always a little dubious.

Posting guidelines faq731-376

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

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[TVP]

Some things to consider:

- Use thread forming screws instead of machine screws

- Use pre-applied thread adhesive with either of the above

- Internal hexalobular (Torx) drives are routinely used for small screws like M1.6 for cell phones and other small devices.


The following links shows some examples:

http://www.sfsintec.biz/internet/sfsinten.nsf/PageID/REMFORM_F_2

http://www.sfsintec.biz/internet/sf...C125784C004F9500/$FILE/El_comp_overview_4.jpg

 

[drawoh]

KENAT,

My spreadsheet shows that a 2-56UNC screw clamping 1/4", is re-usable if you can finger tight + 5[°]. As you note, this is hopeless. You need a gross angle to account for variations in finger tight.

A 1/2 turn is only around 3.5% plastic strain. Soft stainless steel screws can manage this, once.

JHG

 

[racookpe1978]

Seems you've got 3 problems, and solutions to the three are mixing with one another so one solution is making the other 1 or 2 or all three problems worse.

How often do you separate (disassemble) the component, and who does this? You (the professional shop), the customer - on a regular basis like to change a battery that MUST be customer-changeable? Very seldom - and only if it breaks? Often - like if to make a regular adjustment or calibration - where the case MSUT fit back on in order to make the instrument work?

What I've thinkning is that you should try to separate the shear resistance from the screw (over-tightening/thread ripout-and-failure/screw-too-loose) problems caused by trying to accurately and repeatedly "torque" a very, very small screw.

Can you use two pins to catch the shear? The case (the upper part) then need only a minor force to be held on. The larger diameter pins would stop the shear loads. (Fit and hole tolerance need be addressed - maybe by tapered pins larger than the current 2-56 screws.)

Can a locktite plastic hold the screws from coming out and so the torque doesn't need to be as strictly controlled?

Do your workers know the critical nature of "finger-tight" and do they take it seriously? What training have you given them? What is their feedback? Are you using magnetic tip screwdrivers to make assembly easier and more reliable? (The whole operation might be more like watchmaking though - done on a one-of-a-kind basis by experts....)

 

[dinjin]

I would be curious as to how you drill these small tapped holes in aluminum and do the holes shrink after drilling as some of the literature suggests. How close are these holes toleranced and held prior to tapping?

 

[drawoh]

racookpe1978,

#2 screws are not as convenient to install as #8 screws, but they are not particularly difficult. I am a big guy, so most assemblers should have an easier time than I do.

I cannot recall me or anyone else ever breaking any of the stainless steel screws that we use. Either the OP is using brittle material, or he and his operators are ham fisted.

Again, I suggest stainless steel. It is ductile, so lots of stuff happens before the screw actually breaks. The material is non-magnetic, but for a lot of applications, a hex socket head cap screw will hang on your hex key, anyway.

JHG

 

[KENAT]

We don't have our own machine shop so I don't really know much about the process but we end up with quite a few small threads into aluminum without much trouble that I'm aware of.

I had one al part with 4-40 screws that never seemed to be properly free running. They claimed it was gunk from the masking during anodize but I was never convinced. I ended up having to clean up the threads on several of them and it appeared to me to be metallic swarf/chips coming out. Perhaps this was an instance of what you mention dinjin but it was definitely an exception.

Posting guidelines faq731-376

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

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[Synesthesia]

Thank you for all the great comments folks. You really have to appreciate the diversity of experience here! Some context: The systems I'm working on are prototypes and time/money is not unlimited so I like to be able to undo connections to swap in new parts, replace broken parts, etc. That said, fasteners are disposable. A professional machine shop does all the manufacturing, I personally assemble everything. I generally try to avoid high loads on screws in shear by use of keys, interference fits, etc. but oftentimes this is not possible for practical reasons.

Some thoughts:

Overtightening -- Many comments about ductility of my fasteners, this is a good point. Generally I use high-strength hardened fasteners as I am concerned about them failing in shear if shear friction is insufficient. I have found that these are easy to break with a standard long-handled allen wrench in a socket cap head. It's possible that in the cases where I've broken screws they have bottomed out in blind tapped holes, most of my holes are blind. In any event, I should consider more ductile fasteners in the future.

Pins -- I have had trouble combining screws and pins. A recent assembly I took apart had 4 intact stainless steel screws and 3 broken hardened pins. The connection did not fail, it was being taken apart for maintenence. This was surprising, my guess is that this was due to ductility/uneven loading issues (pins were press fit in one part, slip fit in the other) but I am still a bit puzzled on this one. Currently using 7 hardened screws in this assembly and so far so good (designed for 8--huge FoS--but I broke one during tightening).

Screw preload -- The finger tight + 1/4 turn approach sounds like it is worth a try, considering the difficulty in measuring torque and the uncertainty of friction (lubrication is a good idea but I think friction might still be a challenge at this scale). Is there a good reference for more details on this technique? I don't recall seeing anything in my Machinery's Handbook.

@TVP: Your point about thread forming screws is interesting. What is the advantage there?

@dinjin: Holes are simply drilled with the appropriately sized tap drill and then tapped. Material is often 7075-T6. I've never checked my shop's tolerances.

Sorry if I missed anyone's comments, lots to think about.

 

[KENAT]

Synesthesia, I'm afraid the 'finger tight + 1/4 turn' approach is just a slight refinement of the '2 white knuckles' approach of old.

Posting guidelines faq731-376

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

(probably not aimed specifically at you)
What is Engineering anyway: faq1088-1484​

 

[Synesthesia]

Well some other folks mentioned spreadsheets so it sounds like at least some degree of calculating can be done.

 

[drawoh]

...

Overtightening -- Many comments about ductility of my fasteners, this is a good point. Generally I use high-strength hardened fasteners as I am concerned about them failing in shear if shear friction is insufficient. I have found that these are easy to break with a standard long-handled allen wrench in a socket cap head. It's possible that in the cases where I've broken screws they have bottomed out in blind tapped holes, most of my holes are blind. In any event, I should consider more ductile fasteners in the future.

My old Machinery's Handbook claims that mechanics can break any screw up to 5/8". This section has been re-written in my new one. I suspect that a brittle screw transitions too abruptly from tight enough to too tight. According to my spreadsheet ([σ]_y=55ksi) a 2-56UNC screw can clamp 180lb. How much clamping force do you need?

Definitely read your machine design book and prepare a spreadsheet.

JHG