un-explainable breaking fastener 2

[midsidenode]

Soon after receiving a new batch of custom aerospace screws, I received a complaint from some of our technicians that the screws were failing after torquing and needed some help from engineering to investigate the problem. I thought, no problem. This should be one of the easiest investigations I’ve ever come across – they were probably just over-torquing them and needed someone to re-calc. for a proper pre-load. I expected to be a hero when I left.
After examining the properties of the screw, the clamping surface properties and the torque values that they were using, I was a bit confused as to why the screws were breaking. The torque values seemed to produce about 60-70% of a max. acceptable preload – which was OK since that was reflective of how effective the screws had to be. Nevertheless, this was not the most confusing aspect of the problem – The screws only failed around 20 – 30 minutes after torquing.
I verified the screw’s capabilities by tensile testing several for Fyt and Fut and then torqued many more to failure. The data was what one would expect it to be – given the material and screw size. These screws are 4340 steel heat treated to 160ksi Fut min. Does anyone have an idea of what is going on? Could it be that the screws were not tempered properly after heat treating? Thanks in advance.

 

[MikeHalloran]

A search on "delayed fracture screw 4340" brought up "hydrogen embrittlement".



Mike Halloran
Pembroke Pines, FL, USA

 

[Kwan]

Sounds like embrittlement to me also. Are the screws cad plated? You can send some to a lab to verify the embrittlement. You also need to contact the vendor for the screws.

 

[rb1957]

maybe you've found a batch of bogus fasteners ?

 

[TVP]

This definitely sounds like delayed fracture due to hydrogen embrittlement. You need to have a proper metallurgical evaluation performed to confirm this. This alloy and hardness combination will be very susceptible to hydrogen embrittlement if acid pickled or electroplated. Have you specified any embrittlement relief standards on your part drawing such as ASTM B 850, ISO 9588, SAE/USCAR-5, etc.?

 

[CoryPad]

Delayed fracture of high-strength screws is almost always due to Hydrogen Assisted Cracking (or Hydrogen Embrittlement).

Regards,

Cory

Please see FAQ731-376 for tips on how to make the best use of Eng-Tips Fora.

 

[Screwman]

Ditto on the hydrogen cracking. This sounds like a clasic case.
Make sure that all of the parts form that lot are pulled out of any finished assemblies. They are likely to suffer delayed failures in the future when in service.

 

[WKTaylor]

midsidenode...

The key words You used "new batch of custom aerospace screws" raised the hair on my neck. "Spec parts", [NAS, MS, AS, M, etc] have very rigid fabrication processes and tracking requirements [paper-work out-the-gazoo for each production run. "Custom parts", made to customized specs, for a specific client are another story.

A wise (no-so-old) fastener engineer summarized the business of custom aero-hardware in the following statement.

"You deserve to get what You asked for [IE: on the drawings or spec sheet]. However, if You don't ask for very much, then You have no right to 'expect' very much [high quality and performance for a bargin]."

I suspect that several things MAY have happened, if this is truly custom hardware.

Your drawings/specs lacked clarity. There are very good specs for steel [male] fasteners available that may be liberally referenced for fundamental processing methods, tracking, batch testing, documentation, etc. NAS4002 stands-out immediately for fasteners heat treated to 160-KSI. Also, several other documents deal explicitly with fastener metallurgy, grain flow, etc; while several other documents deal with plating processes for high strength steel parts [note: embrittlement is usually a problem with steel parts heat treated above 160-KSI...You never stated what the heat treat was supposed to be].

So where am I going with this? Your problem description is very "hazy", since the nature [shape, composition, coating, etc] of the "screws" was not defined for us. PS: composition does NOT just mean 4340 steel... it means:

4340 or 4340M [X] steel per AMSxxxx, heat-treated to XXX-to-XXX-KSI per xxxxxxxx, maximum grain size X, inclusions and allowed microstructure anomalies of xxxxxx, forged head per XXXX, rolled threads per XXXXX, rolled root radius of xxxx, ground to a machined finish of XX-RA, Magnetic particle inspected before and after plating per XXXX, grade [or class XX], etc....

So what could have gone wrong???

Machined featured VS forged/rolled features

Improper heat treatment [too high, too low, wrong quenchant, etc]. NOTE: AMS2759/1 suggests that tempering temperatures for 4340 vary, depending on initial quench hardness, etc.

No grain controls [size, impurities, flow, etc].

No batch testing for basic metallurgy or embrittlement. NOTE: if actually accomplished on a random sampling of parts, this embrittlement should have been "immediately obvious" to your processor. NOTE: most specs require a "sampling quantity" be included with each production run... IF You ordered a 1000 parts, then X% should have been added to the lot by the processor for destructive testing... if demanded by QA control requirements.

Embrittlement due to improper plating processes [too thick, too much current, improper plating solutions, too long, etc]. OH yeah was it coated with cadmium, nickel, Ni-Cd, Zn-Ni, Zinc, Silver, etc???

Embrittlement relief bake gone bad? Perhaps... but which way...
(a) Pre-plating stress-relief bake not accomplished [or perhaps not specified]?

(b) No plating stress-relief bake required by spec... but improper heat treatment over strengthened parts to point where it was mandatory??

(c) Too low a bake temperature and/or too-short a bake time to eliminate hydrogen? Note: very thick/dense platings demand longer bake times to force H-atom thru the coating.

(d) Too high bake temperature causing liquid-metal embrittlement [cadmium embrittlement]?

NOTE: IF DONE PER AN ESTABLISHED PROCESS, AND PROPERLY DOCUMENTED your [legitimate] processor/vendor should be forthcoming in helping You analyze the failures and come to a "satisfactory" compensation [warranty] for the entire batch of parts. IF you discover holes in the paper trail, or lack of cooperation on the part of Your vendor, then do as others have suggested... send ~100 off to a reputable metallurgy lab for analysis... and scrap the rest... and notify the FAA.

NOTE: You may discover, brutally, that the vendor/processor did every thing EXACTLY as you specified... and the parts failed. This MAY turn-out to actually be Your problem because of the way You spec'ed-out the fabrication process [refer to the "rule" I stated in my intro, way-above].

Lots of things to consider in this situation…


Regards, Wil Taylor

 

[ewh]

Great eye-opener Wil!

 

[midsidenode]

Thanks to everyone that replied to this message. I feel I should at least give everyone a status.
First of all, the good news is that this is the first time these fasteners have been used so there's no logistics problems associated with pulling them.
It is a custom fastener designed by a well known aero-pace company. The date on the eng. drawing is 1969 and only states 4340 160ksi min Fut and passivate iaw QQ-P-35 - nothing else - very incomplete. The fasteners appear to have been cad plated or zinc plated or something - just by there appearance.
I have requested that our mp group look at the failed parts and see if hydrogen embrittlement can be confirmed. We will subsequently request a revision.

 

[WKTaylor]

midsidenode..

Wise choice... this may be self inflicted damage...

Passivation treatments [per AMS-QQ-P-35 or AMS3700] are intended for CRES [stainless steel] and are generally inadviseable for application on steel due to high acidity [etching] potential... with subsequent hydrogen embrittlement: not to mention providing no ["0"] follow-on protection to thr bare steel.

Note: in my previous statement I also forgot to mention potential for temper damage [embrittlement] due to abusive machining or grinding practices [due to friction induced over-heating].

Regards, Wil Taylor

 

[yates]

If the definition dates from 1969, then it is highly likely that the embrittlement relief was either not specified, or insufficient. Then again if the definition says 'passivate', there would be no embrittlement relief specified anyway !. I concur with all the others - sounds very much like hydrogen embrittlement to me.

 

[Screwman]

There are a fair number of drawings which I see where there is a call-out for "passivate" on a plated part and what the customer is asking for is actually for the part to be yellow chromated! It seems to show up on many parts that originated in the UK and other northern European countries.

Just something to be aware of.

 

[swall]

It would be unlikely that 4340 fasteners actually went through a passivation process, at least a nitric acid passivation, as the attack on carbon steel is very rapid and noticeable. I suspect that the fasteners were plated and never baked for H2 embrittlement relief.

 

[plasgears]

Not keeping the head and nut faces parallel under load can introduce unwanted bending moments and failure. This is particularly true of relatively short fasteners. Consider means of building in tolerance for out/parallel like spherical washer sets, soft washers, etc.

 

[Lurchaolot]

Gentlemen - there may be a very simple answer to this. While there may be some material deficiency with the screws and/or their treatment, the way they are installed & torqued up has a large effect on their installed 'strength'. The torque is usually specified as dry torque, for which typically +/- 30% of the residual theoretical pre-load results, due to the variation of friction in the threads.

However, if the threads are lubricated, there will be an increase of up to a factor of 2 (two) in installed screw/fastener end load for the same applied torque. My 2cents.

 

[Kenneth]

Lurchalot:

Re-read midsidenode's original post:

...The screws ... failed around 20 – 30 minutes after torquing.

A delayed failure mecahanism is at work here, most likely hydrogen embrittlement as suggested above. This is not a "simple" overtorque failure.