Ford Upgrade Kit Halfshaft Bolts Found Loose


HighHP

GT Owner
Jun 3, 2019
434
Spokane, WA
This topic has quite a history on this forum, mostly on the OEM halfshaft bolts and Ford’s first fix. But this involves loosening of the Ford Upgrade Bolt kit, which as I understand is Ford’s second and final fix.

I had a leaking output shaft seal and decided to replace both sides. Pretty sure it was the seal and not the halfshaft/flange joint, since the oil was leaking down the transaxle housing from the seal. It was not leaking off the heat shield below the halfshaft. I also cleaned everything and the joints stayed dry, the seal still leaked oil.

Before I removed the halfshafts, I measured the radial play to be 0.010” on the passenger side and 0.015” on the driver side. After installation of the ARP kit below, the radial play was 0.002” – 0.004” on both sides. This was measured on the halfshaft side of the halfshaft/flange mating surface.

Upon disassembly, I found that the two bolts holding the splined flange to the output shaft were loose on both sides. They were positioned to the point where they would be if they were hand tightened. The bolts had, what looked like, factory applied blue loctite. This kit was originally installed at a Ford dealership on April 2008 with 8,200 miles on the car. The car now has 20,000 miles. It is possible, although unlikely, that these were loosely installed and then neglected to be torqued. My assumption is the bolts loosened soon after they were installed; maybe they would have stayed like this forever with no problem?????

Here is a theory on what happened with these bolts:
  • The splined flange is a snug hand slip fit on the output shaft. As this is not a press fit, a micro amount of rotational clearance/play exists between the flange and shaft.
  • The two small bolts do not provide adequate clamping force/friction to keep the flange and shaft from rotating this micro amount of clearance on each other under significant torque. Similar flange/yoke assemblies use large nuts torqued to 150 lb-ft or more that will friction lock the flange and shaft together, except possibly for that first high torque launch which will set the splines tight to each other.
  • So, with these small bolts, each time the axle experiences significant torque flip-flop back and forth, the flange rotates on the shaft the micro amount of rotational clearance.
  • The washer under the two bolts stays tight to the flange and rotates with the flange the micro amount of clearance relative to the shaft.
  • The bolts are threaded in the shaft, so the washer moves a rotational micro amount under the bolt heads each time the axle torque flip-flops.
  • This rotational micro movement of the washer under the bolts slowly loosens the two bolts over time.
  • This rotational bolt loosening continues until the washer is loose enough under the bolts that the washer no longer applies loosening torque on the bolts. Thus, the bolts loosen to about hand tight. The factory applied loctite will then hold them in this position, hopefully.
I used a tap to clean out the output shaft bolt holes. The tap came out with a minimal amount of debris on all four holes. I squirted lacquer thinner in the holes and ran the tap again, used air to blow out the thinner. Once again, I squirted lacquer thinner in the holes and blew them out with air.

I installed the ARP bolt and washer kit sold by Accufab. My feeling is that the ARP is superior to the Ford kit. I know the alloy and specs of the ARP bolts, which doesn’t really make them superior, but I know what I have. AND, the ARP bolts are drilled for safety wire. I did safety wire tie the two bolts and I also used blue loctite for overkill. The safety wire should hold the bolts from rotating. I also used Permatex Gear Oil RTV Gasket Maker (the green tube) to seal the flange/shaft/washer spline area, per Ford TSB 06-23-8. I even safety wire tied the six bolts holding the halfshaft to the flange. Another case of overkill, but I am retired, have time and I needed to freshen up my wire tie skills.

The bolt kit pic attached, shows the Ford update kit that I removed. The other two attached pics show the installed wire tie drilled ARP bolts. Per ARP tech info, the ARP recommend torque applies stress to their bolts to 75% of yield, so these bolts are not torque-to-yield and therefore should be good to reuse. The Accufab instructions recommended torque is equal to the ARP recommended torque for these bolts. Actually, the two small bolts are torqued somewhat less than 75% torque level. You can see the gray Permatex Gear Oil RTV sealer squeezing out around the circumference of the washer.

Hope this is helpful to someone. I read where another forum member, in the long past, had this same issue twice with the Ford upgrade bolt kit. I did not see what the member did for a successful fix. Anyone else ever hear of the Ford Upgrade Kit bolts loosening? Or, any problems with the ARP kit?
 

Attachments

  • Ford Bolt Kit.jpg
    Ford Bolt Kit.jpg
    1 MB · Views: 75
  • Two Flange Bolts.jpg
    Two Flange Bolts.jpg
    1.2 MB · Views: 78
  • Halfshaft Bolts.jpg
    Halfshaft Bolts.jpg
    992.8 KB · Views: 74
Last edited:

GTMD

GT Owner
Mark II Lifetime
Jan 4, 2011
957
NorCal
Looks great and what an outstanding write-up! Thanks for posting.
 

cobra498

GT Owner
Jul 14, 2010
310
Central Ca;ifornia
I have had half shaft bolt problems for years on my 1966 Cobra. I used grade 8 aircraft spruce parts for many years but they still would stretch and loosen after a relatively short number of cycles. I had to replace the bolts 3-4 times per year until two years ago. I found some aerospace grade bolts in AN7 size that came with certification papers and had a small radius under the head that required a washer with a matching cut to the ID. I have had the same bolts in for two years now and they are maintaining torque. I think you are correct in your diagnosis on bolts. I also had an issue with steering arm bolts having tensile failures and switched to ARP stainless 160,000 psi bolts, no more issues. Their products are also very good.
 

HighHP

GT Owner
Jun 3, 2019
434
Spokane, WA
Wow. That sounds similar. Interesting this is happening on other diffs. For comparison: Grade 8 is considered high strenght at 120,000 psi yield, 150,000 psi tensile. The two small ARP bolts (ARP 2000) are 200,000 psi yield, 220,000 psi tensile. Hopefully, they last and the wire tie holds.
 
Last edited:

Indy GT

Yea, I got one...too
Mark IV Lifetime
Jan 14, 2006
2,526
Greenwood, IN
I agree with GTMD, very nice and detailed writeup.

Yes Ford did go thru several iterations on getting this issue fixed. It was difficult to determine the root cause of the bolt failures (Hydrogen embrittlement due to the electro plate applied to the bolt surface). But the final fix appears to have solved the problem.

I understand the theory which you put forth as the cause of the bolt loosening and it may or may not be valid. I do not know. But there have been very few complaints on this topic after installing the last Ford bolt fix or the AccuFab ARP bolts. Both these solutions fixed the root cause failure issue.

And I understand the reliance on ARP bolts as being good (they certainly are), but either the Ford supplied bolt kit or the ARP bolts will equally fix the problem. The problem is not a bolt strength issue, thus if you use a 80 Ksi YS bolt vs. a 150 Ksi YS bolt, it does not matter. Incidentally preloading a bolt to 75% of YS is way under torqued. Typically you want the very highest preload the material can give you without yielding. My experience is to use a torque which puts the fastener tensile area to 90% YS as recommended by J. E. Shigley, "Mechanical Engineering Design". Torque-to-yield or torque plus turn-of-nut tightening procedures are certainly another topic, and very interesting.
 
  • Like
Reactions: PL510*Jeff

HighHP

GT Owner
Jun 3, 2019
434
Spokane, WA
Indy GT - This is really kicking the dust out of the back areas of my brain, so the following may be completely incorrect. The tighter a bolt is tightened or stressed, the lower chances are that it will be exposed to cyclic load fluctuations. Cyclic load exerted on a bolt is not good. The material/parts that the bolt holds together experience the load fluctuations, not the bolt. The bolt clamp force should exceed the cyclic load, thus no load cycles on the bolt. So.....what you note above is ideally the best as defined by standard design criteria. However, this situation, defined by material failure, falls outside of standard design criteria. So the fixes are a result of back-up design criteria - make it as big and strong as limitations allow. You note that this back-up design fix is working, good to know.
I measured my radial deflection before and after to see if that might be a way to determine if the two bolts are loosening, without removing the halfshaft. In my case, it does seem to be a good indicator. I am just rambling, hope I am not wasting forum space. I could go on and on and..........
 

Indy GT

Yea, I got one...too
Mark IV Lifetime
Jan 14, 2006
2,526
Greenwood, IN
Hi HighHP

Your “rambling” is quite good and the back areas of your brain remember quite well…:)

“The tighter a bolt is tightened or stressed, the lower chances are that it will be exposed to cyclic load fluctuations.” Absolutely CORRECT! That is why you want to tighten the fastener to the highest preload (90% YS) you can possibly get from the bolt material.

“Cyclic load exerted on a bolt is not good. The material/parts that the bolt holds together experience the load fluctuations, not the bolt.” CORRECT as well. Bolt threads, by their geometric nature, have high stress concentrations in the bolt head fillet and thread fillets. You do not want these features exposed to alternating (fatigue) loads because fatigue cracks can nucleate at these high stress locations. As Shigley states in his text book, “The importance of preloading of bolts cannot be overestimated. A high preload improves both the fatigue resistance of a bolted connection and the locking effect.”

“However, this situation, defined by material failure, falls outside of standard design criteria.” Not quite sure what you mean with this statement. But the Ford engineers did design the bolted connection correctly, used appropriately sized bolts for the clamp load (which are actually quite small, thus super high strength material fasteners are not necessary) and I am sure were assembled correctly at the (Riccaro/Ford) assembly line. That is why it took the failure analysis team quite some time to figure out why the bolts were failing as all the mechanics of the fastener design looked correct. They finally determined the failure root cause was the plating applied to the sourced bolts which caused hydrogen embrittlement of the bolt geometry.

Basically for general understanding, hydrogen embrittlement of a fastener-

"When high strength steel is tensile stressed, as is the case with a high strength fastener that is under tensile load from tightening, the stress causes atomic hydrogen within the steel to diffuse (move) to the location of greatest stress (e.g., at the first engaged thread or at the fillet radius under the head of a bolt). As increasingly higher concentrations of hydrogen collect at this location, steel that is normally ductile gradually becomes brittle. Eventually, the concentration of stress and hydrogen in one location causes a hydrogen induced (brittle) microcrack. The brittle microcrack continues to grow as hydrogen moves to follow the tip of the progressing crack, until the fastener is overloaded and finally ruptures."
 

nota4re

GT Owner
Mark IV Lifetime
Le Mans 2010 Supporter
Feb 15, 2006
4,176
While the bolts may have failed because of hydrogen embrittlement, the compounding error was that the thick spacer under the bolts actually cupped - becoming concave on the side of the bolt heads. This in turn placed uneven loads on the embrittled bolt heads - "have high stress concentrations in the bolt head fillet and thread fillets" and this would cause them to snap. Unfortuately there is a small gap between the flange and the flat end surface of the splined output shaft. The cupping of the ~1/4" spacer was caused because of this gap. The "fix" in both the case of Ford and Accufab kits was to utilize a spacer ~ twice the thickness of the original along with high-grade bolts.

The "cupping" of the original spacer was true on 100% of the GT's we disassembled. It is very obvious to the naked eye - and certainly visible if you place the spacers on a flat surface.
 
  • Like
Reactions: analogdesigner

HighHP

GT Owner
Jun 3, 2019
434
Spokane, WA
Ahhh.....love reading all of this, very informative, thanks very much. Stress concentrations, stress risers, v-notches, sharp corners, you probably remember seeing this illustrated by refracted light in those flat plate plexiglass models. Hydrogen embrittlement, hydrogen in solution, welders are very aware of this problem. I did not know that this could happen as described above. Metalurgy - interesting things happen in metals. I just had to research this and found this: "As the strength of steels increases, the susceptibility to hydrogen embrittlement increases. In high-strength steels, anything above a hardness of HRC 32 may be susceptible to early hydrogen cracking after plating processes that introduce hydrogen." Apparently, this is a well known problem in high strength steels. That must be why most high strength fasteners/parts are not plated.
So what you guys are saying, is that with the thicker maybe harder washer keeping the bolt load even around the circumference of the bolt head along with correct bolt metalurgy, the problem is solved.
However, that still leaves the micro-rotation effect. I assume that this still takes place with the washer harmlessly slipping under the bolt head periodically. So, as long as the bolt is held from rotating, things should remain secure and tight. Comment??
I recall reading a post by "nota4re" back in July about another bolt failure. Just reviewed it, that was a sloppy install issue. Like most accidents, multiple errors led to that problem. Not cleaning the hole, failure to see/feel the improper torque feedback with a bottomed bolt, and probably more.
Once again, thanks very much for the explanation of things. Very enjoyable and informative.
 

PeteK

GT Owner
Mark II Lifetime
Apr 18, 2014
2,267
Kalama, Free part of WA State
Indy wrote: "Typically you want the very highest preload the material can give you without yielding. My experience is to use a torque which puts the fastener tensile area to 90% YS as recommended by J. E. Shigley, "Mechanical Engineering Design".

Ahhh, good ole Shigley. I have his 3rd edition of Mechanical Design still sitting on my bookshelf. Yup, your and HighHP's comments accord with Shigley's text. Not only surface plating, but other surface treatments that use acid-based techniques will introduce hydrogen to the steel (hydrogen ions define an acid), and therefore should not be used.

I still wonder why Ford/Ricardo elected to use two side-by-side bolts on the end of the transaxle splines instead of the more traditional one big nut or bolt on the end of the spline. All my other cars with transaxles have used the one big nut arrangement, and I've never experienced or read about problems with that design.
 

HighHP

GT Owner
Jun 3, 2019
434
Spokane, WA
Geesh....Even Boeing installs faulty parts that were plated resulting in hydrogen embrittlement.

 

Rick Mustang

GT Owner
Mar 29, 2016
102
Central Texas
The Accufab kit shows out of stock. Anyone know another source or do they replenish their supply? Thanks
 

nota4re

GT Owner
Mark IV Lifetime
Le Mans 2010 Supporter
Feb 15, 2006
4,176
Accufab told me that it would be WEEKS, not days before they had inventory. Until then, Ford is the only source. They seem to be readily available.
 

HighHP

GT Owner
Jun 3, 2019
434
Spokane, WA
Google the part number: gtrab

You will find other other suppliers that show it.

Here are a couple others, hopefully, they don't all drop ship from Accufab.


 

Rick Mustang

GT Owner
Mar 29, 2016
102
Central Texas
Lethal Performance does not stock. They obtain from Accufab when ordered.
 

BIGFOOT

GT Owner
Jan 18, 2012
739
Northeast
Has anyone else had this issue after the Ford fix?