Lower Control Arm Snapped!

[Tangoman]

Here's the thing with aluminum... it doesn't have an endurance limit.

With steel, there is a point where the curve flatlines and no fatigue damage is added with small stress reversals, but with aluminum this isn't the case.

Any stress reversal, no matter the magnitude, will add fatigue damage and the part will eventually fail. So, it doesn't lose tensile strength - to do that it would need to get hot enough to ruin the temper.

This graph shows it very well.
http://en.wikipedia.org/wiki/File:S-N_curves.PNG

 

[sparky]

I updated the previous pic, thank you. I'm still leaning towards the use of non-compliance bushings and over 30,000 miles of street use since they were installed as the culprit.

I don't see street use as a potential contributor. After all what is more gentle on suspension than street use? Track use where the car is fully loaded, with sticky tyres and wheels bouncing off kerbs would be a more likely cause ... but there are many NSX that do exactly this and we don't hear that community saying it's a common issue.

I'd suggest someone damaged the arm in the past through an unusual incident, like sliding sideways into a gutter (true, that could happen in extreme street use!) or perhaps accidentally dragged/scraped it over a sharp edge and started a fatigue point of weakness ?

perhaps you could post a few high res close up images of the broken surfaces?

 

[davidf]

Here's a pic of the broken arm. I'll post close ups of the fracture points tomorrow when I can shoot in direct sunlight. I wonder if my non-compliance bushings might have something to do with this.

View attachment 95358

That closely resembles a chicken wish bone snap. I hope you made a wish.

 

[goldNSX]

I wonder if my non-compliance bushings might have something to do with this.

Nice additional info. Sure it has. Less compliance adds stress on all other parts.

 

[latzke]

Here's the thing with aluminum... it doesn't have an endurance limit.

So if fatigue strength is true to the 10^7 limit...maybe Hugh has been on and off the gas hard 10 million times (about 30 times per mile). Tapping his foot to the Iron Maiden?

 

[Tangoman]

So if fatigue strength is true to the 10^7 limit...maybe Hugh has been on and off the gas hard 10 million times (about 30 times per mile). Tapping his foot to the Iron Maiden?

Ha, no if you had a big drive/brake cycle with a stress reversal of 45ksi, then you would only need to do that around 8000 times before the part failed. Obviously we don't know if that part of the a-arm is stressed up and by how much, but a full brake/drive cycle would push it into tension and then in compression, producing the highest stress range - and that is the main thing that affects fatigue. But, then you add on other damage from other events and you get some rainflow distribution but we would need to know significantly more than we do to figure it out.

Aluminum a-arms will snap eventually, no matter what. Steel wouldn't (assuming no corrosion effects) neither would ductile iron, as long as the part was designed so its max stress reversal was below the endurance limit.

 

[A.S. Motorsport]

I recall there being a portion in the workshop manual on inspection of the aluminum parts.

Update:
Just looked it up it doesn't specify anything further than to inspect for signs of damage or cracks.

Maybe this is why the ball joints aren't available separate so that in cars won't run indefinitely on the same aluminium control arm.
Is there a way to see over-stress or fatigue whilst everything is still assembled?

 

[latzke]

Ha, no if you had a big drive/brake cycle with a stress reversal of 45ksi, then you would only need to do that around 8000 times before the part failed.

Yes, my post was a bit tongue-in-cheek. However, I would assume that Honda in their design work (which included finite element analysis) would have made critical parts like the A arms pretty durable. Probably not 10^7 cycles durable, but also probably significantly more durable than 8000 cycles (I would think closer to the former than the latter).

- - - Updated - - -

That's a very interesting spot to break. Also, it's not the spot you indicated with the red line above.

I updated the previous pic, thank you.

Still not right. Per the picture break is on side with straight-line between frame attachment point and ball joint. The parts illustration with red line shows break on side that is curved between frame attachment point and ball joint. The left/right orientation is off because either the parts diagram is of the other side or your part is upside-down in the photograph (I think the former - that the parts illustration is showing other side).

 

[sparky]

Aluminum a-arms will snap eventually, no matter what.

Here's the thing ... Honda would have designed the arms so that any fatique would not happen in the lifetime of the car. We have one potential data point = one highly modified NSX with > 600whp breaks a control arm after 300,000 miles. That whp is extreme, imparting much higher the ksi peaks than a stock NSX. If and until we see a significant number of failures in cars with normal whp we should not be overly concerned. I for one remain very confident in the durability and design of the A-arms.

 

[Tangoman]

Here's the thing ... Honda would have designed the arms so that any fatique would not happen in the lifetime of the car. We have one potential data point = one highly modified NSX with > 600whp breaks a control arm after 300,000 miles. That whp is extreme, imparting much higher the ksi peaks than a stock NSX. If and until we see a significant number of failures in cars with normal whp we should not be overly concerned. I for one remain very confident in the durability and design of the A-arms.

I completely agree, they definitely did their job. I'm just saying, we'll see these kinds of failures from time to time, and it doesn't mean anything bad about the design or Honda - it just means that aluminum parts fatigue and there isn't crap you can do about it. And yes with that much horsepower and that many miles, I'm surprised it lasted that long - kudos to Honda and their forging supplier.

I like 7075-T6 T651 Aluminum (used in highly stressed structural components, pretty sure it can't be forged though). It was invented by the Japanese in the 1930s (for the Mitsubishi Zero, wikipedia) in an attempt to get rid of this constant fatigue accumulation problem aluminum has. If you look at the S-N curve for that grade, it almost has a real endurance limit like steel. But, it is the most expensive kind grade and temper of aluminum out there. I've designed some parts that use it for that reason because they were absolutely not allowed to fail (expensive application with possible collateral damage). But they were 3' x 2' x 0.5" and they cost 15 grand a pop.

Does anyone make aftermarket a-arms for the NSX? I guess I've never looked...

 

[LJSB]

I completely agree, they definitely did their job. I'm just saying, we'll see these kinds of failures from time to time, and it doesn't mean anything bad about the design or Honda - it just means that aluminum parts fatigue and there isn't crap you can do about it. And yes with that much horsepower and that many miles, I'm surprised it lasted that long - kudos to Honda and their forging supplier.

What a bunch of baloney.
Good grief.

 

[Tangoman]

What a bunch of baloney.
Good grief.

Holy fanboy right? Ha.

 

[LJSB]

Holy fanboy right? Ha.

Not quite. I don't why this happened to Hugh's car. I would need to inspect the car and the part in person and just as important talk with Hugh before I would begin to form a opinion about "how and why". This failure being the result of the normal, expected fatigue of the component would not be something I would decide or even consider without knowing more of the actual circumstance and facts.

 

[vmlinsx]

Interesting!

vmlinsx, how many miles on your car.....and is it turned up to eleven (boosted) like Hugh's?

Brian

In answer to your questions, car is a 1992 with 175,000 miles. Aftermarket turbo installed a few years ago.

I ordered two new lower control arms from Japan. I asked the guys at SoS if they had ever heard of such a failure. They hadn't. My thought was if they haven't heard of it happening, it might be a unique defect.

Was interesting that someone else had one snap. So glad no one else was hurt! I hate to think what would have happened had I been going on the freeway at speed.

 

[goldNSX]

...We have one potential data point = one highly modified NSX with > 600whp breaks a control arm after 300,000 miles. That whp is extreme,...

I forgot those high hp numbers.

For the US guys: I know that I write long posts and they might be boring too but when I did my CTSC I had to do a certification in Switzerland, a very strict one. It's not a simple hop-on like you guys in the US can do with CA legal. It's a complete emissons test ($2k) and a drive-dynamic-test if the car behaves well during fast cornering with abrubtly lifting the throttle and a max-speed-test (no donut-sized brakes). It is expensive but payable and a PITA to do even if you stay below +40% (=380bhp) of the power increase.
Trying to go +40% involves a complete structural test of all suspension parts, yes every single part like bearing, arms etc.
This adds even more tests and I'm pretty sure that trying to 600+ hp in a car that has been built for 274 hp is not doable here in Switzerland.

So, I strongly think that your failure is highly correlated to your high hp/torque, nothing else, not the miles. Cars with very high hp should be aware of this. You might say that high-hp engine are racing for Honda in Japan and all over the world but these cars have not very much in common with a stock car regarding its suspension.

 

[vmlinsx]

I forgot those high hp numbers.

For the US guys: I know that I write long posts and they might be boring too but when I did my CTSC I had to do a certification in Switzerland, a very strict one. It's not a simple hop-on like you guys in the US can do with CA legal. It's a complete emissons test ($2k) and a drive-dynamic-test if the car behaves well during fast cornering with abrubtly lifting the throttle and a max-speed-test (no donut-sized brakes). It is expensive but payable and a PITA to do even if you stay below +40% (=380bhp) of the power increase.
Trying to go +40% involves a complete structural test of all suspension parts, yes every single part like bearing, arms etc.
This adds even more tests and I'm pretty sure that trying to 600+ hp in a car that has been built for 274 hp is not doable here in Switzerland.

So, I strongly think that your failure is highly correlated to your high hp/torque, nothing else, not the miles. Cars with very high hp should be aware of this. You might say that high-hp engine are racing for Honda in Japan and all over the world but these cars have not very much in common with a stock car regarding its suspension.

That's actually a very good point. When I got the turbo kit done on my car, I also had the engine and transmission refreshed, as well as adding a stronger clutch and differential, and better radiator, suspension, and brakes to the car. My philosophy was to over build what I could to add an additional margin of safety to cope with the additional power that my engine is now making.

 

[LJSB]

From first glance(and just a simple understanding of load), it's not even the most load bearing spots of the arm. I'd lean toward an abnormal history/event that assisted in this failure.

This^^^

 

[Hugh]

Sorry for the delay getting high res pics of the fracture points posted. It's been a hectic couple of weeks. Dwight's diagnosis after inspecting the part is that one fracture occurred first and the two broken ends were rubbing against each other and then at some later point in time the other side simply snapped.

Take a look and see for yourselves.....

 

[Midnight_Raven]

From the looks of these pictures you provided it appears to be a fracture caused by shear stess not by tensile or torsional stress. It could be the combination of your non-compliance bushings, aftermarket wheels/tires , and I'm going to assume aftermarket suspension puts more of the load on the control arms vs the stock setup.

 

[Timeless1]

Thanks for posting pics Hugh. Whatever the cause was, you now have given me an idea to reinforced that area.
Time to bring out the Tig welder & Tig up that bad boy & the rest of the control arms as a preventative.

 

[Bravo]

Too bad this is probably too expensive to do.
I know some of these guys and they are very talented.

http://www.imrtest.com/

 

[nigel]

Hugh, Good you didn't wad up the car or do any more damage when the arm failed.

Over at NSXCB, we've spoken about it before, there's a fellow going by the name of "Kaz" that was on the original design team from the beginning !!, when Mr. Honda gave Mr. Uehara the keys to the bank vault to come up with this car. He would be a great contact.
With your mechanical back round, concerning the NSX in particular and the pictures he'd be the best and first one i'd go to with this situation.
He can speak in relevant engineeringese from an NSX design and metallurgical point of view.
Tell him "nigel" told you to contact him.
I'm fairly sure you can only get to him by e-mail or through his "Service Refresh" blog.
He IS the man !
Any help needed get hold of me here or i'm back at the airport, you know where.
All the best.

Cheers
nigel

 

[scorp965]

I recall there being a portion in the workshop manual on inspection of the aluminum parts.

Update:
Just looked it up it doesn't specify anything further than to inspect for signs of damage or cracks.

Maybe this is why the ball joints aren't available separate so that in cars won't run indefinitely on the same aluminium control arm.
Is there a way to see over-stress or fatigue whilst everything is still assembled?

Is there any kind of illustration attached to this inspection reference? Certainly they can't mean to simply shine a flashlight on them and look for cracks?

 

[angus]

Thanks for posting pics Hugh. Whatever the cause was, you now have given me an idea to reinforced that area.
Time to bring out the Tig welder & Tig up that bad boy & the rest of the control arms as a preventative.

I don't think that ids going to work.

 

[Timeless1]

I don't think that ids going to work.

Thinking & trying is two separate experiences, have you try it already to know if it does or not?

2 thick aluminum plates welded on both side should do it (triangle portion). Simple as that & does not need to get any more complicated.