Nissan GT-R LM NISMO (REIB and Welded Steel A-arm)

[Will M]

Was watching this video about Nissan's new FWD car for Le Mans tomorrow and I noticed something odd.

Unless I mistaken it looks like this rather advanced race car has weld steel a-arms with a square cross section and rod ends in bending.
I was very surprised by this as that setup sounds like something we might do for a quick test.
This video may have been recorded a while ago and these components may have since been replaced.

Any way here is the video, you can see the components in the attached photo around the 19 minute mark.
https://youtu.be/fw_2N3tGMEg?t=19m

Thoughts?

-William


(the attached photos are of the same area but at different angles to see around a something partially blocking the view)

[Jonny Rochester]

This "rod ends in bending" seams to be standard practice to build A-arms in supercars and prototypes from before the 60s right to present day it would seam. (And of course hot-rods copy or do whatever they like).

The first pic is a pushrod, the rod end itself is not in bending, but there would be some bending stress in the bracket below it.

The second pic seams to show rod ends to the upright, these would be "REIB" as you point out. But worth noting they don't have heaps of thread showing, and they are not a 5/16" rod end, but much bigger. Also possibly "single shear" bolt attaching rod-end to upright.

This style of connection would work fine in FSAE with ~7/16" rod-ends, and that's what UTAS had in 2001. It didn't fail, but our design was slammed and we redesigned wishbones for 2002 with 7/16" and 5/16" sphericals which where lighter and more elegant, but less adjustable.

[CWA]

The new Nissan car was also 20 seconds a lap off the pace during the last practice run I paid attention to. Causation / coincidence? I know I couldn't possibly say.

[njgedr]

He also said that it was a "show" car.

[JT A.]

The new Nissan car was also 20 seconds a lap off the pace during the last practice run I paid attention to. Causation / coincidence? I know I couldn't possibly say.

Something tells me REIB isn't going to cause a 20 second per lap difference in speed.

[Will M]

The first photo was just to show that they had push rods rather than pull rods meaning that the other rod end was pretty heavily loaded.

My surprise comes not from the added weight which would be fairly small but from the lower strength/reliability.
FSAE courses are nice and flat and there is basically never contact between cars.
While Le Mans is bumpy and contact is fairly common.

So while a rod end could be sized for your expected dynamic loads wouldn't a encapsulated bearing be more robust?
Especially if you go off track or someone bumps you?

Nor would the threads offer easy adjustment during the race.
Maybe on the top a-arm, but why the bottom?
Unless like njgedr said it is only a show car.

-William

[CWA]

Something tells me REIB isn't going to cause a 20 second per lap difference in speed.

Haha, yeah I know bud.

I was trying to imply that the same attitude that lead these designers to fit rod ends in bending may have lead them to make the other poor design decisions which in turn lead to a car completely off the pace. A loose insult perhaps, but it was supposed to be half-funny too. It is Friday after all, and I have beer

[Jonny Rochester]

I think it's worth noting that wishbone/pullrod failure is common in FSAE. And it's not REIB that is causing the failure, but adhesives failing when a carbon part is bonded to a metal part. If the judges can ban REIB (can't remember if its an actual rule or not), then they could just as legitimately ban adhesives in suspension and steering. I think it's better to relax the rules and student designers can make the call.

Auburn failed with this at Michigan.

[Ahmad Rezq]

Jonny
REIB isn't banned in FSAE. Although Judges consider it bad engineering. I agree with those who don't use (Bad Engineering) to describe REIB as long as you know what you are doing.

[Tim.Wright]

I think we need to better clarify what's actually in those pictures. The GT-R LM has pull rod front and rear from what I remember.

Rod ends in bending are like single shear interfaces. They offer advantages in asssembly, manufacturing, tolerance conttol and adjustability in exchange for a small weight penalty.

There's nothing iherently wrong wm, them.

[stever95]

Tim:

I disagree - commercial rod ends are not rated for shear loads. At least not any that I've come across.

You'd have to make your own calcs using measured geometry and material grade assumptions with a rather large factor of safety (stress concentrations around the end of the threads) to justify their use.

Beyond that, yes... nothing inherently wrong, I agree.

[Z]

I reckon the REIB pictures are undoubtedly of a "show" car. Just enough to allow the car to be pushed around for photo shoots, etc. The upright top-BJs visible in other parts of the video are more likely representative of the racecar. Also, pushrods at front, IIRC (?) (Edit: Correction, pullrods at front but with spring-dampers mounted high and horizontal above gearbox, so tall rockers with pivots mid-way up gearbox-sides...)

This latest effort from Ben Bowlby is a cracker of an attempt to exploit a loophole in the rules. The current rules make it difficult to get high levels of efficient downforce on the rear wheels, but the front of car is less restricted. Hence the idea of a predominantly front-drive car, with big front tyres and front biased downforce (~65%F, or more?).

The "20 seconds off the pace" that someone mentioned is what it should be at this stage. This is supposed to be a multi-year effort (2 or 3 years?). If they were front of pack now, then big Rules changes very soon. NO MORE EASY FRONT-DF!.

Z

[Will M]

While the design was very out of the box thinking two of their cars retired and the third was not classified.
http://blackflag.jalopnik.com/whats-with-the-slow-pace-on-the-nissan-gt-r-lm-nismos-a-1711215832

They had a number of issues that lead to the two cars retiring but I suspect that it was because they were not using their first choice of tires that led to them being 20 seconds off pace.

My understanding is they disconnected their hybrid system from the rear axle.
That meant less regen, so bigger brakes, different wheels, and different tires.

Getting any car into LMP1 is an achievement especially since they designed and built everything in 8 months.
I really hope they come back next year!

-William

[Tim.Wright]

I disagree - commercial rod ends are not rated for shear loads. At least not any that I've come across.


This just means that you need to do some of your own testing or analysis if you want to use a rod end in this loadcase.

Very few spherical bearings have published ratings for axial loads but pretty much everyone are loading them that way in their upper and lower ball joints. Why are these applications not raked over the coals?

[Will M]

Tim:

I disagree - commercial rod ends are not rated for shear loads. At least not any that I've come across.




Would these not be what you're looking for?

http://www.mcmaster.com/#rod-end-bearings/=xndf53
(look under 'Thrust-Rated Ball Joint Rod Ends')

-William

[smokebreak]

648

REIB is common place in most racing. However they are much larger than the rod ends commonly used in fsae. Maybe the DJ's are trying to get engineers away from them due to strength or reliability issues. Or they're just trying to spark innovation and grow development in future racing machines. Photo is of LMP2 car after finishing 24hr race.

[Alumni]

648

REIB is common place in most racing. However they are much larger than the rod ends commonly used in fsae. Maybe the DJ's are trying to get engineers away from them due to strength or reliability issues. Or they're just trying to spark innovation and grow development in future racing machines. Photo is of LMP2 car after finishing 24hr race.

First off, where are the REIB in your photo? This has always been a concern of mine since since the design judges have drilled the "NO REIB" into young students minds without a proper explanation of why or what one even is! How many teams this year showed up with a turn-buckle based toe-link system just to avoid having a rod-end they assumed was "in bending?"

Secondly, that's an old photo of a Riley Daytona Prototype. A car built with cost very high on the priority list. So, as far as I'm concerned, while it may not be as elegant or strong, maybe a bit heavier, it was a good engineering decision as it has clearly met the design requirements and function!

P.S. I find it hard to believe that it just finished a race since it is missing the rear anti-roll bar. ;)

[CWA]

Alumni, do you not agree that what is pictured (as I have highlighted for clarity) is a rod end bearing joint?

If you agree that it is, can you explain how this joint in its current application will not be under some form of bending stress during cornering, vertical wheel loading, or at the very least, braking? I will likely remain confused until someone can. Thanks.

[smokebreak]

The picture should have been explanation enough but I'll clarify, the upper a-arm is fitted with a rod end connected to the upright, which is "in bending" from both braking and acceleration forces. Here's a better view of the same cars' transmission and suspension which was replaced in the night. You know because they can repair their cars if they break during a race.

654

P.S. Ask any race engineer and they will explain that in the case of not needing a anti-roll bar one link can simply be disconnected as all the P cars were at this years Rolex 24.


653

Cheers,
John

[CWA]

The picture should have been explanation enough but I'll clarify, the upper a-arm is fitted with a rod end connected to the upright, which is "in bending" from both braking and acceleration forces.

I thought so. But Alumni's comment does not tie in with what the picture shows, with my understanding, nor your comments. Hence my confusion..

[Alumni]

My apologies, I'm an engineer not a writer damn-it!

In my many re-writings of the angry jumbled thoughts in my head when I see something like that I removed my own question of where are the REIB in that photo? Many, many students, including myself at a much younger age were so afraid of rod-ends we assumed any on a control arm are in bending (including the inboards of a non-sprung arm.)

Either way, my main point is and forever will be my second statement as above.

[Alumni]

Also since I am paying more attention now - in regards to the initial topic discussed, while I haven't been able to spend as much time "spying" as I used to , to my knowledge nobody that raced at Le Mans last weekend did so with composite control arms.

The times we've thought somebody was have been few and far between, and never confirmed.

Carbon shrouds a-la IndyCar - yes. But not structural carbon.

[CWA]

I am still confused Alumni, sorry

[APayne_723]

Last time I read through the LMP1 rules I'm pretty sure I saw that the suspension arms are required to be steel.

[Will M]

welded steel a-arms with a square cross section


I pointed out the a-arms not for being steel but for their square shape.
My understanding is that for a given weight round (or oval) tubing would be stiffer and stronger.
And they would not be appreciably more difficult to fabricate.

-William

[Alumni]

I pointed out the a-arms not for being steel but for their square shape.
My understanding is that for a given weight round (or oval) tubing would be stiffer and stronger.
And they would not be appreciably more difficult to fabricate.

-William

Stiffer and stronger in which direction? (That's a rhetorical question for the budding race car engineer to answer for themselves.)

Also once again, I think you'd find the vast majority of cars that ran the LM24 had square lowers and circular uppers, with the occasional aero-tube thrown in the mix. Just something to think about...

[stever95]

Would these not be what you're looking for?

http://www.mcmaster.com/#rod-end-bearings/=xndf53
(look under 'Thrust-Rated Ball Joint Rod Ends')

-William

William - I do not mean "thrust-rated" when I refer to rod ends that are rated for shear loads. I mean a rod end that has been fully threaded into it's a-arm tube housing, used as a ball joint, in accel or braking.

EDIT: Or, if you've got a spring / damper / pushrod / pullrod force putting an a-arm in bending, then you can consider bump forces as well.

[CWA]

William - I do not mean "thrust-rated" when I refer to rod ends that are rated for shear loads. I mean a rod end that has been fully threaded into it's a-arm tube housing, used as a ball joint, in accel or braking.

EDIT: Or, if you've got a spring / damper / pushrod / pullrod force putting an a-arm in bending, then you can consider bump forces as well.

Steve, isn't the issue with this kind of application of a rod end joint the 'bending' stress, rather than the 'shear' stress? I mean yes the rod end shank will still be under some shear stress, but the majority of stress is going to be caused by the bending moment acting on the part. This is what the forum post is all about. I'm curious as to why you've chosen this descriptor over the more appropriate one, the one that everyone is talking about.

[Will M]

Steve,

Take a look at this photo:
http://images1.mcmaster.com/mva/Contents/gfx/small/8405kc3s.png?ver=1312187806
I'm not sure we're using the same terminology.


I just sent McMastCarr this email to get some more information.

***

Hello,

I have a question about your Thrust-Rated Ball Joint Rod Ends.
Specifically, is the thrust rated for the rod end threaded into something or just for the ball 'pushout' force?
(ie for the whole assembly like it would be used or an idealized situation where the rod end is supported all around the ball)

Thanks,
William

***

-William

[stever95]

Folks.. let me preface this with the following - I am not trying to be coy or get you hung up on terminology. Just expressing my opinion (which I believe to be factual :D)

To reiterate: My purpose in this discussion is to highlight that it is possible to avoid rod ends in bending but still have a rod end in shear, a load case for which they are not rated or designed as far as I am aware.


Steve, isn't the issue with this kind of application of a rod end joint the 'bending' stress, rather than the 'shear' stress? I mean yes the rod end shank will still be under some shear stress, but the majority of stress is going to be caused by the bending moment acting on the part. This is what the forum post is all about. I'm curious as to why you've chosen this descriptor over the more appropriate one, the one that everyone is talking about.

CWA,

As I said before - a rod end which is fully threaded into its housing, loaded in the X direction with the rod end in the Y direction, will be in single shear, NOT bending (unless you decide that the ~5mm offset from the bolt force to the threaded shank constitutes a moment arm). If this still doesn't make sense in words I can make a quick diagram to better show what I'm trying to say.


Steve,

Take a look at this photo:
http://images1.mcmaster.com/mva/Contents/gfx/small/8405kc3s.png?ver=1312187806
I'm not sure we're using the same terminology.


I just sent McMastCarr this email to get some more information.

***

Hello,

I have a question about your Thrust-Rated Ball Joint Rod Ends.
Specifically, is the thrust rated for the rod end threaded into something or just for the ball 'pushout' force?
(ie for the whole assembly like it would be used or an idealized situation where the rod end is supported all around the ball)

Thanks,
William

***

-William

William,

I did see this image when I checked out the McM link. Thrust loads in the Z direction are not my concern item. However, if my understanding is correct, they should really rename "Radial" as "Axial."

Thoughts?

[Jay Lawrence]

Steve,

If you do some calcs you'll probably find that the bending stress is always greater than the shear stress, even when you reduce the moment arm to its smallest possible configuration.

[CWA]

Folks.. let me preface this with the following - I am not trying to be coy or get you hung up on terminology. Just expressing my opinion (which I believe to be factual :D)

My favourite lecturer in uni would often, on the subject of terminology, refer to a Lewis Carroll quote from Alice on Wonderland; "when I use a word, it means just what I choose it to mean — neither more nor less". It was to help us understand that whilst large efforts can be made to standardize engineering definitions, often some artistic license is used by engineers (or sometimes people just straight up do not know what they are talking about). Thus there is nothing wrong with, in fact it is often critical to, as Z would agree, establish your definitions. Hence why I asked.

And yes, everyone believes their own opinion to be factual. And when some people hold opinions that aren't quite as factual as others, it is either proven so with results, or deemed so by general consensus.


To reiterate: My purpose in this discussion is to highlight that it is possible to avoid rod ends in bending but still have a rod end in shear, a load case for which they are not rated or designed as far as I am aware.

And I believe it is not possible to avoid fitting a rod end bearing in bending when it is stuck in the outer end of a wishbone as pictured previously. Certainly I am not aware that it is possible to fit a rod end joint shank in shear without bending by any conventional means.

Rod end bearings are usually rated for neither 'bending' nor 'shear' loads. In my experience of designing for REIB, you have to seek the yield stress of the RE housing material from the manufacturer (or try and derive it from available radial ratings), make some predictions of the bending (and shear) and tensile stresses that will be seen under your operational loads, then apply a decent enough factor of ignorance to make yourself feel less nervous. Pretty normal stuff, is it not?


As I said before - a rod end which is fully threaded into its housing, loaded in the X direction with the rod end in the Y direction, will be in single shear, NOT bending (unless you decide that the ~5mm offset from the bolt force to the threaded shank constitutes a moment arm). If this still doesn't make sense in words I can make a quick diagram to better show what I'm trying to say.

Of course you consider the moment arm. As Jay said, the bending stress caused by this very moment arm can often work out to be an order of magnitude larger than any shear stresses seen. This is exactly why the FSAE judges place such scrutiny on REIB! It is what this whole thread is about, and all the other legacy threads on the subject in this forum. The issue has never been known as REIS! But please draw a diagram anyway, maybe I have misunderstood what you are saying.

Also, the actual moment arm will be "5mm" (I assume you refer to the wall thickness of the RE body between ball joint hole and shank here) PLUS ~half the diameter of the ball insert. A moment arm is defined as the perpendicular distance between the pivot (thread base/outer end of wishbone in this case) and the force line of action (which you have to assume is acting through the centre bore of your RE ball). Perhaps this is pedantic, but let's not be misleading.

So for this M8 rod end, your '5mm moment arm' is actually 17mm - http://www.mcgillmotorsport.com/images/SAK2(2).jpg

[CWA]

I did see this image when I checked out the McM link. Thrust loads in the Z direction are not my concern item. However, if my understanding is correct, they should really rename "Radial" as "Axial." ?

No. If anything, they should rename what is depicted in the image as "Thrust Load" to "Axial" load, as this is generally more standard terminology in my experience (the axis being the axis of the ball joint insert with no angular displacement). The image depicts 'Radial Load' correctly; this is a force acting on the radius of the hole in the body. Other rod end joint suppliers will go on to clarify that the direction of this rated radial load is usually only in line with the shank axis (putting the shank in pure tension or compression only) and thus giving no useful clues as to the 'bending strength' of the part.

McM's image depicts the direction that 'thrust' or 'axial' load has been applied during their tests. Will is asking for clarification as to how the rod end is constrained to react the thrust forces applied during McM's rating tests. If the piece is screwed into something (effectively becoming a cantilever as per our discussions) then the thrust load limit may be established in McM's tests when the part shank fails in bending.

Or, the part may be constrained by placing the RE body flat onto a solid surface, with a hole in the surface larger than the ball insert of the joint. The load limit in this case will be when the thrust/axial load causes the ball insert to 'pop' out of the rod end body. In this case, the rod end shank is not being tested 'in bending' so the test results would give no clues as to the part shank's strength in bending.

Will, I look forward to a response to your question from McM. My guess would be that the test is more in line with your latter suggestion where the rod end body is supported around the ball. I suspect this mainly because this is equivalent to how spherical bearings (without shanks) are usually rated for axial 'pop-out load' too, so the method of constraint would be easy enough to apply similarly to a rod end. But if this is not the case, and the part is actually tested to failure in bending, then this is very interesting to me.


Thrust loads in the Z direction are not my concern item.

ALSO Steve, unless a person's spring damper unit / pullrod / pushrod outer joint mounts directly to their uprights, at least one of their rod end joints they have fitted into the outer end of their wishbones will certainly be seeing a good amount of axial / thrust loads under vertical force input from the tyres (static weight, weight transfer, bumps), and the associated rod end shank bending/shear stresses will be seen too. I don't know if you have plans to fit REIB on a car or not, but if you do, these loads should certainly be one of your concern items.

[Will M]

So McMaster Carr never responded (meaning that most likely their vendor never responded; MC is normally pretty good about these things).
Sounds like we need a REIB test consortium!

-William

[DougMilliken]

Sounds like we need a REIB test consortium!
Just in case you are serious...relative to ten years of TTC experience:

+ Some uni should have suitable test equipment available for use for free (not true for flat belt tire testing).
+ There are many more rod end suppliers & sizes than there are FSAE tire designs, and rod ends are relatively low price.
+ Volunteers could build some test jigs, run the tests (interesting work) and write up the results.
+ Membership could be as simple as sending in samples to be tested, and making results available to all that contribute from a secure website.
+ From TTC experience, collecting money from many different countries is time consuming. If you can run on a barter system I would strongly recommend it.

-- Doug

[Will M]

It would be a seriously good idea... but I graduated a while ago.
I like the idea of a barter system.
A 'universal' test fixture would help with that
I'm thinking steel block with a large hole in the middle and two through holes on the sides.
The rod end would be screwed into a shaft which would be inserted through the side holes.
One hole would be in mm and the other in inches so that teams could send a new shaft if the test site did not already have their thread size.

Maybe I'll do a rough draft tomorrow...

-William

[Goost]

I have a 100kN tensile testing machine.

If someone made the jig I would gladly test any rod ends provided and organize data.

For a jig I picture three setups that cover relevant loads:

1) axial tension/compression
A) an axial threaded hole in the base
B) a clevis mounted in the cross head with a bolt through the rod end bearing.
If this is free to pivot I could do compression tests too
2) bending (~short beam shear) mode 1
A) perpendicular threaded hole in base, offset such that rod end bearing is centered on machine axis
B) the clevis from 1B - this loads the race of the spherical bearing
3) bending (~short beam shear) mode 2
A) perpendicular threaded hole in base, offset such that rod end bearing is centered on machine axis
B) A bolt/threaded rod mounted in the cross head 1B - this attempts to slip the ball out of the spherical bearing

Perhaps 2 and 3 deserve a way to adjust the cantilever length?

May be easiest to simply machine fixtures for each of the few common rod end sizes.
We only used AN 3,4, and 5 sizes I believe. Maybe some teams use similar metric sizes?

[Tim.Wright]

I would actually be interested in also seeing the stiffness of the joints that comes out of this because I suspect the joints are a lot stiffer than people think. Often I hear people speaking about joint compliance whereas in a "ball jointed" suspension I'd estimate that most of the compliance is coming from the parts themselves not the joints.

However, there's not a lot of joint stiffness data available in the public domain.

[JT A.]

There is an SAE article where they tested compliance & friction in rod ends: 2006-01-3650, has some pretty good info.

[stever95]

Yikes! Been a while since I checked back. Couple things.


Steve,

If you do some calcs you'll probably find that the bending stress is always greater than the shear stress, even when you reduce the moment arm to its smallest possible configuration.

Jay! What an interesting point. I'll have to look at the solid mechanics again before I can respond intelligently, but I suspect this is not always true.



ALSO Steve, unless a person's spring damper unit / pullrod / pushrod outer joint mounts directly to their uprights, at least one of their rod end joints they have fitted into the outer end of their wishbones will certainly be seeing a good amount of axial / thrust loads under vertical force input from the tyres (static weight, weight transfer, bumps), and the associated rod end shank bending/shear stresses will be seen too. I don't know if you have plans to fit REIB on a car or not, but if you do, these loads should certainly be one of your concern items.

CWA,

I should have been more clear... The load case which puts rod ends in bending due to a Z force has been talked about ad nauseam, and it's easy for me to understand - so it's are not my concern item. I'm not gonna blow up this conversation again without providing some diagrams, I think we're getting stuck on semantics.

On the rod end test consortium... that sounds like a WONDERFUL idea!!!

[CWA]

I should have been more clear... The load case which puts rod ends in bending due to a Z force has been talked about ad nauseam, and it's easy for me to understand - so it's are not my concern item. I'm not gonna blow up this conversation again without providing some diagrams, I think we're getting stuck on semantics.

Then I have not understood your point at all, I look forward to your diagrams.

[Goost]

I'm not sure, maybe this has been discussed 'ad nauseum' but I don't see much data about it. Google search only found a couple news articles.

I threw some notes together and made a plot of (strength / axial-strength) vs (loading angle) for a common size (5/16-24) rod end.
Sorry it's a bit blurry and in a compressed folder, I hit the file size limit for the forum.

I was thinking about it and this concept may not be the issue in general;
it seems that the more difficult thing is to explain to people why their rod-ends are in bending.

I am also looking forward to your diagram, Stever95.

~~~

EDIT: if you don't want to open the image, the major take away is that more than 3 degrees misalignment halves the strength.

[CWA]

Can you attach a bigger pic, Goost?

[Goost]

Let me know if that is not better - again?

As always, I'm sure this has errors, please let me know - if your response requires Book/Chapter/Verse heading I likely won't read it all.

[CWA]

Let me know if that is not better.

I was hoping to be able to read your assumptions and equations too :)

[Goost]

CWA,

haha, fair enough! That edit should be a better image, and I corrected part of the small shear assumption calculation.

[Goost]

Got a second opinion on the clarity of the slide; this has some visual improvements but all the same calculations.

[CWA]

CWA,

haha, fair enough! That edit should be a better image, and I corrected part of the small shear assumption calculation.

Thanks Austin. Looks good to me. I like the way you've normalised against shear strength.

In support of Austin's post, I've made my own (attached) which just displays 'tensile', 'shear' and 'bending' stress for an arbitrary rod end diamater, load, and 2 arbitrary moment arm lengths, against Theta which ranges from 0 to 90 degrees. Consider my results alongside Austin's diagram; I've borrowed Austin's premise, assumptions, equations and naming conventions.

Assuming my calcs are correct, it is clear how bending stress is far more prominent than shear stress when a force is applied in any direction other than the axis of the rod end shank (i.e away from pure tension or compression of the shank). An effective 6mm shank diameter and a 10mm moment arm (L1) gives a bending stress level of more than 13 times the shear stress level. Proportionally, doubling the moment arm length (L2) doubles the bending stress level, making it around 26 times the shear stress level.