mx5 diff problems
 

evening all. Was having a think about the know issues with the diff arm snapping, and obviously I want a solution like the rest of you. So I have an idea and i'd like your thoughts on the subject.

in the mx5, the diff, gearbox and engine are connected in one solid structure, tight and stiff. where these connect to the subframes, they are on rubber bushes to allow the vibration and road shock to disperse.

On the roadster, you have no subframe. I think Stot may almost have the solution with his solid mounted bracket. the theory being that the solid mounting will prevent the diff from twisting. however, I think that by having the supporting bolts running straight down to a new plate at the bottom of the diff cage, any rotational force is transferred laterally down the bolt, preventing any shearing force to the support bolts. I would still use rubber bushes to the diff arms, so any small twisting of the diff is absorbed by the bushes, as in the original car.

What are your thoughts?:confused:

IMHO it needs to be all rubber mounted, or all hard mounted.
If it's half and half the hard mount is going to take all the stress and is likely to fatigue fairly quickly.

I've gone for a hard mounting solution now, similar to Stots.

If I were starting from scratch though I'd probably use polybushes in the arms, and bolt the nose to the chassis via bolt isolation bushes.
That way it'd be nice and solid, but still have a little isolation to reduce vibration and potential fatigue.

Whatever you do, just don't use landrover mounts in shear :D

I saw that thread - oops!

normally, id agree with a solid mounting all round, but I wonder if any flex in the chassis is going to damage the diff with no absorption. but would rubber mounting give too much flex? Afterall, in the mazda, the nose is solid mounted, but the arms are bushed onto the subframe.

I'm at the point in the build ready for the stripdown to paint, and I don't want to damage the coating on the chassis by changing my mind!

Yes, but the nose is solid mounted to the gearbox, which is then rubber mounted to the chassis, so effectively it is still rubber mounted all round.

Have you had a look at the locostusa site? The MX5 (or Miata as they like to call it) is a very common donor over there, and there's numerous threads/builds/ideas on mounting the diff.

Yes, but they are connected in the mazda via the ladder beam, to keep the diff from twisting. the rubber connections are where the diff/box/engine structure join the subframes, so they are isolated from the main loads on the body. Remember the gearbox mount on the mazda is not rubber, its stiff. I was thinking by using the transmission tunnel as the ladder beam (maybe with a few additional diagonals) and solid mount the drivetrain to the tunnel, it MAY act as the original ladder beam in the mazda. rubber mounting as the original car will give a little flexability.

for instance, if you hit a bump on the road with the rear left wheel, the rear left of the car would twist slightly. if the diff was solid mounted instread of bushes then this would twist the diff arm with the framework, stressing the diff arm.

On another note, can I ask why you revised the handbrake plate at the back of the tunnel a few times? im setting mine up now, and I cant see any problems. it cant be that easy can it?

I changed it because I didn't like the angle of the cables at that bracket.
With the bracket following the seat backs the cable outers point upwards into the tunnel, but the inners have to go down to the lever.
Think it took me a few attempts to get it right without the cables rubbing on the chassis, or the plate bending too much.
Probably would have been fine the way it was originally :rolleyes:

This is my take on the problem. In the original application all is well unless silly hp is applied. If you look at the way the diff is mounted as standard it relies on the cross bar top mounting and what is effectively a tether for the nose of the diff. As mounted in the haynes as soon as you side step the clutch the diff tries to buck instead of pulling on the tether as in the original and its this that is breaking the beam. ( I think). The nearest thing I can think of that would allow the diff to be rubber mounted and effectively tether the diff nose is by using something like this http://www.torquesolution.com/v/vspf...S-MC-R56-2.jpg and not restrain the diff nose. This is the third failure of the mx5 beam and in a minute someone will get hurt by a driveshaft stabbing them in the back.

Bob

Keith Tanner of Flyin Miata seems to have found no problems with his set up, which was a welded bracket.

Eventually someone will hit on a solution. I think the key is to prevent that drivers side arm from twisting, as others have said it's design to fail in the event of a rear impact I believe ( correct me if I'm wrong). The arm if you look at it is a H-beam which is a strong section but if it's twisted it will fail from that notch they have cut.

Only other option is to reinforce the arm.

Ugly as it is, if/when mine fails I'll be interested to see how.



These are volvo engine mounts one top and one bottom with big end bolts going through the centre to 10mm thick steel sections one on the left and one to the right. If the nose lifts enough it will then stop after 10mm (I'm guessing), which I think the rear diff mounts can accommodate, and fingers crossed that's not enough to cause a failure. Problem is will it then try to pull the rear of the diff downwards around the new centre of rotation as torque is continued to be applied and fail anyway. I think this is the scenario for failures in high powered MX5's

I think Skov is right though if you are solid mounting anywhere then it has to be all round.

Useful link...

http://foxed.ca/rx7manual/2003mazdar...101025T09.html

1 comment on that - the bolts are in shear. eventually, it will snap them if there not tough enough.

As for solid mounts all round, you need to allow SOME movement due to chassis twist. No matter how braced it is, the chassis will twist in corners and over bumps. the mild steel diff cage will twist, but the cast diff wont. To be honest, I think a stiff mounted diff is more likely to fail than a bushed diff. Think how much damage a car would get with no suspension at all?:confused:

A while back i'd read this problem could be solved in a single stroke by throwing out the MX5 diff casing and housing the diff in a simpler box that would mount up the same as a Sierra diff would.

Has that idea gone out of the window ?

Phil has posted this up before with the extra chassis parts required but he removed it so I guess its not available yet.

Other problem is that many of us already have saturn based chassis so need a solution for that.

Cheers
Stot

Not out the window, but people who already have the MX5 diff cage cant simply install that box. It requires a whole new diffcage and changing the transmission tunnel. - Not an option for some.

beat me to it Stot..........

One common mod on MX5s is to stiffen the diff rubbers, because tired mounts can give a reduction in gear change quality due to excessive movement (if I remember what I read correctly, it was some time ago and I can't find the bookmark). The mounts could be stiffen by filling with shoe goo, fitting diff void fillers or stiffer mounts (which I think Mazdaspeed did along with uprated engine mounts).

The Sierra diff is solid mounted in these cars. AFAIK there have been no issues with diff mounting failures.

There are several kit car manufacturers using solid mounts on their MX-5 based builds; GBS and Roadrunner Racing are two certainties because I have seen their cars IRL. Well, GBS cut off the arms and solid mount the centre part of the diff upper.

Can someone not maufacture a brace to go over the existing arms. Both of the breaks have occurred on the right hand arm where it has had a notch taken out. The original mounts etc could then be used

David

Maybe, they are engine big end bolts, but as I say I'll be interested to see how it fails.

Agreed - but the Sierra diff is mounted at the nose, not on thin arms 300mm away from the application of torque. The mazda diffs are designed to fail. that's why the O/S arm has 2 notches taken out the arm.

Please don't think im shooting you all down - im certainly no expert!

heres a few shots of a simpe idea:

Viewed from rear



viewed from below/rear



viewed from above


Its simply 25x25x1.2 (but I will use 40x20x3) tube. I will of course use bolts to secure it, and weld it into the chassis with stiffener plates.

Can anyone see any issues?

I know, this subject has been done to near death. The arm is designed to fail to prevent the PPF from acting like a battering ram in an accident.

My point is that the Sierra diff is solid mounted (the MX-5 diff has a nose mounting as well) and doesn't suffer from chassis flex related breakages.

I think its safe to say we all know that the MX5 diff needs to be mounted at the nose, but its a case of 1) how to fix it to the chassis, 2) to bush mount, solid mount or a combination of the 2.

Were all in agreement that the Sierra diff is a superior mount, but that doesn't resolve the owners/builders with a Saturn design rear cage who are nervous about how to mount there diff.

This is my point.

The mounting options have been done to death. The Miata has been the donor of choice for a number of years Stateside, Keith Tanner has a blog which runs into hundreds of entries. They've already gone through this, we're not breaking new ground here.

Can you provide a link so we can all have a read? Cheers.

I have removed the bushes and bought Flow-Flex poly bushes which arrived today, they are like Stot's solid. Ones I was thinking of placing a large washer at the bottom so that they would be like solid mounted but with the poly bush giving a small amount of movement, I am thinking of mounting the nose of the diff with rosé joint type eyelets with poly inserts if that makes any sense.
Regards Mick.

Keith Tanner http://www.cheapsportscar.net/diary....ate=1980-01-01

After a visit to gbs ( great British sportscars & kit spares - same company ), I had a conversation with a chap called Richard. They cut the arms off their mx5 diffs & use a bracket system, curing the problem. The mx5 diff brackets are available from kit spares/gbs. :D

have you got a link to the bracket?

No mate sorry. It's not in the catalogue on the web site either. If you ring GBS on 01623 860990. It's not an item that they usually sell, but if you ask them specifically for it, they will supply one. Obviously you will have to adjust your rear cage to suit, but Richard guarantees that this will completely eliminate the problem. I have ordered my one to pick up early next week. :D

Can you post a picky when you've got it? And how much are they?

I've sat on my hands for long enough now and I've to get this out so I can sleep at night.
The chassis designs:
First off the in the beginning there was the standard Sierra diff mounting as per the book, any body remember what that looked like? Yes mounted to a pair of SB4's and a pair of 5mm thick chassis plates mounted in the middle of the chassis. There are no known brakages of this design to date even with some of your more suspect welding.
The mount sits in the middle of the chassis surrounded by that material that makes up the seat back and the diff cage,thats not going to twist unless you smack a curb side ways at speed at which point a broken diff is the least of your worries. It's solid mounted does not move and if you can feel any vibration from it after you've fitted your car with 350 and 250lbs springs your up there with the princess and the pea.

The Saturn design...
It's been no great secret amongst those that I have talked to that I am not a fan of this design as it assumes to many things in the way of peoples ability to make it, the design also over looks some very basic engineering understanding or should that be lack of it.
The design of the lower wishbones that uses the M20 adjusters is fine for the top rear Sierra set up as it just holds the upright in place, the Saturn design puts these under greater loading by putting these on the bottom. Your thinking so what theres 2 of them, at this point I'd like to point out that there is now 2 high stress points acting on 2.5mm thick tubing that may have been MIG welded by a less than proficient person.
Top rear wishbones:
What a great idea to put the end of the damper in the middle of the wishbone, it means you need a spring so hard that the car will knock your teeth out and tend to bounce over the bumps rather than absorb them. To my knowledge no one has as yet to have a Roadster corner weighted and had a set of coilovers valved to the correct requirements of the suspension. If your using cheap single adjustable bump and rebound coilovers then getting your suspension to react quick enough is going to be a problem.

The MX5 diff and it's cover...
In the Mazda shell the diff is mounted to the sub frame which could be argued does not move or twist, at the other end it's bolted to the ali spine which in the grand scheme of things only moves by 1-2 mm ever. Mazda have designed a car which smooths out the vibration of road noise to give a better ride,thats not why you are building a Roadster.
People have got caught up that the diff is rubber mounted on the arms and decided that the nose should be too, if you can move your choice of rubber mount with your fingers it's too soft, if you can move it with 2 pairs of mole grips it's too soft, if it has a metal plate bonded top and bottom it's too soft. By the time you have found the right grade of rubber mount you will have run out of cast ali diff covers.
The nose of the diff needs to be held in place top and bottom with a one piece bracket made from 5-6mm plate that is bolted to 5mm chassis plates and SB4 with M12 bolts. This will stop the nose from nodding under load and stop the arms of the diff from breaking.

GBS design:
If any body is looking at that and thinking ah they have cracked it just bolt through the ali casting at the top.... DON'T! The top of the casting is not flat at any point, there will only be the edge of the nut sat against the casting causing a stress point. All you will have is another broken casting in a different place.

Talon's diff cover...
Many said it was not needed and this may be true but consider this...
it fits the standard Roadster chassis although you have to move SB4 abit, it gives 4 large areas in which to put 4 M12 bolts, it's made of 2mm steel so you can weld extra mounting lugs to it, I doubt that any body will ever manage to brake the 6mm main plate, it has no in built fail point (well not intentionally any way), it will fit the Saturn design the same as it fits the standard chassis.
I will be publishing the drawings for the extra tubes and mounts when I have fully tested the design by trying to break it... by vigorously driving around the rough farm track and taking it drifting.

Like you Phil I will only bite my tongue so long, I am sorry for the pressure you are under in your personal life and your business, but that does not give you the right to rubbish other peoples efforts, maybe not everyones welding is top notch, but they do have a go.
I wonder how you would fare if you were to try the jobs that some of the others on here have, I am qualified to make comment on their work but wont as I know the effort that these guys put in and good for them, each time they do something on their build they are learning something new OK

my rant in now over

Arfon

Thanks for saying that Arfon, I was just about to go out to the garage and cut up my chassis, that I've spent 2 years putting together, and wishbones up for scrap as apparently there's nothing good about them. I think all too often people forget, most of us are putting a together a car in an 8x18 ft garage with the tools we have available and can afford with our own less than perfect amateur skills. But that's ok, these cars may not be perfect but it's about building a car and challenging yourself and not so much about the end product as good or bad that may be.

I wanted something to keep my mind busy and that wouldn't cost a fortune as times are hard. I also wanted my two young children to see what was possible so they don't grow up thinking what they want to do is out of reach or not possible like I did. But lets not get too deep about this.

Basically Phil you've just demotivated every MX5 Saturn design builder on here!

Nice one!

Seriously thinking of not using this forum anymore it's too depressing!

I think you have missed the point some what, go back and read my post again and try and pick out the concerns over the designs from a safety point of view rather than it being a personal attack on your self.

I didn't think it was a personnel attack on myself. I understand your concerns, but it doesn't do anybody with a finished chassis/car much good though does it. Unless they decide not to drive it. Or do it all again differently.

And you aint qualified enough to criticize me Phil, believe me you aint.

Arfon

CTWV50 you have a p.m. mate

Thanks for that Mike; saved me the effort :) To save pople flicking through, Keith Tanner's nose mount design is very similar to Phil's design.



On another note, I think people are being too sensitive here. I didn't see Phil criticise any one person. Bear in mind, I've got a SSC based MX5 chassis (well, I have 2) and I've actively contributed to discussions with other builders both on this forum and IRL on different aspects of improving the design. I don't feel offended.

Phil does have a point though about welding quality. If your welding isn't up to scratch (which is a very real possibility with a home build, not saying anyone in particular is not up to it), then should you really be building a car capable of going over 100mph?

I think he's also spot on with the diff mounting designs. It is incredibly frustrating to have the same topics come up again and again with carbon copy proposals and outcomes (not that I'm blaming you Short in the slightest, you are contributing with an alternative design). Plus, like I said, the Americans have already examined this in detail. It seems counterproductive to not use their findings as our basis and tweak the designs to suit the Roadster.

Phil has also highlighted the one thing about the SSC design I believe could be bettered; the threaded adjusters. At one level, I would look at using a different thread pitch on the wishbones so that a standard rose joint can be used.

An alternative could be to use the adjustable wishbones to set up the car, then make a second set of fixed (or semi fixed) wishbones to those settings. This is what Formula Student builders are advised to do rather than have rod ends acting as outboard ball joints, which usually fail. Although, I would point out I saw a Northern saloon car championship RS200 (replica I assume) at Stoneleigh this year with a very similar threaded adjuster design to that laid out by SSC and the book.

Another option might be to see if an adjustable system similar to OEM Mazda, or shims, could be implemented with fixed wishbones.

WRT the rear upper wishbones, I think it will be difficult to find a way around this without redesigning the outboard end of the wishbone a la Johno's latest design. I do think that it should be possible to have an acceptable set up for the rear suspension using the SSC design.

If Phil post was constructive about the possible failures of the Saturn chassis then that would be fine or peoples welding but it doesn't come across like that at all to me. Apparently he can't sleep at night thinking about it's short comings! A bit dramatic. No?

We all know this and it's a good point so do destructive tests on your welds and make sure they're not going to let you down. That's what I did. There's no need to describe possible builders as "a less then proficient person" which is very polite insult to my mind considering the tone of his post.

Thank you, this is useful information and something I hadn't thought of.


I don't see this as much of an issue and like you think it's a workable design.


With regard to the diff support design, I see this as one of the fun/interesting bit's of the build, something we can all throw ideas around with. I enjoy the discussion. I'd like to avoid solid mounting if I can but if that's what it takes then so be it.

With regards to how solid a rubber mounting needs to be, I can jack the rear of my car up using the nose of the diff (obviously only a part built car at the time)
and it does not move at all in relation to the car chassis but that might change during a hard 1st gear launch, maybe 2nd gear launches could avoid this issue completely which might not be as daft as it sounds considering the weight of the vehicle.:D

I read the not being able to sleep bit as either tongue in cheek, or he's like me. I sometimes go over what I need to do with this bit or that bit of the build.

I remember Albert posting up a youtube vid of someone showing you how to weld, where it was more like how NOT to weld. People may not be as thorough as you are with weld testing etc. And Phil did say 'may' not 'is'.

Once again, hugely constructive, thanks Phil!

So I will make this my last input on this subject (well at least for tonight as the cider monster is taking hold) but Phil have you ever completed the build of a haynes because I have never seen anything posted with a completed car built by your good self if you have I appologise if not don't slagg off other peoples efforts and designs

Good night

Arfon