Clutch Failure - Sense Check

[revilla]

I know it shouldn't happen. But it did!

My comment about more torque was a bit tongue-in-cheek, there no way my engine is producing the sort of torque that should tear apart a clutch cover. I know that.

Any ideas on why it might happen? I think it's a pretty unusual mode of failure.

The cover was properly located on the dowels and the bolts were properly torqued and all still in place, intact and tight.

There's no sign of anything else wrong, there hasn't been any unusual vibration (well, until it actually failed), the spigot bearing was fine, the CRB was still pretty much OK, the engine is running smoothly, the flywheel was fine, there was no obvious foreign object damage, the friction plate is only lightly worn and intact, and yet the cover was coming apart at the seams all the way around. Just disintegrating.

I know it's done at least 32000 spirited road miles with me; if it's as old as the car it could have done 43000 maximum.

[rj]

Oh,

Have just had a look at the last dyno plot: 195.7Nm which equals to 144.3 lb ft. (This was in 2007 and a lot has happened since then)

 

[John Vine]

Glad to hear you've found the problem, Andrew.  Mind you, I don't think any of us predicted that type of failure!  I've never seen anything like it.

Out of interest, what make was the cover plate?  You might want to send those photos to them for comment. 

PS: Don't forget the pre-load.

JV

[revilla]

The cover plate was the standard (at the time) Caterham AP Racing offering.

[Wrightpayne]

OMG!! Lucky it didnt snap off completely!

[rj]

I would contact AP and ask them why this could happen.

[Alastair B]

The cracking from the photos all seem to be starting from the same location in the apertures, could be a material issue such as porosity or contamination and/or coupled with the final machining finish which could have induced micro-cracks, then with load and vibration the cracks propagate, then as ligaments fail the load is transferred to the remaining ones and the problem deteriorates more rapidly.....just a wild theory

[John Vine]

I agree. They were very good over my failed R400 driven plate.

JV

[7 wonders of the world]

and a good source of useful knowledge when I needed to change my plate too,

[7 wonders of the world]

It does look like the pressing process to form the steel cover has induced stress risers - either a raw material or maybe a heat treatment issue.

[revilla]

I will contact AP to see what they say.

I took the car out for a good test drive this morning. Other than having to tweak the clutch cable adjustment a bit, everything was fine. All of the juddering and all of the buzzing vibration in the pedal has gone. No slippage or anything to suggest any kind of problem now.

Just a couple of observations on the Helix clutch vs. the AP clutch:

1. The Helix clutch (surprisingly, to me) gives a much lighter pedal action than the AP, much less force needed to depress the clutch. Given that I use the car entirely on the road, sometimes inevitably getting caught in slow traffic, and given that I have a medical condition that results in my leg getting rather cramped and painful if I have to work a heavy clutch a lot, I like that. The AP clutch always did feel very heavy, ever since I've had the car, whereas the Helix feels a lot more like a "normal car" clutch to operate.
2. The Helix clutch does however have a sharper biting point than the AP, there's less travel between the point where it starts to bite and where it is fully engaged. With the AP clutch it seemed to take a bit of movement before the clutch started to actually disengage whereas with the Helix the biting point is more at the top of the travel. Some of this may just be that the clutch needs to settle in a bit. It's not a problem, not saying these's anything wrong with it, just a different characteristic and I need to re-educate the muscle memory in my left foot a bit!

I will report back if AP have anything to say.

Cheers, and thanks for all the input,

Andrew

PS: Big thanks to ex-member and still-lurker Mankee Cheng for coming over and giving me a hand yesterday, much easier taking the engine in and out with an extra pair of hands and much quicker disconnecting and reconnecting all those hoses and connectors with two people on the job!

[thompster]

I fitted a Helix clutch when my AP one failed about 18 months ago.  No complaints about performance and no issues. That said I was surprised to see the friction plate was LUK and the cover had the manufacturers name and part numbers ground off, which got me wondering if I'd bought a standard clutch with an expensive Helix sticker.  The AP clutches use Valeo friction material (I think) but the cover plate is stamped as AP. No idea if AP actually bespoke the covers or just stamp a pattern one though.

PS.  Good to see Mankee still about 

[revilla]

Response from AP:

Hello Andrew,

 

When we have seen a failure like this before it is normally down the flywheel. Lightweight flywheels tend to be flexible causing fatigue in the areas where your clutch has failed. It looks from your picture the flywheel is cast iron. If you can confirm this, it probably means this is not the cause.

 

Secondly, a misalignment issue can cause a failure, as each revolution of the engine can load the clutch in an abnormal way. This can come from the gearbox input shaft either angularly or axially misaligned. Do you have a spigoted flywheel to locate the input shaft?

 

Do you have a picture of the contact area on the diaphragm spring fingers and the release bearing? Does everything look concentric?

 

Regards,

Jeremy Govan Key Account Race Engineer

[garybee]

That's an interesting email, nice to see they're not treating you like a moron with their questions.

[rj]

AP!

[revilla]

Yes I tried to frame my question to them in an intelligent and polite way rather than as a complaint, and gave them as much information as I could.

In answer to their points, yes I do have a standard Rover cast iron flywheel, unmodified, and the car doesn't spend much of its like at silly RPM so I doubt flywheel flexing was the cause.

I'm running a spigot bearing that I inspected when changing the clutch and found to be healthy and correctly installed, supporting the input shaft nose.

I'm also running a standard Caterham setup with a Type 9 gearbox and a stock Caterham bellhousing dowelled to the block. All fasteners were secure so I doubt there would be any significant misalignment.

Mankee has the broken clutch parts (payment for a day's help, never let it be said I'm not a generous man ) but will get me those photographs for them over the weekend.

My guess at this stage is that one of those cracks was caused my some kind of fatigue failure, maybe from residual stress from the manufacturing process, and that once that had failed it left the clutch out of balance, slightly out of alignment or unevenly loaded and that caused the kind of cyclic loading that AP are talking about which led to the other failures.

I did notice a buzzing vibration through the clutch pedal around 4000rpm pretty much as soon as I got the car back in the road after the winter which I was sure wasn't there at the end of last season. I didn't think a lot of it at the time but looking back it was the start of this problem (it's gone completely with the new clutch) and the initial failure had probably occurred at that point, since when I've done 1200 spirited miles or so. When it started juddering badly on the Fish and Chip run, I still had several hundred miles to get home and to be honest I thought I just had a worn friction plate or something so carried on driving it normally, just using a few more revs when pulling away. So if the clutch was already broken it probably suffered a lot subsequently. I'm probably very lucky it didn't come apart completely which could have been nasty.

[John Vine]

That's a promising response from AP, Andrew.  Do they want to inspect the cover?  (I'd hope so, as I'm sure they'll have the equipment and expertise to examine it in fine detail).  Out of interest, do AP manufacture the cover plate themselves or do they buy in a pattern part and then stamp it "AP"?  

I'd be very interested to see the photos of the CRB and diaphragm fingers when you get them.  

I'm probably very lucky it didn't come apart completely which could have been nasty.

I'd say you were very fortunate indeed.  I quail at the thought of that cover disintegrating at 7000+.  There'd be high-speed shrapnel sprayed out all over the place (your legs included)!

JV

[Wrightpayne]

In my youth I remember mini racers having 1/4 plate welded on the bulkhead as the flywheel in line with your gonads! :-0

[nickh7]

It was my plate that let go and took all the friction material with it . 

[revilla]

These are the pictures. I can't see anything obvious other than some scoring of the release bearing face where it has been spinning relative to the fingers ... the issue that John Vine's "Preload Mod" addresses. Can anyone else see anything of note?

[elie boone]

There isn't much wrong at first glance, are all the fingers on the pressure plate in the same plane i.e. if you place the fingers on a flat surface does the pressure ring of the pressure plate is also level ?

i just would have liked to see a bit more grease inside the CRB.

[John Vine]

Thanks for the pics, Andrew.  Interesting.  The diaphragm fingers look in good nick to me.  That CRB is the SKF one that CC adopted to replace the standard Ford one (rather than fix the basic pre-load design problem).  It appears to have a concave surface, or is that a trick of the light?  It indeed looks scored, and also displays what looks like heat discoloration -- both symptoms of inadequate pre-load. 

I looked up CP3748-4 on the AP Racing website.  It mentioned only the -6 version.  I don't know the significance of either the -4 or the -6.  But AP's advice for the CP3748 family of covers is to use a flat-face CRB (which the Ford one is).  Whether that has any bearing on the issue, I couldn't say.

JV

[revilla]

I've got Mankee to check the old clutch and as accurately as he can measure, everything is square and level. The face of the driven plate site level when the clutch cover it sitting on the fingers. Nothing looks wrong at all (apart from the cracks!).

Yes the CRB does look a little worn and a bit dry but I don't think that had much to do wit the eventual failure as it was still running smoothly. I'm aware of the fact that one of the bearing face and fingers should be curved and one flat. I think it mostly the photographs that give the impression it was a curved bearing, coupled with h the face that the outer shoulders were rounded, but the actual operating face was flat. The new Helix clutch came as a three piece set, the plate had curly fingers and the bearing was more obviously flat.And JV - "Whether that has any bearing on the issue..." ... Oh dear!

[Wrightpayne]

It will be interesting to see what AP say next as their various modes of failure have been disproven 

[revilla]

I've emailed all the pictures and information off to them tonight, will report back if and when when I get a further reply.