Spheres and failure modes

This is the Forum for all your Citroen Technical Questions, Problems or Advice.

Moderator: RichardW

User avatar
CitroenCrazy
Donor 2020
Posts: 124
Joined: 20 Mar 2005, 19:17
x 41

Spheres and failure modes

Post by CitroenCrazy »

I've been working on a couple of Xantias of late and ended up with a pile of spheres that had low or no pressure.
To amuse myself while it was raining, I cut them open on the lathe.

This first picture is of a sphere I replaced because it was down to half its nominal pressure. It was fundamentally healthy, just old.
You can see here just what's inside one of these mysterious green balls.
Sectioned sphere 01.JPG

This shows the metal section that closes the membrane.
The idea is that this metal dome is secured by the screw at the top, creating a gast-tight seal between the edge of the membrane and the upper body of the sphere.
Sectioned sphere 02.JPG

Now we have a comparison between a good membrane and one from an accumulator that was completely dead, probably as a result of heat.
As well as being black, the bad one was also much harder and less flexible than the other.
Sectioned sphere 03.JPG

This shows the split in the membrane
Sectioned sphere 04.JPG

So far, so predictable.....
Next, I sectioned two dead rear spheres; and when i say dead, I mean zero pressure and full of LHM.
When I split them, the membranes were still green and flexible.
The fault lay in the metal domes, which both showed fatigue-induced cracks, starting at the edges of punched holes. As a result the dome has deformed and the seal between the membrane and the sphere body has failed.

This shows a comparison betweem the domes of aftermarket spheres and what I think are genuine Citroen parts (which have the pressure stamped into the shell, in addition to ink-jet printed part numbers).
Sectioned sphere 05.JPG

A slightly blurred close up shows the cracks a bit better. What's not very clear in the photos is there's distortion around the D shaped holes, and rather than the curve of the D being smooth, it comes to a point in the middle. If you were setting out to design a stress-raiser, it would be had to do a better job. The Citroen parts are significantly better finished.
Sectioned sphere 06.JPG

So, mildly interesting ?
People talk a lot about spheres losing pressure over time, but I haven't seen any discussion about fatigue failures before.

User avatar
GiveMeABreak
Forum Admin Team
Posts: 21577
Joined: 15 Sep 2015, 19:38
x 2163

Re: Spheres and failure modes

Post by GiveMeABreak »

Thanks for posting that up. I've often suspected that aftermarket ones are inferior quality and this looks to show just that. The poor finish of the inner metal housing has clearly left the rectangular shape in the membrane and the distorted shape may well of contributed to the cracks and ultimate failure.

User avatar
CitroenCrazy
Donor 2020
Posts: 124
Joined: 20 Mar 2005, 19:17
x 41

Re: Spheres and failure modes

Post by CitroenCrazy »

Hi Marc
Yes, it's been an interesting afternoon.
In total, I've cut open 5 spheres; these are the membranes.

On the left is the good one; it was in a sphere that still had 20+ bars of pressure.

In the middle are two from Xantia accumulators, where I think the heat has killed them. Both were quite rigid and hard compared to the first.

The 2 on the right are the ones where the metal domes suffered fatigue cracking. They are still flexible, like the the one on the left, though deformed because they've been compressed against the sphere body when there was no gas to act as a cushion.
comparison of membranes 1.jpg

Gibbo2286
Donor 2020
Posts: 4772
Joined: 08 Jun 2011, 18:04
x 877

Re: Spheres and failure modes

Post by Gibbo2286 »

I think you're kidding us, you've been in Katie Price's dustbin. :)

User avatar
white exec
Moderating Team
Posts: 6292
Joined: 21 Dec 2015, 13:46
x 1060

Re: Spheres and failure modes

Post by white exec »

That's a splendid bit of sphere dissection, done really tidily on the lathe.

I opened up a genuine Citroen sphere here some time ago - cut it into cross-section with a thin angle-grinder disc...
DSCF2158.JPG
The punchings and fold-over in the metal cap were also nicely smooth, and would have added strength to the hole, I guess.

I ought to cut open some more, really. Have a few here, all genuine Citroen.
Interesting to see your different membrane materials.
The mechanics' handbook (Xantia-XM-Synergie) lists the membrane type (Multicouche, Desmopan, Urepan) for each sphere, along with its 96....... production number, which is printed on the sphere.

The M diaphragm (the most heat-resistant, iirc) is the subject of a Citroen patent, so I guess aftermarket manufacturers would not be using it.

User avatar
CitroenCrazy
Donor 2020
Posts: 124
Joined: 20 Mar 2005, 19:17
x 41

Re: Spheres and failure modes

Post by CitroenCrazy »

Nice work Chris, I don't think I could be that precise with a cutting disc

User avatar
ksanturion10
Posts: 156
Joined: 14 Oct 2018, 21:59
x 16

Re: Spheres and failure modes

Post by ksanturion10 »

Quick question if I may :),
lots of fluid leaking on unscrewing is definitive for flooded nitrogen chamber or this is normal?

BR
Martin

User avatar
white exec
Moderating Team
Posts: 6292
Joined: 21 Dec 2015, 13:46
x 1060

Re: Spheres and failure modes

Post by white exec »

Not necessarily. Could just be that the system (or that part of it) wasn't completely depressurised when the sphere was unscrewed. Fluid could get pushed out by either system pressure, or sphere gas pressure pushing against the diaphragm.

A sphere full of fluid (gas pressure completely gone) won't squirt when unscrewed. If it squirts, likely that some gas pressure was still there.

User avatar
ksanturion10
Posts: 156
Joined: 14 Oct 2018, 21:59
x 16

Re: Spheres and failure modes

Post by ksanturion10 »

Pretty logical actually, so I'm presuming left system pressure, as it started leaking directly past the thread at lets say 1 turn. Not sure, but at least 300 ml went over.
Will try to diagnose the sphere by removing the valve and pushing the membrane with something spherical at the tip.
Depressurizing was done at low with engine off and opening the 12 mm screw at the accumulator.
Could it be the anti-sink valve sticking and so preventing proper operation at the spheres?
I've seen a couple of sections and I think, that it connects left and right corners, when in operation.

BR
Martin

User avatar
white exec
Moderating Team
Posts: 6292
Joined: 21 Dec 2015, 13:46
x 1060

Re: Spheres and failure modes

Post by white exec »

If you want to re-use the sphere, it might not be a good idea to push against the diaphragm with the top bolt removed.
Without the bolt tightly in place, the established seal between the diaphragm and the upper sphere surface will likely be broken, which might not help subsequently.

When I fit Valprex valves to existing spheres, I'm always careful to remove the plug/bolt gently, so as not to push down on the dome plate.

User avatar
ksanturion10
Posts: 156
Joined: 14 Oct 2018, 21:59
x 16

Re: Spheres and failure modes

Post by ksanturion10 »

Yup, my English continues to be slightly dubious :D.
Actually, finally had the time to change and the rear spheres, first was the SC.MAC (anti-sink), as I wanted to see if it will cure the problems I have with the rear suspension. With no apparent improvement, I changed and both rear corner spheres, first was the rear right, and there it leaked a lot from the main thread, where it bolts onto the cylinder - not the "valprex thread". At the rear left it leaked like 3 spoons, so I thought, that there must be something different between the spheres (flat, punctured or so) as I depressurized the same way for the right and the left. I did them one at a time, not both simultaneously at first depressurization - so it was depressurization -> change of rear right corner -> up to highest so to jack the other side -> low again with depressurization -> change of rear left corner. The different LHM amounts, that leaked is what puzzles me, but probably it is normal.
I saw a thread somewhere, think it was yours, about removing the damper from the sphere (sort of a valve for me) - I was thinking to push the membrane from there.
Actually they are genuine ones, and someone fitted "valprex" valves onto them, so I will try not to damage them too much, as I expect, that they are still in good condition.

BR
Martin

User avatar
white exec
Moderating Team
Posts: 6292
Joined: 21 Dec 2015, 13:46
x 1060

Re: Spheres and failure modes

Post by white exec »

I don't think you'll be able to remove the damper insert from the sphere neck, as it's a factory fit in there.
I did cut off a sphere neck a while ago, to look at the dampers in detail, but the sphere was destroyed by doing this:
DSCF3728r.jpg
DSCF2714p.jpg
With the sphere off the car, if it's got a Valprex valve, and you have the Tecnosir re-gassing tool, you'll be able to check whether the sphere is good or not. Otherwise, you'll need a new sphere, but rescue those Valprex valves, because you never know when they might be useful.

I wouldn't attach too much importance to the amount of fluid escaping from those two rear spheres when you unscrewed them, because there were so many unknowns - suspension down/up/down etc, system probably not bled (at the brake bleed screws) each time, etc. etc.

There's no easy way to check whether a sphere is good or not, unless you have the proper testing kit, or the Technosir C720 tool (€150 or less). All you can do is do a bounce test on the car, and if not good, then replace the spheres. Note: Only the classic round green or grey spheres are easily regassed; the newer 'saucer' spheres cannot.

User avatar
ksanturion10
Posts: 156
Joined: 14 Oct 2018, 21:59
x 16

Re: Spheres and failure modes

Post by ksanturion10 »

Very informative, thanks.
Well, right now I'm a little bit tired from the repair I've done, so if they can not be dismantled properly I will not thinker with them at all :)
white exec wrote:
25 Sep 2020, 16:02
I wouldn't attach too much importance to the amount of fluid escaping from those two rear spheres when you unscrewed them, because there were so many unknowns - suspension down/up/down etc, system probably not bled (at the brake bleed screws) each time, etc. etc.
Here you are on the spot. Have never bled the rears, nevertheless this was the first suggestion I got from this forum, only the fronts after replacing the handbrake cables and the brake pipes to the callipers. So, the stiffing of the rear suspension may be solely due to lack of bleeding at the breaks as they are also connected to the rear corner spheres. Will see :)
Do you suggest, that the system should be bled also at all callipers whenever there is a removal of some component from the hydraulic system - I mean, depressurization followed by "citrobics" (lowest->highest->lowest->highest->normal) is not enough???

P.S.: will post separate question about bleeding of an ABS system, as I see from the provided by the forum .pdf, that Lexia can be used.

BR
Martin

Big Eb
Posts: 9
Joined: 27 Feb 2019, 15:11

Re: Spheres and failure modes

Post by Big Eb »

CitroenCrazy wrote:
27 Jun 2020, 16:54
I've been working on a couple of Xantias of late and ended up with a pile of spheres that had low or no pressure.
To amuse myself while it was raining, I cut them open on the lathe.

This first picture is of a sphere I replaced because it was down to half its nominal pressure. It was fundamentally healthy, just old.
You can see here just what's inside one of these mysterious green balls.
Image


This shows the metal section that closes the membrane.
The idea is that this metal dome is secured by the screw at the top, creating a gast-tight seal between the edge of the membrane and the upper body of the sphere.
Image


Now we have a comparison between a good membrane and one from an accumulator that was completely dead, probably as a result of heat.
As well as being black, the bad one was also much harder and less flexible than the other.
Image


This shows the split in the membrane
Image


So far, so predictable.....
Next, I sectioned two dead rear spheres; and when i say dead, I mean zero pressure and full of LHM.
When I split them, the membranes were still green and flexible.
The fault lay in the metal domes, which both showed fatigue-induced cracks, starting at the edges of punched holes. As a result the dome has deformed and the seal between the membrane and the sphere body has failed.

This shows a comparison betweem the domes of aftermarket spheres and what I think are genuine Citroen parts (which have the pressure stamped into the shell, in addition to ink-jet printed part numbers).
Image


A slightly blurred close up shows the cracks a bit better. What's not very clear in the photos is there's distortion around the D shaped holes, and rather than the curve of the D being smooth, it comes to a point in the middle. If you were setting out to design a stress-raiser, it would be had to do a better job. The Citroen parts are significantly better finished.

Image


So, mildly interesting ?
People talk a lot about spheres losing pressure over time, but I haven't seen any discussion about fatigue failures before.
Hi,as you've been working on a few Xantias ,i have a 2000,hdi 110 estate and looking for a hydraulic steering hose,the bottom one under centre sphere ,would you have 1 floating about,mine is a LHD in Portugal,so probably different to RHD.
Cheers Eb

User avatar
CitroJim
A very naughty boy
Posts: 42615
Joined: 30 Apr 2005, 23:33
x 1351

Re: Spheres and failure modes

Post by CitroJim »

Forgive me but I've only just seen this incredibly interesting thread!

Never know about fatigue failures :shock: Every day's a school day :)