<b>Xantia Heater Matrix Leak : Cause & Prevention</b>
Good News : there is reason to think that some leaks from the Heater Matrix are preventable, and it has been observed that when a leak does occur, it can be self-healing.
On holidays, dodging kangaroos a thousand miles from home, I had an overheating incident in my 1994 Xantia 2.0i. This mishap has led to insights that are not widely known. Considering the onerous expedition of disassembly required to access the heater matrix, and the care and equipment required to replace a headgasket, I have documented my discoveries in a way that enables the reader to evaluate the merits of my observations. I hope that this may help to save others from these potentially expensive repairs.
The overheating incident was originally caused by my not understanding the peculiar radiator-cap engagement mechanism on my model of Xantia (in which the cap does not seal when turned fully clockwise).
So when I went for a holiday drive through the outback with the radiator cap inadvertently unsealed, the stored heat in the head would cause localized boiling of coolant and thus blowing water out of the expansion chamber after the engine was shut down.
NOTE : Not all Xantias have a low-coolant-level warning light. Mine lacks this.
So it was only when the temp warning light came on, that a critical loss of coolant showed itself.
An immediate shutdown saw the gauge temperature peak at 117 degreesC, so there was no reason to believe that the head would distort due to heat-induced stress-relief.
After a cooling period, the riddle of the peculiar radiator cap was recognized, the lost water was replenished, and we went on our way with the engine running normally.
But all was not well. Unknown to me, the head gasket had been compressed more than usual by the differential thermal expansion of the alloy head and the steel head bolts, and a tiny leak of combustion gases was pressurizing the cooling system to the relief setting of 23 psi.
A few hours later, we had a puddle in the footwell, and I figured out its probable cause. So I released the radiator cap, and we continued the tour by replacing lost coolant after every stop.
After a week or so, the leak into the footwell slowed and eventually stopped (which was much appreciated). And it turned out that the leak from the headgasket was so marginal that it too shut off . . . and after I could no longer detect gas coming from the cooling system, I sealed her up.
A few days later, the heater leak started again. And the test for gas from the cooling system was positive.
So I re-torqued the head bolts. To do this, I removed one bolt at a time (staring from the centre and working out to the ends), greased the shaft & thread, reinserted and torqued to 85 Nm. (the published procedure does not take account of existing compression in a head gasket, so I chose 85 Nm from past experience with a DS motor).
And it worked. The bubbling in the test bottle stopped (after a bit of burping), and has stayed that way for a good number of months.
And the leak from the heater matrix again stopped of its own accord. This strongly suggests that the leak was from a disturbed seal, rather than a split or pinhole in the hoses or matrix itself.
The relief pressure on my Xantia radiator cap is 1.5 Bar (23psi). Which indicates that the sealing at connections to the heater matrix are vulnerable at or below 23 psi. Presumably, this kind of leak results from distorting an elastomer seal, which can over time reseat itself.
I have always used water and soluble oil (anti-corrosion) in my cooling systems. Where I live, the need for antifreeze is marginal (11am is an early start for me, Shane), and I have been wary of the intrinsically corrosive nature of glycol (decay of the corrosion buffering system imposes the 2 year/100,000km replacement period).
It figures that the design of the head in the Xantia (and presumably other modern engines) allows more thermal storage by increased distance of the water galleries from the combustion surfaces. This would be designed in for the purpose of holding a higher temperature at the combustion surface (which then leads to higher engine efficiency due to reduced heat losses into the cooling system). This would explain the trend to higher relief pressures in modern cooling systems and the trend for manufacturers to specify anti-freeze coolant : in order to squelch localized boiling in the head after shutdown.
To prevent higher pressures in the cooling system from overwhelming the seals at the heater matrix, two options present themselves.
1: Fit a header tank with a larger expansion capacity to accommodate the extra fluid that is temporarily displaced by the localized boiling, and fit a radiator cap with a lower relief pressure to vent air compressed by this process.
Or
2: use a coolant with a higher boiling temperature (as recommended by Citroen, but without any reason given).
I went with option 2 (because it takes less time to implement), using a coolant with 33% glycol. This is claimed (by Valvoline) to have a boiling point of 126 degreesC at 1.5 psi above the outside atmospheric pressure.
CONCLUSION :
These maintenance details have the potential to substantially reduce occurrences of the common and horribly inconvenient heater matrix leak.
1: Adjust the coolant fluid level to the minimum point when the engine is cold (ie allowing the maximum expansion volume in the header tank).
2: Use a 33% glycol coolant fluid.
Together, these should keep the pressure on the heater matrix to a minimum (and presumably well below the deadly 23 psi).
Best Regards to All,
Peter Mann
Xantia Heater Matrix Leak : Cause & Prevention
Moderator: RichardW
-
tomsheppard
- Posts: 1801
- Joined: 19 Dec 2002, 14:46
- Location: United Kingdom
- My Cars:
I have often wondered if topping up to the max mark in the header tank when cold would create alot of pressure when the coolant was at operating temp (and maybe force a hole in the rad or matrix) But would the manufacturer put the max mark at the top of the header tank if it was potentially bad for the system? Isn't that what rad caps are for anyway- to relieve excess pressure during an overheat and prevent pressure damage?
Peter,
Interesting to hear about how things have turned out since that long telephone call I had with you when you were in the middle of your troubles. I find my ventilation system smells less now that I have sealed the leaks around the thermostat housing (both to head and the hose) so that I don't randomly lose coolant. I still maintain the if you keep the expansion bottle topped up, the lesser air volume in the bottle keeps the pressures in the overall system down to more acceptable level. I also havn't worked out a simple way to lower the over pressure relief down to a more normal 15psi or so to reduce the stress on the heater matrix. I am running 50% glycol to give increased safety margin on the overtemp side of things.
Keep up the good work.
Ken W
Interesting to hear about how things have turned out since that long telephone call I had with you when you were in the middle of your troubles. I find my ventilation system smells less now that I have sealed the leaks around the thermostat housing (both to head and the hose) so that I don't randomly lose coolant. I still maintain the if you keep the expansion bottle topped up, the lesser air volume in the bottle keeps the pressures in the overall system down to more acceptable level. I also havn't worked out a simple way to lower the over pressure relief down to a more normal 15psi or so to reduce the stress on the heater matrix. I am running 50% glycol to give increased safety margin on the overtemp side of things.
Keep up the good work.
Ken W
It says 1.5 Bar on my radiator cap. 1 Bar is approx 15 psi.
The gas volume absorbs water displaced from the cooling system (by thermal expansion of the water). The pressure of the gas volume increases in proportion to the amount of water entering the header tank, but the defining variable is the gas volume in the header tank when cold. For example, if the air volume is ten times that of the displaced water, the gas pressure will have increased by about 12% when the engine has reached operating temperature. (heating of the air in the header tank will additionally increase the gas pressure).
If the air volume is twice that of the displaced water, the gas pressure will have doubled when the engine has reached operating temperature. (If I remember my matricualtion Chemistry OK, this is a principle described by Boyle's Gas Law). The same notion applies in our suspension spheres, where the ride becomes stiffer as the gas spring shrinks from depletion past the membrane.
I tried running it with the radiator cap from the BX, but the spring was not compressed by the normal amount, so it retained virtually no pressure. The Xantia Radiator cap appears to have an internal relief valve, rather than allowing excess pressure lift the spring-loaded rubber gasket and vent itself. Modifying a radiator cap by adding a relief valve (perhaps from industrial pneumatics) which vents out the top is a possibility I have yet to explore.
The gas volume absorbs water displaced from the cooling system (by thermal expansion of the water). The pressure of the gas volume increases in proportion to the amount of water entering the header tank, but the defining variable is the gas volume in the header tank when cold. For example, if the air volume is ten times that of the displaced water, the gas pressure will have increased by about 12% when the engine has reached operating temperature. (heating of the air in the header tank will additionally increase the gas pressure).
If the air volume is twice that of the displaced water, the gas pressure will have doubled when the engine has reached operating temperature. (If I remember my matricualtion Chemistry OK, this is a principle described by Boyle's Gas Law). The same notion applies in our suspension spheres, where the ride becomes stiffer as the gas spring shrinks from depletion past the membrane.
I tried running it with the radiator cap from the BX, but the spring was not compressed by the normal amount, so it retained virtually no pressure. The Xantia Radiator cap appears to have an internal relief valve, rather than allowing excess pressure lift the spring-loaded rubber gasket and vent itself. Modifying a radiator cap by adding a relief valve (perhaps from industrial pneumatics) which vents out the top is a possibility I have yet to explore.
Very interesting post!, i have read it twice to soak up all the valuable information!,
If you did fit a lower pressure relief valve, then you could maybe hook it into the widscreen washer system to give a visible indication of the valve venting so that you knew when it was venting system pressure and that it was time to back off the throttle or pull over and stop ??
If you did fit a lower pressure relief valve, then you could maybe hook it into the widscreen washer system to give a visible indication of the valve venting so that you knew when it was venting system pressure and that it was time to back off the throttle or pull over and stop ??
Aha ! Nifty idea if your coolant contains no glycol (which can do nasty things to paintwork). In my experience, boilover happens only when there has been a loss of coolant (so that it quenches dry areas in the head, and so causes localized boiling, which blows coolant back into the header tank), or after engine shut down (as described above).
In the Good Old Days (before radar gave the saftey nazis a stick to beat us with) I drove my DS at 90mph through outback Australia during a heatwave with water as coolant . . . no problem.
In the Good Old Days (before radar gave the saftey nazis a stick to beat us with) I drove my DS at 90mph through outback Australia during a heatwave with water as coolant . . . no problem.