MGF Head Gasket Failure

MGF Head Gasket Failure.

Root Cause Analysis, Solution Proposal and Validation.

Introduction.

Head Gasket Failure in the MGF is the main concern for owners and prospective buyers. Most MGF’s seem to suffer at least one failure, and often, more than that.

Head Gasket Failure is not limited the MGF. All K-Series engines, including the KV6 suffer – but none as bad as the rear engine MGF.

To understand the root cause of the failure we should look to the design of the cooling system, we should also think about what features are unique to the F, and any commonality in the individual failures.

Firstly I will review the system, and instrument the vehicle as necessary to determine the exact behaviour of the system.

The MGF Cooling System:

1. When cold, the coolant cycles around as shown in Figure(1). I.e. The radiator is excluded from the circuit.

2. As the coolant flowing in this ‘bypass’ approaches normal operating temperature the thermostat begins to open.

3. Coolant in the radiator circuit is now able to flow – the coolant flows across through the open thermostat and into the engine. Figure(2).

4. The thermostat opens and closes accordingly to regulate coolant temperature by regulating the flow through the radiator circuit.

Fig(1): System flow diagram before thermostat opens:

(Ignore the narrow pipe running from the jiggle valve to the header tank, and the narrow pipe running from the cylinder head to the header tank – these merely allow air bubbles to disperse and maintain coolant level respectively. They are not involved in the flow)

Fig(2): System flow diagram after thermostat opens:

Fig(2.)

At this point, it should be noted that the configuration of the K-Series cooling system is unusual. Normally the thermostat would be located on the ‘outward’ flowing side of the block. It should also be considered, that the volume of cold water in the radiator circuit is far greater than a front engine vehicle.

Consider the effect of locating the thermostat here, coupled with the high volume of water in the system.

As the coolant in the bypass circuit approaches operating temperature, the thermostat begins to open.
Cold coolant from the radiator circuit now flows across the thermostat and into the block.
The cold water closes the thermostat, and once again hot water in the bypass flows over the thermostat into the head.
The thermostat begins to open again and so the cycle repeats.

The volume of cold water available is very large, so this cycle repeats longer than on a conventional front engine car.
The bulk of the cold water, i.e that in the radiator is at ambient air temperature. The radiator is not heated by conduction or radiation as it would be in a front engine vehicle.

The thermostat location allows for no mixing / blending of the cold & hot water before the temperature regulation or before the coolant enters the engine.

HGF Failure Root Cause Theory.

HGF tends to occur in the winter more often – non intuitive at first thought.
Though the front mounted K-series does suffer failures, they are not as frequent as the F.
Water pump impeller failure is common – as described in Roger Parkers’ article ‘Overheating MGF – Don’t Just Assume – Check!’ (Enjoying MG 2006).
Dislodged cylinder liners often blamed.
Stretched, loosened head bolts also often blamed.

All of these factors are easily accounted for in the theory.

Due to the location of the thermostat, cold coolant is allowed to surge into the block without being mixed with hot coolant first. The thermostat closes and the cycle repeats.

The result: a repeated thermal shock to the head. The cumulative effect of this thermo-mechanical stress is the mechanical failure of one or more components leading to HGF.
A large volume of cold coolant is present so this cycle can repeat a large proportion of your journey – particularly during the winter months.
In the winter the coolant in the radiator rapidly cools if it has had a chance to warm.
Sudden variations in the temperature differential across the head cause thermo-mechanical stress to all components attached. The wet liners are stressed repeatedly and the asymmetrical forces across the head can cause the head bolts to stretch and loosen. In the worst case the head will warp, or the gasket gives way under the changing forces.
The water pump impeller fastening is in the direct firing line of this thermal shock – leading to fatigue failures.

The significant difference here is caused by the location of the thermostat. Compared to conventional engines (e.g. Rover T-Series or O-Series) with the thermostat Located on the outward side of the engine, the cold coolant mixes with the hot bypass coolant before entering the block.

The temperature differentials and variations in these differentials are smaller and so the thermo-mechanical stress is much smaller.

There are two further negative effects caused by the thermostat location.

The ‘feed’ to the water pump is restricted. Flow is either through the bypass tube or the thermostat opening. I.e the pump operates less efficiently and will suffer cavitations.
Due to the ‘unusual’ location of the thermostat on the inlet side of the engine, the thermostat must be placed ‘backwards’ in the housing to allow the bypass flow to cross the bulb – i.e. the thermostat is orientated in the incorrect direction for most efficient hydrodynamic flow.

Proving the root cause theory.

I instrumented my MGF with a Dewetron 3000 Series digital/analogue data logger. 4 thermocouples and a GPS receiver. The following information was logged:

1. Coolant temperature at the thermostat housing – i.e. cooling inlet temperature.

2. Coolant temperature at the outlet elbow.

3. Coolant temperature in the radiator.

5. Engine bay temperature.

6. GPS position – to derive vehicle speed.

All experiments were conducted with an outdoor ambient air temperature of +3oC in late December / early January.

The left hand side of the unit consists of 8 + 256 analogue inputs. The four pairs of connections on this panel link to the remotely mounted thermocouple amplifiers. The logger is sat on top of a 200AmpHr YUASA battery to ensure reliable long term logging.

Results.

The vehicle was started cold (left overnight).

The run starts with some mild town driving, a blast up the motorway, followed by a long run through the countryside. 1hr in total.

The light blue line shows the vehicle speed. This determines the air flow across the radiator.

The town driving can be seen clearly up to 700 seconds – the erratic changes in speed, maximum 60KMH, consistent with normal town driving.
From 700 – 1300 seconds is a quick dash up the motorway, average speed of about 115KMH.
The countryside drive that follows shows an average higher speed than the town drive and not such severe speed variations.
The engine is stopped at 2000seconds – and the radiator temperature falls, and the heat soak on the head can be seen before they, too, begin to cool.

The thermostat housing (effectively the input hose to the block) and the outlet elbow temperature both steadily rise during the town drive up to about 80oC.

The thermostat begins to open and the cold surge of water, as predicted, can be seen at about 950seconds – the thermostat housing temperature suddenly falls.

The thermostat then closes and settles into the predicted cycle. (The housing temperature fails to fall any lower and the dark blue line – radiator temperature – levels off slightly).

A large differential between inlet temperature and outlet temperature is clear. It is also not a stable differential.

The red line shows the difference in temperature between the inlet and outlet side of the block. The initial climb is expected, but the continual erratic variation that continues will lead to the predicted thermo-mechanical stress. The spike at 500seconds corresponds to a sudden increase in the temperature differential, similar variations follow.

The data acquired supports this theory. So now to the solution.

The solution.

We now know the significant factors that cause the failure:

Thermostat location on inward side of the block.
The volume of cold coolant in the system.

These two factors must be eliminated:

So, the objective is the relocation of the thermostat and reduction in the volume of cold coolant. N.b. This is not a reduction in the volume of coolant, merely are reduction in the volume that is not warmed during the warm up cycle in Fig1 above.

The third (and important!) factor – the modification should be affordable.

Relocating the thermostat as above has the following effects:

· Reduction of cold coolant in the system. Only the coolant actually in the radiator is not warmed up gradually during the warm up cycle.

· Cold coolant in the radiator is mixed with warm coolant in the pipe work before it reaches the block.

According to the theory, the thermal shock caused by temperature transients across the cylinder head is now minimised, and so the thermo-mechanical shock is also minimised. The setup should now be much less prone to HGF.

Practical Application of the Theory.

Remote thermostat devices are available. They are very expensive.
The photograph below shows my solution to this problem – two ‘normal’ thermostat elbows can be turned back to back to create a remote ‘in-line’ thermostat housing.

The donor, in this case, was two Renault Megane 1.4 petrol engines at the local scrap yard.

Of course, this design leaves a spare ‘bypass’ connection – I used this for a thermocouple to support later data logging – you can just block this off.

This design also allows for a smooth laminar flow once the thermostat is open, the bypass tube is fairly large bore – more than adequate for the bypass during the warm up cycle.

(An 8mm thick hard aluminium ‘gasket’ has been added since this picture – the o-ring alone didn’t make a perfect seal against the second plastic housing.)

The remote housing is easily installed on the F – there is even access from above with the plastic trim in the wheel well removed:

A suitable T Junction for the bypass circuit was sourced from a scrap Porsche 944, complete with a suitable length of hose.

From Above:

The use of two thermostat housings means a blank ‘washer’ is required to fill the gap between the mounting face of the housing and the recess for the thermostat it originally held.

The original MGF thermostat was replaced with a blank in exactly the same way – created by cutting the centre out of a scrap thermostat:

Choice of Thermostat.

The normal MGF Thermostat is set at 88oC. The new location of the thermostat must be accounted for when deciding what thermostat to use.

To regulate the engine at the same temperature, the temperature drop along the pipe to the new location must be taken into account – this was measured as about 10oC in the previous experiment.

A Rover 620Ti thermostat – set at 78oC is ideal. The T-Series thermostat housing requires a ‘valve’ type thermostat, which closes the bypass when open – the valve plate is easily removed for our application.

The results:

The same route was repeated after the modifications had been made.

The car was left overnight, to ensure it was completely cold, and the ambient was again about 3oC.

It can be seen immediately that the temperature differential across the head is now smaller, and the variations in the differential virtually non existent. The warm up cycle is smoother, and the thermostat only opens once – this time smoothly. (It is not forced shut by an inrush of cold coolant).

Comparison pre/post modification.

The temperature at the inlet to the cylinder head has less erratic variation and a smoother warm up cycle.

Outlet temperature also has less erratic temperature differential.

The radiator now operates at a higher temperature – variations are again less erratic. The higher operating temperature reflects the improved performance of the water pump.

Processing the data shows the drastic improvement in head temperature differential. The key here is that the blue line (post modification) is lower and more stable – the erratic changes in the red line are the root cause of the HGF – the temperature differential is also 40% greater before the modification.

The difference is clear – and the warm up time is not noticeable affected.

An additional test was conducted post modification.

This graph shows the various temperatures, and speed, but also shows altitude (in dark green).
The journey is a return trip along an uphill section of the A3(M) – the return journey can be seen clearly in the symmetry of the altitude vs. time graph.
The car was hot at the start of this test, and shows that the new setup is more than capable of regulating engine temperature over a fast uphill struggle.
The car was stopped (engine running) for a period of time in the middle – the thermal soak can be seen on the yellow and pink lines.
This test was conducted post modification to give confidence that the new cooling system could effectively regulate the engine temperature under heavy load.

Conclusion.

The modification has had the desired effect. Without any ill side effects.

The temperature variations have been minimised and the volume of cold coolant drastically reduced.

The cumulative effect of repeat stress cycles will eventually lead to a mechanical failure – be that the wet liner sinking, the bolts stretching or the water pump failing – with this stress now minimised, HGF should be much less likely.

Extra Supporting Evidence.

Drilling of the thermostat is a successful modification.

The theory and results above explain why: the 4mm hole added to the thermostat maintains a continuous flow even whilst the thermostat is closed. The cold coolant in the system is warmed up during the warm up cycle. However, the warm up cycle is extended in this case.

MGF Cup Cars seldom experience premature HGF.

Intuitively you might think they would be first to go. However, the thermostat is removed on these cars to reduce restrictions in the coolant flow to allow better cooling when driven hard – the knock on effect is elimination of the thermal shock experienced by the cylinder head.

Other Causes of HGF.

It should be noted, that the poor coolant system design is not the only cause of HGF in the MGF.

Any defect in the coolant system can potentially lead to premature HGF. I.e. Split hoses, rusted pipe work, leaking radiator can all lead to a shortage of coolant in the system and so lead to HGF.
Split hoses are common place, the F has a labyrinth of complicated pipe work – all hoses should be regularly inspected.
The metal pipe work is prone to rusting from the inside out. This should be regularly inspected and replaced if necessary.
The safest option is to fit a coolant level sensor – alerting you to sudden loss of coolant. The Rover dealer can supply you with a coolant tank and a level sensor for a 2005 edition MG-TF, this can be wired directly to a buzzer in the cab – this will cost under £20. Alternatively a pre-made kit is available but is considerably more expensive.
If the engine is cut before the coolant drains out risk of damage is significantly reduced.

David Monks. Uk 2006.

Credits and Acknowledgments.

This article has not been written for financial gain!

Credit is due to the following websites and articles:

‘Overheating MGF – Don’t Just Assume – Check!’ - Roger Parker, Enjoying MG 2006.

Rob at http://www.mgf.ultimatemg.com/

Dieter at http://www.mgfcar.de/

Dave at http://www.rovercarhospital.co.uk/

Carlos at http://web.tiscali.it/elise_s1/index.htm