Vanishing Coolant of Rotax 912 Engine

Rotax 912 engine: upper coolant elbow with flange, cylinder #3. The liquid seen on the bolts is Corrosion X added after installation.

By 2017, our Rotax 912 engine had run 1,815 hrs over 22 years. 

During that time, the engine always had some small loss of coolant on long trips, as was evident from a slightly lowered coolant level in the overflow bottle. 

We attributed the change to evaporation. So during each oil change, we added a bit of coolant to that bottle. 

In recent years, the level in the overflow bottle dropped somewhat more rapidly, and we had to add coolant after each extended cross country tip.

We could not see any obvious leaks: 

The water pump did not show any leaking coolant in the bleed hole or at its four elbows, all hose connections were dry, and the upper and lower coolant elbows of the cylinder heads showed no loss of coolant.

Coolant elbow at bottom of cylinder #1

Then came a surprising change. 

We had flown from Aero Country, our home base near Dallas, to Deming, NM, to visit friends Lynn and Philip. 

Due to strong headwinds at 2,000 ft AGL and above, we cruised 1,000-1,500 ft AGL, with coolant temperature 220 deg F. The entire trip required 7.5 hrs in the air.

During the preflight for the return trip, we were in for a shock: The overflow bottle was empty! 

Yet we hadn't seen any loss of coolant during the intermediate stop in Andrews, TX, or when parked in the hangar at Deming. 

Upon closer inspection, we did see two or three drops of coolant near the top spark plug of the #3 cylinder. There wasn't any other indication of a leak.

We added coolant to the engine to bring the level up to the filler cap and also refilled the overflow bottle.

Next we warmed up the engine using high idle. Philip stood next to plane and monitored the head of cylinder #3 for leaks. 

Lo and behold, he saw one or two drops of coolant come out between the flange of the upper coolant elbow and the cylinder head.

Could this seemingly small leak result in loss of the entire content of the overflow bottle during the trip from Dallas to Deming? 

Answer

Indeed it did, and in such a way that during the refueling stop in Andrews and after the arrival in Deming no coolant loss was detected. 

When a trip is started, the cooling system of the engine contains only coolant and no air. 

As engine and coolant warm up, the expansion of the coolant creates a hydraulic pressure that almost immediately reaches the 12 psi limit of the filler cap. 

The excess coolant then goes to the overflow bottle. 

At the same time, the increased pressure pushes one or two drops of coolant out of the system at the upper elbow of cylinder #3, as seen by Philip.

Now comes the takeoff and subsequent flying en route. 

There is always some slight pressure in the coolant system. It pushes out drop after drop at the elbow. These drops are replaced by coolant from the overflow bottle.

Cylinder #3 is at the rear of the engine. The air stream through the cowl grabs the emerging coolant and converts it into a spray that leaves the engine cowl without a trace.

The leakage per minute of flight was minute. But during the course of 7.5 hours of flight, the overflow bottle was completely drained. 

Possibilities for Leaks

The Rotax engine may leak coolant at the water pump, the hose connections, and the various elbows. 

In the water pump case, the problem is a leaking seal of the impeller or leaking threads of its four elbows.

Hose connection leaks are due to insufficiently tightened clamps.

For the elbows of the cylinders, coolant may leak past the threads, or it may move past the O-rings used to seal the flanges of the upper elbows.

Repair 

If the water pump has a leaking impeller seal, do not delay repairs. 

When that seal fails, coolant escapes through a bleed hole. 

There is a secondary seal that prevents the leaking coolant from entering the engine. If that seal also fails, coolant contaminates the engine oil and may produce catastrophic failure.

For the leaking thread case of the elbows at the pump or the cylinders, Rotax has an excellent video showing the repair process; note that the link requires a paid subscription. 

We only mention that Loctite 648 specified for the repair should not be replaced by any other sealant.

The failure of O-rings used with the elbows on top of the engine is easy if it involves cylinders #1 or #2. The flanges of those elbows are readily accessed. 

Upper coolant elbow and flange of cylinder #1. The liquid seen on the bolts is Corrosion X added after our repair.

But the flanges for #3 or #4 cannot be readily removed since they are tightly covered by the intake manifolds; see the above photo for #3, where the leak occurred in our case. 

You may be tempted to try removing those elbows directly, by inventing some clever way to loosen and remove the Allen head bolts. It seems like a very difficult task.

Instead, we viewed the repair as an opportunity where we replaced the O-rings of all upper elbows as well as the hoses connecting the elbows to the center reservoir.

With a major repair such as this, we keep a log of the steps of disassembly where we also note needed replacements or adjustments as they are discovered.

Steps

Note: The size of any bolt mentioned below is the wrench size, not the diameter of the bolt.

1. Remove the carburetors. 

2. Remove two bolts so that the ignition module is no longer attached to the engine block. 

One of the bolts has a 13 mm hex head, is rather long, and uses a spacer tube. It goes vertically into the engine block. 

The second bolt has a 5 mm Allen head and attaches the module horizontally to the rear angle bracket.

Second bolt of ignition module, with 5 mm Allen head, is horizontal. It attaches the ignition module to the angle bracket.

3. Remove the 5 mm Allen bolts holding the intake manifolds. 

On the left hand side, you can easily remove the manifold. 

On the right hand side, the ignition module is still attached to the manifold. But you can lift both together by 3-4 inches and hold them up with a wooden spacer without stressing the wiring of the ignition module. 

Push rags into the intake openings in the engine block so that nothing can accidentally fall in.

4. Drain enough coolant so that the upper part of the engine is without coolant. 

That way the seal of the water pump is still surrounded by coolant and does not dry out. 

You may consider this process to be overly cautious. But such care is never misplaced when an engine has been running for many years.

5. Slice the four hoses lengthwise at the center reservoir so that you can twist them off. 

Do not cut the hoses crosswise since you want to use them for accurate measurement of the replacement hoses.

6. Remove the 5 mm Allen bolts holding the flanges of the elbows. 

The elbows come easily off the engine block. Push rags into the four openings so nothing can accidentally fall in.

7. Remove the hoses from the flanges by gentle twisting. If needed, slice them lengthwise. 

Do this very carefully so you do not damage the delicate elbows.

8. Remove the O-rings of the intake manifolds and elbows. Clean everything carefully.

9. Prepare the new O-rings by coating them with a small amount of Vaseline. 

This helps the O-rings to adapt and flow into the correct position, so to speak, when the elbows and intake manifolds are installed. 

10. Clean the threads of all bolts used for elbows and intake manifolds and coat them with anti-seize compound. 

Do this also for the 13 mm hex bolt attaching the ignition module to the engine block. 

We also brushed Corrosion X on the heads of the Allen bolts after installation since they have a tendency to corrode.

11. Install everything as noted in your log. 

Use tie wraps to prevent chafing due to vibration. 

For the 5 mm Allen bolt attaching the ignition module to the rear bracket, use blue Loctite as thread locker.

12. You have opened the fuel system and the coolant system. 

This is a major repair requiring utter attention and checking, including rebalancing of the carburetors.

13. Before the first flight, do two extended run ups on the ground where the engine warms up and then cools down. 

We carried out the rebalancing of the carburetors during the second run up. 

Carefully check for any leaks. The run ups eliminate any air trapped in the cooling system, so check the overflow bottle and add coolant as needed.

Test Flight

Stay in the airport pattern during the first flight and monitor the water temperature gauge. 

Land again, remove the cowl, and inspect everything. 

Given careful assembly, it's almost certain that everything will be okay. 

We carried out a second, more extensive test flight away from the airport, again followed on the ground by careful inspection of the engine.

And now you are ready for flying.

Learning a Lesson

When landing after a long leg of a cross-country trip, always wait a few minutes and check the coolant level in the overflow bottle. This can be conveniently done after refueling. 

If the coolant level is low, you know that a significant amount of coolant was lost. 

You can start investigating the cause right then and there, instead of being surprised by an overheating engine while flying over hostile terrain, or by an empty overflow bottle during next day's preflight check.

Another Symptom Triggering Hose Replacement

For several prior years, the overflow bottle of the cooling system seemed to collect a thin layer of blackish dust in the bottom within, say, 25 hours of service. 

Our reaction: We would clean out the bottle and refill it.

After we replaced the upper hoses during the above repair, such blackish deposits have not occurred. 

This means that these hoses were somehow shedding minute amounts of the blackish material.

This is an important insight for our repair/replacement strategy, which uses condition instead of time in service: 

When such blackish deposit occurs, it is time to replace the hoses connecting the cylinder heads with the central reservoir or the water pump.

Another Loss of Coolant

The discussion so far has addressed one particular coolant loss. There are other explanations. 

In particular, the cap of the small reservoir on top of the engine may have become defective and needs to be replaced. This manifests itself as follows. 

On flights where the coolant temperature stays well below the boiling point of the coolant, around 223 deg F at atmospheric pressure, and where the flight is at low level, there is no coolant loss. However, at higher altitudes and/or higher temperatures, there is such loss.

The defect of the cap is obvious if the gasket has become worn and no longer seals. 

But there can be another, nasty defect that cannot be seen. The cap has a hidden vacuum valve. It allows coolant to flow from the reservoir  to the cooling system when the latter system develops a vacuum. 

When that vacuum valve fails, you have the above situation. We had that failure recently, with the cap more than 28 years in service.

BMW motorcycles use that cap, too. Look for "BMW 17112345074 radiator cap" on the Internet to find a dealer. The price is about half of the Rotax price. Yet it evidently is the very same cap.

Have any questions or feedback about vanishing coolant of the Rotax engine? Please share your thoughts in the comments.