This winter I have made more flights than usual with OAT in the range of 10-30 degrees. Since the temperatures here are generally not considered hot, I had retained the bypass thermostat in the 1986 13B and used the heater core from the donor car for cabin heat. Recently the heater core developed a slow leak so it was replaced with a new one. The new one developed a leak in a short time too. While it is nice to have heat in the cockpit, obtaining it from hot coolant dripping onto the center floor section wasn't the preferred method.
Finding the source of the leaks in the heater cores proved difficult until the inside was rinsed several times with alcohol and dried with compressed air. Then the cores were pressurized to 10 psi with air and immersed in water. The leaks in both cores were from cracks in the end tank where it was bent to form the double pass function. Soldering the cracks only stopped the leaks for a short time after which the same area failed again.
My cooling system uses an expansion tank with a 14 lb cap which has never had a problem containing the system pressure. The expansion tank is connected to the area of the water pump just below the thermostat with a 1/4" hose with a 1/16" diameter constriction. Another 1/4" hose connects the expansion tank to the water pump inlet. This allows air to be purged to the expansion tank since the expansion tank hose connections are below the liquid level. During phase 1 testing, measurements of pressure in the area just below the thermostat showed pressures up to 40 psi under conditions where the thermostat was modulating the system pressure by being only partially open and engine RPM over 4500. The constriction in the line between the high pressure area below the thermostat and the expansion tank prevents the expansion tank from seeing that high pressure and also limits the coolant flow that bypasses the radiators.
Similarly, a 3/16" constriction was placed in the 5/8" hose to the inlet of the heater core to limit the pressure in the core. Another 5/8" hose connects the outlet of the heater core to the water pump inlet.
Since it was possible that the pressure was still too high in the heater core and causing the leaks, a pressure gauge was connected to the heater core inlet with a 4 ft section of 1/8" id Teflon tubing. Under cruise conditions where the thermostat was controlling the coolant temperature by being partially open, the average pressure at the core inlet was about 10 psi which should be reasonable. The term "average" is used because the pressure on the gauge was observed to be oscillating 2 to 3 times a second from about 5 to 15 psi. The coiled Teflon line to the gauge was also observed to be squirming in sync with the gauge pressure changes.
The conclusion is that the heater core tanks fail not because of the average pressure but because of fatigue due to the pressure oscillations.
So why is the coolant pressure oscillating?
One possibility is that the water pump is unstable. It is known that centrifugal pumps can exhibit behavior where they can have two different flow rates at the same head. Systems with these unstable pumps can sometimes vibrate due to oscillation between the two flow rates.
Another possibility is that the bypass thermostat is oscillating in the partially open position. Calculations of the area of the bypass passage combined with the measured pressure difference across the thermostat bypass blocking plate and the measurement of the thermostat spring force show that this is possible. The oscillation could occur because the velocity of the coolant between the blocking plate and the bypass inlet increases as the thermostat opens and Bernoulli lowers the pressure in this region. The blocking plate is pulled down against the seat at which time the bypass flow is blocked, the pressure under the plate rises and the plate opens again. The sequence of events is then repeated. It should be noted that the oscillation would be due to pressure and flow, not temperature change since the thermostat cannot respond to temperature changes fast enough to open or close 2 to 3 times a second. Bernoulli would be working at the top of the thermostat, too, contributing to the problem.
A possible solution is to block the bypass passage and install a non bypass thermostat. A flight with this configuration on Friday showed that the pressure oscillations had been eliminated.
Another possible solution is to block the bypass passage and eliminate the thermostat completely. A flight today showed that the pressure oscillations had been eliminated with this configuration, too.
If the coolant temperature doesn't remain too cold, the preferred solution would be to eliminate the thermostat. Otherwise, the coolant temperature could be controlled by throttling the cowling exit area or with a non bypass thermostat, the choice being a compromise between the reliability of the two methods.
In hindsight, the coolant pressure oscillations may have been the cause of several other issues. One is that I always had a problem with chaffing of the coolant hoses between the engine and the radiators. Tying the hoses down and installing supports to them was somewhat effective, but resulted in having to replace the supports made from 3/8" aluminum tubing periodically due to cracking. I had attributed this to air flow through the cowling and engine vibration, but seeing the writhing Teflon line to the pressure gauge has caused me to rethink this.
Another issue was a clicking noise of several times a second frequency that sometimes would occur during cruise at OAT's where cooling was more than sufficient. This may have been due to the movement of the coolant hoses or the sound of the thermostat motion itself.
A third issue was an airframe vibration with a frequency similar to the clicking noise. I searched for a cause of this with a possibility of an oil canning surface or engine roughness, but could not isolate the cause.
I have not been able to initiate the pressure oscillations, clicking noise, or airframe vibration since removing the bypass thermostat. With only two flights so far, though, it is too soon to draw definite conclusions on the noise, vibration, or chaffing, but the coolant pressure oscillation is definitely gone.
This is experience is reported in the light of the recent thermostat discussion, with the caveat that my problems may be unique.
Steve Boese
RV6A, 1986 13B NA, RD1A, EC2