The graph below shows results of a cold start and drive test done to show how quickly the MGA cooling system will warm up from a cold start with low ambient temperature. With a little thought this will provide more information than is obvious at first glance. I will put the qualifying notes first.

The car is my 1958 MGA in mostly standard form. Engine is a 1600 block bored +.030 for 1625cc displacement. Cylinder head is an "18" head from early MGB, which is nearly identical to the "16" head from the MGA 1600-MK-II engine (1622cc). The head is shaved to reduce combustion chamber volume from 43cc to 38cc to have original compression ratio (8.3:1). The only significant change is a Crane Cams fast street cam with a mild 260 degree duration and about 1/16-inch extra lift (very smooth idle and mild manners). It has a Mallory Dual Points non-vacuum distributor, which likely has little or no difference for this test. Aside from having a tad more torque at high rev's, this engine is very much similar to a stock MGA 1600 engine. This engine has accumulated 76,000 miles since rebuild as new, currently with reasonably good compression all around (135-125 psi range).

The radiator is not the original cell core unit. It has been rebuild with a vertical tube (VT) core with high density tubes and fins. It likely obstructs air flow more than the original cell core and is likely no better at dissipating heat (maybe worse). For cool weather the type of radiator core is probably insignificant, especially for this test where the thermostat remains closed to prevent fluid circulation in the radiator until it is up to normal running temperature.

Thermostat is a 180 degree "Chevy" type (plain Jane) with a 1/8-inch hole drilled in the flange to vent air when filling fluid. The thermostat was previously checked in hot water with a cooking thermometer (verified accurate at 212dF in boiling water). The thermostat would start to open a few degrees late around 182-183dF, and be fully open 10 to 12dF later at 192-195dF. Observed data during this test at point of highest fluid temperatures seems to correlate well with these opening points.

There is also a BMC style blanking sleeve installed that is "intended" to improve summer cooling, but has not proven to make much difference in that respect. The bypass port in the head is generally obstructed but not completely sealed off. During warm up a small amount of fluid will still get through the bypass port, while only a tiny flow (negligible amount) could get past the thermostat to the radiator.

The temperature gauge was previously verified accurate within about +/-2dF at 212dF in boiling water (700-feet altitude). In prior warmer weather (90dF or higher) after a 24 hour non-running soak the gauge reads very close to ambient temperature, so it appears to be close to accurate near low end of gauge as well. It is assumed that the non-linear markings on the face dial should closely follow linearity of the actual needle motion, implying reasonable accuracy for the full scale.

Coolant fluid is 50/50 Ethylene-Glycol and water mixture (recently fresh flush and refill). Radiator pressure cap is 7-psi relief. That would be irrelevant for this test, as fluid temperature never goes above boiling at atmospheric pressure.

The car was parked outside for 24 hours (6:30pm to 6:30pm) prior to this test. Low temperature night before was 8dF, daytime high was 20dF. Ambient temperature at time of this test (two hours after sunset) was 15dF with a regular light breeze around the car. We then have a 15dF cold start, and one minute of fast idle to reduce choke appropriately. Then back out of driveway, a 2-block slow drive, right turn at a traffic light, and run up an entrance ramp onto expressway. Progress from there is generally steady 55 to 60 mph, never exceeding 1/2 throttle or 3500 rpm engine speed. At 5 miles there is slight slow down (40 mph x 1/2 mile) for interchange to another expressway. This accounts for the relationship between time and distance on the graph. In short, the car is being driven casually under normal load during warm up, not idling.

MGA Warm-Up Curve -- Click for larger image.

Cold start temperature is know by ambient air temperature. Temperature data points at one minute and two minutes are estimated after the fact in relationship to other known data. Temperature points from 90dF and higher (3 minutes and later) came from the car's temperature gauge. Navigator (my adult daughter Teresa) was carefully monitoring the clock and temperature readings and recording the data, and data is confirmed by driver (myself). The heater valve was full open and heater fan running throughout the test, so any effect of the heater would tend to slow the warm up process slightly.

Now for the interesting observations. Before the thermostat opens, coolant temperature is ramping up at a rate of 30 degrees per minute. The thermostat opening point (180dF) is very obvious while watching the temperature gauge. At this point fluid in the engine is warm while fluid in the radiator is still quite cold. When thermostat starts to open, the rate of temperature rise subsides dramatically to just 5 degrees in the next minute. Then there is a sudden drop of 5dF in barely more than 5 seconds, followed by gradual increase again. This is apparently a result of the first slug of cold water from the radiator reaching the thermostat (and temperature sensor). This oscillating temperature process repeats at about 1/2 minute intervals, gradually diminishing as radiator fluid is warming. Temperature stabilizes within a few minutes when all fluid is mixed and warmed, final settling in regularly between 186-188dF.

The heater had the windscreen defrosted by the time fluid temperature hit 120dF. It was putting out notable heat for the passengers at 150dF, and shortly after it hit 180dF the hat and gloves came off. At the 24 mile mark we grabbed an exit ramp, turned around and headed back enjoying the "cool" evening drive. 25 minutes later we checked in for natter 'n' noggin with several local MG club members, once again driving the only MG in the car park. I suppose it could have been more fun if it was snowing. Maybe next time.

Another casual observation: If the thermostat failed to open, coolant temperature might hit the boiling point (250dF+) around the 10 minute mark. Food for thought.