The Ford Barn

krswen · 08-03-2010, 03:15 PM

I’ve been pondering the advisability of using a thermostat. Here’s why.

The Model A cooling system is very primitive compared to more modern cars. Today, when we design an engine, the water pump is situated so that it pushes coolant into the engine, usually the block first, then through the cylinder head(s) and finally to the radiator. The impeller is designed with very specific blade shapes so that the “incidence angle” of the incoming flow is minimized. The engine represents a resistance to flow which plots as an upward turning parabolic when plotted on a pressure vs flow “performance map” of the pump. We match the resistance curve with the pump size to center the curve in the “sweet spot” of pump efficiency.

In the case of the Model A, the pump is at the engine output side of the flow circuit and all the pump does is to raise water to the height of the upper radiator header tank. Density changes, which result from the heat exchange process in the tubes and fins, cause to water to sink down to the bottom of the radiator. The pump DOES NOT push water through the radiator. It cannot since there is “head-space” in the upper header tank and it is not pressurized.

On modern engines with a thermostat, it is plumbed so that when the thermostat is closed there is an “escape route” back to the engine inlet so the flow is never stagnated and the pump is never “deadheaded”.

In an A with a hot water heater (rarely done as far as I can tell) the heater circuit would provide similar relief.

However in a typical A with a thermostat in the upper hose, when it is closed the pump will be deadheaded and the impeller will be churning in non-flowing water. For modern water pumps this would be a VERY bad thing, leading to possible cavitation and impeller erosion. The A impeller has only three blades and seems to run with rather generous (by today’s standards) clearances. Maybe it doesn’t care.

There seems to be a lot of discussion in the threads I’ve found about the desirability of “slowing the flow down” in the cooling circuit. Let’s think about that for a minute. In a compact heat exchanger (the term we use in engineering to describe automotive radiators) one of the performance parameters is the flow velocity through the water tubes. The heat exchange generally improves with higher, not lower, velocities. Balanced against this however is what we call “residence time”. The longer the coolant is in the radiator the more heat it can reject to the air (in a cross-flow design which is what radiators are). However, that same increased residence time means that coolant in the engine has more time to gain heat … maybe enough to reach boiling. Also slower flow through the engine means that the temperature difference from inlet to outlet is larger, possibly leading to increased thermal stress and distortion.

So here is my question: has anyone done controlled experiments to determine if adding a thermostat has any beneficial, or deleterious effects on steady-state cooling? No argument that during warm-up (what we call “transient conditions” in engineering) you can get the engine up to temperature faster. But is it worth the possible downside effects of impeller wear and the greater thermal gradients that have to occur in the engine with longer residence time and slower flow?

Thoughts or experiences from those who have done experiments with PROPERLY running cooling systems before the thermostat was installed?

08-03-2010, 03:15 PM	#1
krswen Junior Member Join Date: Jul 2010 Location: Eureka, CA Posts: 24	Thermostat data, not opinions I’ve been pondering the advisability of using a thermostat. Here’s why. The Model A cooling system is very primitive compared to more modern cars. Today, when we design an engine, the water pump is situated so that it pushes coolant into the engine, usually the block first, then through the cylinder head(s) and finally to the radiator. The impeller is designed with very specific blade shapes so that the “incidence angle” of the incoming flow is minimized. The engine represents a resistance to flow which plots as an upward turning parabolic when plotted on a pressure vs flow “performance map” of the pump. We match the resistance curve with the pump size to center the curve in the “sweet spot” of pump efficiency. In the case of the Model A, the pump is at the engine output side of the flow circuit and all the pump does is to raise water to the height of the upper radiator header tank. Density changes, which result from the heat exchange process in the tubes and fins, cause to water to sink down to the bottom of the radiator. The pump DOES NOT push water through the radiator. It cannot since there is “head-space” in the upper header tank and it is not pressurized. On modern engines with a thermostat, it is plumbed so that when the thermostat is closed there is an “escape route” back to the engine inlet so the flow is never stagnated and the pump is never “deadheaded”. In an A with a hot water heater (rarely done as far as I can tell) the heater circuit would provide similar relief. However in a typical A with a thermostat in the upper hose, when it is closed the pump will be deadheaded and the impeller will be churning in non-flowing water. For modern water pumps this would be a VERY bad thing, leading to possible cavitation and impeller erosion. The A impeller has only three blades and seems to run with rather generous (by today’s standards) clearances. Maybe it doesn’t care. There seems to be a lot of discussion in the threads I’ve found about the desirability of “slowing the flow down” in the cooling circuit. Let’s think about that for a minute. In a compact heat exchanger (the term we use in engineering to describe automotive radiators) one of the performance parameters is the flow velocity through the water tubes. The heat exchange generally improves with higher, not lower, velocities. Balanced against this however is what we call “residence time”. The longer the coolant is in the radiator the more heat it can reject to the air (in a cross-flow design which is what radiators are). However, that same increased residence time means that coolant in the engine has more time to gain heat … maybe enough to reach boiling. Also slower flow through the engine means that the temperature difference from inlet to outlet is larger, possibly leading to increased thermal stress and distortion. So here is my question: has anyone done controlled experiments to determine if adding a thermostat has any beneficial, or deleterious effects on steady-state cooling? No argument that during warm-up (what we call “transient conditions” in engineering) you can get the engine up to temperature faster. But is it worth the possible downside effects of impeller wear and the greater thermal gradients that have to occur in the engine with longer residence time and slower flow? Thoughts or experiences from those who have done experiments with PROPERLY running cooling systems before the thermostat was installed?