View of the inside floor board of the car with the seats out.  Engine is the classic 20 hp Stanley, two cylinder double acting 4” bore and 5” stroke slide valve and Stephenson reversing link that also acts as a cutoff control.  This Pontiac Tempest was chosen because it had independent suspension, meaning that the original differential and integral transmission were bolted to the floor boards.  The stub axles were affixed to the suspension, thus greatly reducing unsprung weight.  The Stanley engine, as mounted in a Stanley automobile, is supported in the front by a ball, fitted into a socket, so that it can move up and down and rotate with the rear axle.  The original Stanley-made Locomobile had a chain drive from the engine to the axle.  This arrangement had been patented by the Stanleys and then sold to the purchasers of the Locomobile company, thus they had to come up with another way of transmitting power from the engine to the wheels, hence the straight cut spur gear on the engine meshing with the same style of gear teeth cut into the ring gear of the differential.  This was a long way around the barn to making a steam powered car with good suspension by minimizing the unsprung weight. 
            There have been no mass produced steam automobile engines and so when first working with modern steam power, the inventors/developers used existing Stanley or Locomobile engines and in a few cases Bryan Steam tractor engines.  The best made from scratch auto steam engines were the Charlie Keen and Williams Brothers engines from the mid-1950s.  These were both one off designs.  I should mention the Leslie in-line six from the late 1930s as being one of the best designs ever and also the Besler designed Kaiser Manhattan V-4 steam engine from the 1960’s—again one-off designs.  We hear rumors that Derr made a very nice V-4 steam engine, but that has been lost in the mists of time.

These are photos of the controls for the monotube boiler that was made by George Sandin of Northern California.  It is identical to the boiler he made for his Studebaker, likewise with a Stanley engine in it.  You will notice the various temperature sensors—expanding rods, and the pressurestat—a Bourdon tube.  There are numerous holes in the bell cranks for adjustment as experience was gained with this boiler.  If you know what to look for you can see in the top right corner of H four microswitches that are actuated by a lever that moves across them. 
In an automobile, steam demand is continuously variable depending on traffic, hills, speed limits, and stop signs, making it very complex to control a boiler.  Boiler control is always aimed at a constant steam pressure and steam temperature.  Basically, steam temperature is controlled by input water flow (more water cools it) and steam pressure is controlled by burner heat production.  Usually, with an atomizing gun-type burner, there is no modulation of the fire—it is only on and off.  Sandin had a two speed burner control in his Studebaker with the lower output for city driving and the higher output for highway speeds.  Anyone who has worked with a monotube steam generator will state that it is a miracle if there is any control at all with a monotube as hunting, flooding the boiler, and running it dry are the usual. 
The reason is because there is a several hundred foot length of tubing between the cool water input and the hot steam output, giving many opportunities for hysteresis.  Ideally, reading output steam temperature should give one good control.  In reality, by the time the cool input water traverses the many feet of tubing to get to the end where the steam temperature is used to control the water flow, the boiler is flooded which starts a great oscillation of steam temperature almost impossible to stop.