Shunt wound or permanent magnet DC motors are inherently constant torque
devices because the magnetic field is constant at or below base speed.
Any way to get variable armature voltage will produce variable speed at
constant torque. snip

Thanks for all the good info. But it is my understanding that there is a
difference in the output torque of DC motors depending on the means by which
the speed is controlled. In fact, isn't that why expensive PWM drives are
preferred over other supply circuits?

Almost all industrial DC motor controls are phase controlled SCR drives. The
current ripple resulting from a phase control SCR drive can cause commutation
problems with big, low inductance motors (like 1000 HP and larger), but is not
usually a problem with smaller motors. Having to wait for the next zero
crossing of the 50 or 60 hz line in order to make a control command change can
theoretically be a problem for some high performance servo applications, but is
not typically a problem. The current ripple also causes torque ripple which
can cause an audible buzz and, in a few aplications, process problems. I'd
wory about this torque ripple if I were designing a machine to grind telescope
mirrors, for instance, but probably not for a power drill controller. PWM DC
drives can be faster responding and they can minimize torque and speed ripple,
but they are more expensive particularly in larger sizes, they are less robust,
have less thermal overload, can cause interference problems, can lead to
insulation system stress, etc.

As to the claim that PWM is somehow better because the pulses are always at
peak voltage, I do not buy that. For a given magnetic flux level, motor torque
is proportional to current, not voltage, and current in an inductive motor
circuit cannot change at the PWM carrier frequency.

Mill