Alright. I know that WWII-era submarines could vary their underwater speed, and that they were more efficient at lower speeds–which means they had some way of efficiently controlling the speed their electric motors were running at. And they were able to do this without semi-conductors.
So how exactly did they do it? Some sort of mechanical pulse-width modulation? Mechanical transmissions? What? 
I have no personal knowledge, but they must have had many, many batteries on board so it would be an easy matter to have wiring set up to have many different voltages available.
Having the batteries wired in series and then “tapping” the chain at different locations would certainly change the voltage efficiently, but can you easily recharge such a setup without having to measure the specific gravity of each battery and recharge it accordingly?
Now that I think about it, I guess you could also just switch between groups of batteries, e.g., if you need half-power you could run off half the batteries, then switch to the other half, then back, etc.
I’d be really interested in details if this is, in fact, the case.
You could configure the batteries in banks so that various series and paralell combinations could be selected.
You could also vary the field current to a shunt connected motor. Reducing the field current is very much like changing gearing. When field current is reduced, speed increases, but torque is reduced. Since field current is modest compared to armature current, it would be practical to control this either with battery bank taps, or variable or tapped resistance.
Alternativly, two field windings could give 3 speeds when supplied from a fixed battery voltage:
1-Low n * I winding Highest speed
2-High n * I winding medium speed
3-Both windings in paralell (or maybe series) …lowest speed.
The effect of series connection depends on the winding turns and resistance. It might be useful as a fourth speed.
Am I misunderstanding the question? Why not control the current with a simple rheostat?
I’m guessing that this would be wasting far too much energy - longevity of the power in the batteries would surely be a priority.
From here
I would guess that they could tie in banks of batteries, or switch them from parallel to series circuits, to control the voltage over and amperage available to the motors. It’s possible they could have used some kind of mechanical modulation; heck, that’s how (most) anti-lock braking systems work today.
Stranger
To elaborate on what GorillaMan said, these were BIG electric motors, from the site that Stranger on a Train linked to:
If you used a rheostat you could end up dissipating 100’s of kilowats in it. That’s a LOT of heat and wasted power.
Too bad it doesn’t go into detail on how the electric motors were controlled…
(I should also point out that various sites I found along the way suggested the top underwater speed of U-Boats was 7 knots, so we’re only talking about a difference between ‘slow’ and ‘very slow’)
Sweet! Thanks, GorillaMan.
Looks like Kevbo’s guess was closest.
As long as we are guessing I’m goint to go witn an amplidyne system. In this system you use a DC generator to supply armature current to a DC motor which has a separately excited field. That is, the motor field is connected to some constant source of DC voltage so that the field current is constant. By controlling the armature current you control the motor torque and thus the amount of power being put into the ship’s screws. By using a DC generator to supply the motor armature current you can use a low powered rheostat, or perhaps battery taps, to control the generator field current which determines how much armature current is produced. A tachometer tells the ship’s crew the rotation rate of the screws and they can adjust it as desired.
A more elaborate system could have the amplidyne generator field current controlled by the output from a rheostat operated by a small control motor. The input to the control motor would then be the difference between an input control voltage (which is a command for a certain output rpm) and the output of the tachometer. When you first start up the output tachometer output is zero and the control motor runs the amplidyne generator field control rheostate up to some level which starts the screws turning resulting in a tachometer output. Eventually, depending upon the time constants in the system, the input command and the tachometer output are equal and the amplidyne field control motot now has zero input voltage so it stops with the rheostat at the generator field voltage that is required to produce the armature current for a given output rpm.
At half speed a simple rheostat is dissipating (as heat) the same amount of power (more actually) as the motor is delivering mechanically to the shaft. Not only do you have to get rid of the heat (not too hard for a sub) but you are wasting battery power…so you can’t stay under as long, or you need larger, heavier, more expensive batteries.
What many people know as a rheostat today is the circuit used in a modern light dimmer switch. This is a fairly sophisticated circuit which uses two components ( a triac and a diac) that were not invented until several decades after the U-boat era. It actually switches an AC waveform on for part of a cycle. Because it is a switch, power dissipation is minimal. In contrast, a true rheostat is a variable resistor, which is not an efficient way to control power.
I’m a little confused, David Simmons, let me know if I’m understanding this right:
The main batteries provide current to the armature of the main motor, and power a smaller, rheostat-regulated (possibly automatically regulated) motor that drives a generator that supplies the current for the field of the main motor?
No, you’ve got it backwards. Amplidyne control is standard method of motor control. I didn’t dream it up on the spur of the moment. But I’ve developed a problem with my amplidyne which I don’t yet see how to get around. I’ll get back to you either with an explanation or a withdrawal of the idea.
Field current control of the main motor speed is the only alternative that I see to an amplidyne. Controlling the main propulsion motor armature current directly with a rheostat requires a rheostat that has to handle what - 1000 horsepower?
Let me know–I hadn’t heard of amplidyne control before. I didn’t mean to reverse the armature and the field windings in my summary; just a mixup–other then that, was I understanding what you were saying?
And a rheostat to directly regulate the voltage on a motor that big would have to handle at least half a megawatt. It’d work great if you needed to do a lot of baking. 
1,000 horsepower would roughly be about 750,000 watts (about ¾ megawatts).
A rheostat that could handle that much power would probably be bigger than the submarine. Never mind the fact that rheostats are an inefficient way to control electricity, they are also not capable of controlling huge amounts of electricity either.
Here is a diagram ofa simple amplidyne motor speed control system as described in paragraph one of my first post.
The sticky part of it is that the power to drive the submarine comes from the generator. That means that there must be a motor driving the generator that is big enough to drive the submarine and make up for the inefficiencies of the screw drive motor and the generator. However, the system is so far superior to direct field control of the screw drive motor that it might be worth it.
Just so we can see what kind of voltages and currents are involved this site gives the numbers. Scroll down to “The Batteries.”
The US subs had two banks of lead-acid batteries that gave 250 V and a max of 4000 Amp. So the maximum power out of the batteries would be 1 megawatt or 1340 horsepower.
Using a tappled battery to control armature current would mean switching potentially very large currents with all of the attendant problems of arcing. It could be done I guess but doesn’t look attractive.
Controlling drive motor field current would mean switching much less current but the speed control might not be all that precise. However this is the simplest method.
The amplidyne system givers excellent speed control (which might not be a requirement) but means having 3 large electric machines instead of just one.
