I'm kinda curious as to how these were done (or if they were) before servo driver ICs became available?

I mean, I know how you could do it with a microcontroller - that's not that complicated.

...and I have some idea how it could be done conceptually with analog electronics (with the exception of the pulse stretcher part).

But I haven't found any "contemporary" (ie -1960s) books or magazine articles detailing a circuit design; usually just block diagrams, or more generally, a design is given with an input potentiometer for the position, and an output potentiometer at the end of the gear train to drive the error input to the "summing amplifier" (which drives the h-bridge to turn positing/negative.

I'm wanting to know how the whole circuit could be built from discrete parts plus some basic ICs; I figure that you'd need a op-amp or two, maybe something generating a ramp for the position control, to generate a positive or negative output voltage that would cause the h-bridge to move in the proper direction...

But I've yet to find any kind of articles contemporary to the era around when PPM/PWM hobby servos first popped up on the RC hobby scene; that is, going from certain older control systems (ie, "galloping ghost") to something closer to a modern analog servo? I know that there were servos that had the control electronics "off-servo" - most usually with 5-pin/wire designs, similar to how some of the cheaper RC proportional control cars use as "servos"...

Does anyone have a reference or a schematic of one of these old systems (assuming such a thing exists)?

 

One of the earliest servo control systems is the steam engine governor as early as 1788, long before electricity was put to use. To put that into perspective, Michael Faraday was born in 1791.







Reference: https://en.wikipedia.org/wiki/Centrifugal_governor

 

Servo control is achieved by applying negative feedback.
Think about it. When you have positive feedback in a system, it can result in "run away", for example, population explosion.
You need the opposite in order to control a system.

A simple transistor amplifier uses negative feedback. In this circuit negative feedback is applied through R1.
As the transistor conducts more current, the voltage across R2 increases. The voltage on the collector falls. Thus the base bias voltage supplied by R1 from the collector to the base also falls. This reduces the current through the transistor.




This op amp circuit uses negative feedback through the feedback resistor RF.
When the input voltage grows more positive, the output voltage goes more negative. This counteracts the input voltage and thus reduces the effects at the output. The voltage gain is lower than the open loop gain.


In a nutshell, here is how an op amp circuit works:
The error voltage is detected between the two input pins of the op amp. This voltage is amplified >100,000 times and the op amp settles to an output voltage when the error voltage is zero. Thus, when the op amp has settled, the voltage at the inverting input is the same as the voltage at the non-inverting input.

The op amp circuit itself is a very fast servo control system.

 

Your typical old analog MFC is a good example of an analog servo.

Before that we used the old motorized needle valve gas flow controllers. They were designed as PID controllers.

120XP

E220
These were later modified to use analog servo flow control valves with gas chamber vacuum as the feedback instead of true mass flow.

The maintenance manuals for these old guys had the complete theory of operation, details schematics and calibration procedures. The E220 series is still running today making lagging edge semiconductor devices.

 

WAY BACK, before transistors were common, I got hold of a TUBE type servo amplifier. The output fed the "control phase" of an AC servo motor. The inputs to the amplifier were the analog control signal and the sensor feedback signal. JUST LIKE the block diagrams.
But the audio gain was way too much and so it would oscillate and arc inside the output tubes.

 

The theoretical servo system, such as a position servo set by means of a knob rotated potentiometer, used to set a command signal such as zzero to ten volts, uses an amplifier with an input similar to a summing junction,.
The positive input is set to some voltage and the difference between the command voltage input and the position feedback input, which varies as the position changes, results in a voltage output that drives a motor. The motor moves the position and also moves tghe position feedback potentiometer in a direction to reduce the difference between the command voltage and the feedback voltage. to "zero".

 

https://sound-au.com/articles/servos.htm

 

Attached is the Servomechanisms chapter from the 1956 NAVPERS Basic Electricity textbook produced by the US Navy to help train technicians who had to maintain servo-based weapons systems. No digital computers even hinted at here, and a nice theoretical discussion.

I extracted the one chapter from the book, but it has some very good material and the complete text is available from The Internet Archive along with many other NAVPERS publications.

 

It’s actually quite simple, I built one myself.

The geared down motor shaft rotates a potentiometer to give a voltage analog feedback for position.

That is compared to a control voltage. If less the motor is rotated one way, if greater the other way. When the feedback matches, it stops. Of course some damping and proportional power are needed. My implementation simply used opamp with gain to drive a bipolar output to the motor. The minimum voltage to make the motor run gave you a dead zone around the set point, and the gain of the opamp controlled how large that was. It worked surprising well with the gain, being the P component of a PID controller. More sophisticated ones would use 3 terms.

 

that is exactly how analog hobby servos work. the only additional thing is that they convert duty (pulse width) to analog setpoint.

 

Another old technique that didn't use active electronics is the Magnetic Servo Amp, which used saturating magnetic cores to provide gain.

 

Another old technique that didn't use active electronics is the Magnetic Servo Amp, which used saturating magnetic cores to provide gain.

We also used Amplidynes on fire-control systems.



https://www.wolvertonbailey.com/amplidyne.htm