Control of Gun Mounts and Negative Feedback Systems

[From Fred Nickols (2015.01.31.1523.EST)]

[From Fred Nickols (2015.01.31.1454 EST)]

I’ve been trying to follow the homeostasis thread. Not sure I’m doing so well but it does raise a question; namely, whether or not a system that uses negative feedback is necessarily a control system. I have in mind the gunfire control systems I tended to when I was a technician in the Navy. So I changed the subject line.

One of the pieces of equipment I tended to was the computer. It calculated gun orders which served to position the gun mount. The circuits controlling the gun mounts were clearly what we typically mean by a control system, i.e., the gun orders (train and elevation) were the reference signals, the actual position of the gun mount were the perceptions, and the amplifiers and motors that controlled the gun mount treated any difference as a discrepancy and operated the motors so as to bring actual gun mount position and ordered gun mount position into correspondence.

However, in the computer, there were also servomechanism with amplifiers, etc., and in some cases negative feedback did not constitute a control system; at least not in the sense I understand it. To be specific, the gun order modules themselves were controlled by servos driven by signals from small amplifiers. The amplifiers made use of positive and negative feedback. The positive feedback served to increase the torque so that the gun order module moved very quickly and with consider strength. The negative feedback served to stabilize that movement so that it was jittery (no sense in jerking around a five ton gun mount). I don’t think the presence of the negative feedback in the gun order amplifiers constitutes a control system all by itself.

Am I wrong?

Regards,

Fred Nickols

Distance Consulting LLC

www.nickols.us

Regards,

Fred Nickols, CPT

Distance Consulting LLC

Assistance at a Distance

The Knowledge Workers’ Tool Room

“Be sure you measure what you want.”

“Be sure you want what you measure.”

[From Rick Marken (2015.02.01.0830)]

···

Fred Nickols (2015.01.31.1454 EST)–

FN: I’ve been trying to follow the homeostasis thread. Not sure I’m doing so well but it does raise a question; namely, whether or not a system that uses negative feedback is necessarily a control system.

RM: Don’t get confused Fred. If it controls it’s a negative feedback system. If it doesn’t control it’s not.

FN: I don’t think the presence of the negative feedback in the gun order amplifiers constitutes a control system all by itself.

FN: Am I wrong?

RM: It’s only a control system if it controls some variable. If the gun order amps controlled some variable then they were negative feedback systems; if they didn’t control some variable then they were not.

RM: Remember. Phenomena phirst! If a variable is not being controlled by a system then there is no reason to look for an explanation of the behavior of the system in terms of negative feedback.

Best

Rick

Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble

[From Fred Nickols (2015.02.02.1238 EST)]

Thanks, Rick. Allow me a shot at clarifying something.

All control systems make use of negative feedback.

Not all systems making use of negative feedback are control systems.

Are those two statements true? I think they are.

Fred

···

From: Richard Marken (rsmarken@gmail.com via csgnet Mailing List) [mailto:csgnet@lists.illinois.edu]
Sent: Monday, February 02, 2015 11:28 AM
To: csgnet@lists.illinois.edu
Subject: Re: Control of Gun Mounts and Negative Feedback Systems

[From Rick Marken (2015.02.01.0830)]

Fred Nickols (2015.01.31.1454 EST)–

FN: I’ve been trying to follow the homeostasis thread. Not sure I’m doing so well but it does raise a question; namely, whether or not a system that uses negative feedback is necessarily a control system.

RM: Don’t get confused Fred. If it controls it’s a negative feedback system. If it doesn’t control it’s not.

FN: I don’t think the presence of the negative feedback in the gun order amplifiers constitutes a control system all by itself.

FN: Am I wrong?

RM: It’s only a control system if it controls some variable. If the gun order amps controlled some variable then they were negative feedback systems; if they didn’t control some variable then they were not.

RM: Remember. Phenomena phirst! If a variable is not being controlled by a system then there is no reason to look for an explanation of the behavior of the system in terms of negative feedback.

Best

Rick

Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.

Now available from Amazon or Barnes & Noble

[Martin Taylor 2015.02.02.13.07]

As you know, I can't speak for Rick, but I'd like to clarify

something in my own voice.
As general statements your first statement is true. The second is
not; a control system exists only in one special case of negative
feedback loop.
A feedback loop has between zero and N places where influences from
the outside world combine with influences circulating in the loop.
The case of zero is uninteresting. The case of 1 is the equilibrium
system in which the energy to restore the resting state is provided
by the action that moves it from the resting state.
The case of 2 includes control systems, but not all negative
feedback systems that have two places where external influences
interact with loop influences are necessarily control systems,
depending on how one defines control systems. With the usual
definition, to be a control system requires that there is
amplification powered by an external energy source, and that the
amplification is greater on one branch between these two places than
in the other. One could, however, say that this is the definition of
an effective control system, and that any negative feedback loop
with two places of interaction between external and internal
influences is, by definition, a control system. This is a question
of language usage.
The case of N>2 is a more general homeostatic system, in which
one might say that N-1 variables are controlled (the -1 comes in
because the control of the others determines what happens at the
last one, just as control of the perception in a 2-place control
system determines what happens at the environmental correlate of the
perceptual function. If you want to call these N>2 systems
control systems, there is an argument for doing so, but I think it’s
not a good one, since their structure and behaviour is different
from that of a PCT-standard control loop unless one particular
quasi-perception is in some way favoured by the distribution of
energy sources and amplification around the loop. Biological
homeostatic systems, so far as I know, have energy supplies all
around their loops, and a variety of possible influences on their
variables, so they wouldn’t look like PCT-type control systems. The
greater the value of N, the less like a control system is the
negative feedback loop.
Bottom line–Are negative feedback loops with N places where
external influences interact with loop influences all control
systems, where
N=1? No.
N=2? Yes or no, depending on how you define a control system.

···
        [From

Fred Nickols (2015.02.02.1238 EST)]

        Thanks,

Rick. Allow me a shot at clarifying something.

        All

control systems make use of negative feedback.

        Not

all systems making use of negative feedback are control
systems.

        Are

those two statements true? I think they are.

Fred

[From Rick Marken (2015.02.02.1610)]

···

Fred Nickols (2015.02.02.1238 EST)–

Thanks, Rick. Allow me a shot at clarifying something.

FN: All control systems make use of negative feedback.

RM: Yes. Indeed, only systems organized as negative feedback systems can control.

FN: Not all systems making use of negative feedback are control systems.

RM: The only way this could be true is if the negative feedback loop is in some way defective. If the loop gain is too high a negative feedback system it will be unstable and oscillate, so it’s an example of a negative feedback system doesn’t control. I would still call it a control system; just one that has to be adjusted.

FN: Are those two statements true? I think they are.

RM: They can be if you are talking about negative feedback as the effect of the output of a system on the cause of that output. So negative feedback exists when o = f(e) and e = -g(o) so that o = f (-g(o)). With negative feedback the output variable is on both sides of the equation.

RM: If o = f(e) but e is not a function of o then there is no negative feedback. This is the case in equilibrium systems such as the mass spring system where the restoring force F = -kx but x is not simultaneously a function of F. Calling the restoring force “negative feedback” is just using the term in a metaphorical way.

Best

Rick

Fred

From: Richard Marken (rsmarken@gmail.com via csgnet Mailing List) [mailto:csgnet@lists.illinois.edu]
Sent: Monday, February 02, 2015 11:28 AM
To: csgnet@lists.illinois.edu
Subject: Re: Control of Gun Mounts and Negative Feedback Systems

[From Rick Marken (2015.02.01.0830)]

Fred Nickols (2015.01.31.1454 EST)–

FN: I’ve been trying to follow the homeostasis thread. Not sure I’m doing so well but it does raise a question; namely, whether or not a system that uses negative feedback is necessarily a control system.

RM: Don’t get confused Fred. If it controls it’s a negative feedback system. If it doesn’t control it’s not.

FN: I don’t think the presence of the negative feedback in the gun order amplifiers constitutes a control system all by itself.

FN: Am I wrong?

RM: It’s only a control system if it controls some variable. If the gun order amps controlled some variable then they were negative feedback systems; if they didn’t control some variable then they were not.

RM: Remember. Phenomena phirst! If a variable is not being controlled by a system then there is no reason to look for an explanation of the behavior of the system in terms of negative feedback.

Best

Rick

Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.

Now available from Amazon or Barnes & Noble


Richard S. Marken, Ph.D.
Author of Doing Research on Purpose.
Now available from Amazon or Barnes & Noble