Shameless Plug

On Tue, 30 Nov 2004 10:59:17 -0800, Tim Wescott
<tim@wescottnospamdesign.com> wrote:

>Guy Macon wrote:
>
>> Tim Wescott wrote:
>> 
>> 
>>>Guy Macon wrote:
>>>
>>>
>>>> "Why do some controller boards have an option to reverse the
>>>> phase of the D?  What is that good for?"
>>>
>>>OK, I'll bite -- what _is_ it good for?  I've never done closed-loop 
>>>control with prepackaged controllers and I've never seen that done 
>>>elsewhere.  I can certainly see reversing the phase of the whole thing, 
>>>or reversing the phase of the D term if it's coming from some other 
>>>feedback source (which would imply a second input) but I _can't_ see the 
>>>point in intentionally establishing an unstable zero in your control system.
>> 
>> 
>> My experience is more hands-on than theory, but here is the answer 
>> I gave my classes:
>> 
>> In N years of setting up servos, I have never once found a use for
>> it, nor have I found any literature that explains when it might be
>> of some use.  I think that somewhere back in the early days someone
>> was told to put in a jumper that reverses the phase of the entire 
>> servo (quite handy when someone miswired a section that is really 
>> hard to get to), got it wrong, and some other manufacturers have 
>> been copying the "feature" ever since.
>> 
>> If one of the theory boys has a better answer, I am all ears.
>> 
>Being able to reverse the sign of the whole thing is good -- one of the 
>old curmudgeonly engineers from whom I learned practical control liked 
>to say that when designing one of these things you should count up all 
>the sign changes in the loop -- then throw in one extra for the one you 
>missed.  His circuits always had at least one spot where you could 
>rearrange the inputs to an op amp and reverse the sense of a signal.  I 
>follow that now:  there's always at least one place in my software where 
>one can insert a '-' and change the sign of the whole thing.

All controllers I've seen, be they in software or hardware, have a
flag or switch to reverse the controller action. Some even let you
reverse the output action. Eases up a lot in initial implementation.

I've seen the ones that let you reverse the terms individually but
haven't found a need for trying them out. One accidentally had the
integral term reversed and I went bonkers trying to tune it before
finding the error.

The derivative term  is useful when you have a slow process with
inertia. Once the controlled variable starts to respond it puts a lid
on further controller action which might cause unacceptable overshoot.
On the few occasions I do apply it it's usually in homeopathic doses.

- YD.

-- 
Remove HAT if replying by mail.

Guy Macon wrote:
> 
>  "Why do half the engineers call it Proportional-Integral-Derivative"
>  and others call it "Proportional-Integral-Differential?"  When I 
>  did a Google search on "proportional integral differential" I got
>  18,600 hits while "proportional integral differential" only had
>  3,060 hits, but most of the "proportional integral differential"
>  hits seem to be by scientists and equipment manufacturers. 
>  Which is correct?" 
> 

The "D" part of PID refers to the presence of a rate-of-change term, 
often able to be interpreted as a derivative of a position (or similar) 
quantity.  Derivatives are a concept deriving (!!) from the differential 
calculus.  Either term refers to the notion of infinitesimal differences 
whose ratios are considered as they approach (under suitable existence 
conditions) some limiting value.

Thus either version of the acronym is fine.
Elsewhere in this thread is a timely reminder about the danger of 
reliance upon statistics (eg: Google hits) to "prove" something.
I've come to the conclusion that if the majority agree upon something, 
it's probably false (works great in the stock market, but you have to be 
careful about your audience when defining the principle as "the fallacy 
of democracy"......<grin>)

Geoff.

On Tue, 30 Nov 2004 10:59:17 -0800, Tim Wescott wrote:

> Guy Macon wrote:
> 
>> Tim Wescott wrote:
>> 
>> 
>>>Guy Macon wrote:
>>>
>>>
>>>> "Why do some controller boards have an option to reverse the phase of
>>>> the D?  What is that good for?"
>>>
>>>OK, I'll bite -- what _is_ it good for?  I've never done closed-loop
>>>control with prepackaged controllers and I've never seen that done
>>>elsewhere.  I can certainly see reversing the phase of the whole thing,
>>>or reversing the phase of the D term if it's coming from some other
>>>feedback source (which would imply a second input) but I _can't_ see the
>>>point in intentionally establishing an unstable zero in your control
>>>system.
>> 
>> 
>> My experience is more hands-on than theory, but here is the answer I
>> gave my classes:
>> 
>> In N years of setting up servos, I have never once found a use for it,
>> nor have I found any literature that explains when it might be of some
>> use.  I think that somewhere back in the early days someone was told to
>> put in a jumper that reverses the phase of the entire servo (quite handy
>> when someone miswired a section that is really hard to get to), got it
>> wrong, and some other manufacturers have been copying the "feature" ever
>> since.
>> 
>> If one of the theory boys has a better answer, I am all ears.
>> 
> Being able to reverse the sign of the whole thing is good -- one of the
> old curmudgeonly engineers from whom I learned practical control liked to
> say that when designing one of these things you should count up all the
> sign changes in the loop -- then throw in one extra for the one you
> missed.  His circuits always had at least one spot where you could
> rearrange the inputs to an op amp and reverse the sense of a signal.  I
> follow that now:  there's always at least one place in my software where
> one can insert a '-' and change the sign of the whole thing.

Stumbled on this while looking for something entirely different -
something about heavy boots on the moon leading to physics humor, over on
rec.puzzles - anyway, when I hit this on the physics joke page, I thought
of this post.
http://www.xs4all.nl/~jcdverha/scijokes/2.html#12

Cheers!
Rich

OTOH ...

One of the early power system simulation packages developed (I think) by 
IBM (dates from the days of punched cards so that gives an indication) 
had a reasonably well documented (for those days) program, luckily...

In the main body of the program prior to calling a subroutine to 
simulate the voltage regulator was a line of code that reversed the sign 
of a variable, with a brief note to state that the standard equation 
assumed the quantity was positive when it was actually negative (or 
something). Immediately after the start of the subroutine was a line of 
code that reversed the sign of the same variable, with a brief note to 
state that the standard equation assumed the quantity was positive when 
it was actually negative.

We found this after several days of wondering why the simulation of a 
power system transient indicated an unstable voltage regulator.

But I guess if you get the action of controller wrong it's nice to be 
able to reverse it very quickly - hopefully before the operators notice!

Bruce

Tim Wescott wrote:
> Guy Macon wrote:
> 
>> Tim Wescott wrote:
>>
>>
>>> Guy Macon wrote:
>>>
>>>
>>>> "Why do some controller boards have an option to reverse the
>>>> phase of the D?  What is that good for?"
>>>
>>>
>>> OK, I'll bite -- what _is_ it good for?  I've never done closed-loop 
>>> control with prepackaged controllers and I've never seen that done 
>>> elsewhere.  I can certainly see reversing the phase of the whole 
>>> thing, or reversing the phase of the D term if it's coming from some 
>>> other feedback source (which would imply a second input) but I 
>>> _can't_ see the point in intentionally establishing an unstable zero 
>>> in your control system.
>>
>>
>>
>> My experience is more hands-on than theory, but here is the answer I 
>> gave my classes:
>>
>> In N years of setting up servos, I have never once found a use for
>> it, nor have I found any literature that explains when it might be
>> of some use.  I think that somewhere back in the early days someone
>> was told to put in a jumper that reverses the phase of the entire 
>> servo (quite handy when someone miswired a section that is really hard 
>> to get to), got it wrong, and some other manufacturers have been 
>> copying the "feature" ever since.
>>
>> If one of the theory boys has a better answer, I am all ears.
>>
> Being able to reverse the sign of the whole thing is good -- one of the 
> old curmudgeonly engineers from whom I learned practical control liked 
> to say that when designing one of these things you should count up all 
> the sign changes in the loop -- then throw in one extra for the one you 
> missed.  His circuits always had at least one spot where you could 
> rearrange the inputs to an op amp and reverse the sense of a signal.  I 
> follow that now:  there's always at least one place in my software where 
> one can insert a '-' and change the sign of the whole thing.
> 

["Followup-To:" header set to sci.electronics.design.]
On Tue, 30 Nov 2004 08:59:24 -0800,
  Tim Wescott <tim@wescottnospamdesign.com> wrote
  in Msg. <10qp9n47qbbp076@corp.supernews.com>

> I fear that my mind was poisoned long ago by a German instructor who 
> pointed out that modern linguistic theory doesn't much recognize a 
> "right way" and a "wrong way" -- it just records prevalent usage, and 
> tries to keep out of the way of the steamroller.

This reminds me of one of my favorite entries in Strunk & White, Modern 
English Usage on "flammable" vs. "inflammable". The correct term is 
inflammable, but on trucks that hold dangerous goods you'll always see 
"flammable". Quoting from memory: "Unless you drive such a truck, and are 
hence concerned with the safety of children and illiterates, use 
inflammable".

According to my pedantic mind, there's no such thing as a flammable 
substance, but the general public seems to think otherwise.

--Daniel

Geoff wrote:
>
>Guy Macon <http://www.guymacon.com> wrote:

Please don't follow "Guy Macon wrote" with something that I
clearly labeled (in the part you snipped) as being not my
question, but rather a question typical of a student in a
class I taught. 
 
>>  "Why do half the engineers call it Proportional-Integral-Derivative"
>>  and others call it "Proportional-Integral-Differential?"  When I 
>>  did a Google search on "proportional integral differential" I got
>>  18,600 hits while "proportional integral differential" only had
>>  3,060 hits, but most of the "proportional integral differential"
>>  hits seem to be by scientists and equipment manufacturers. 
>>  Which is correct?" 
>
>The "D" part of PID refers to the presence of a rate-of-change term, 
>often able to be interpreted as a derivative of a position (or similar) 
>quantity.  Derivatives are a concept deriving (!!) from the differential 
>calculus.  Either term refers to the notion of infinitesimal differences 
>whose ratios are considered as they approach (under suitable existence 
>conditions) some limiting value.
>
>Thus either version of the acronym is fine.

So you derive from deriving the derivative from the differential that
one should not differentiate between differential and derivative?
That's different.

>Elsewhere in this thread is a timely reminder about the danger of 
>reliance upon statistics (eg: Google hits) to "prove" something.

Google hits prove commonness of usage on the World Wide Web.
All else is derivative - an important difference.

Guy Macon wrote:

  ...

> So you derive from deriving the derivative from the differential that
> one should not differentiate between differential and derivative?
> That's different.
> 
> 
>>Elsewhere in this thread is a timely reminder about the danger of 
>>reliance upon statistics (eg: Google hits) to "prove" something.
> 
> 
> Google hits prove commonness of usage on the World Wide Web.
> All else is derivative - an important difference.

Bravo!

Jerry
-- 
Engineering is the art of making what you want from things you can get.
�����������������������������������������������������������������������

Guy Macon <http://www.guymacon.com> wrote in message news:<10qoea0ci864ef2@corp.supernews.com>...
> Tim Wescott wrote:
> >
snip
> I also found it helpful
> to show how to use a stopwatch and odometer to derive speed with 
> no speedometer, a stopwatch and speedometer to derive distance 
> without an odometer, and a speedometer and odometer to derive
> elapsed time with no stopwatch.  Your audience is different, 
> of course - this worked really well with mechanical engineers,
> but software engineers are quite different.
> 

Maybe it's just me, but shouldn't this be obvious 
to anyone who's had even basic physics in school?


-Lasse

Lasse Langwadt Christensen wrote:

>Maybe it's just me, but shouldn't this be obvious 
>to anyone who's had even basic physics in school?

Only those who went to school back when they were still teaching
how to apply basic physics to real-world problems.

To be fair, some schools do a great job of this, but I have 
personal experience of a person who got an EE degree from a
state college without ever tumbling on to the fact that when
you send current down a wire there has to be an equal current
through a return path.  :(

That engineer was put to work maintaining COBOL programs.
This was in the '90s, not in the age of COBOL.

Guy Macon wrote:
> Please don't follow "Guy Macon wrote" with something that I
> clearly labeled (in the part you snipped) as being not my
> question, but rather a question typical of a student in a
> class I taught. 
>  

My apologies.

> 
> So you derive from deriving the derivative from the differential that
> one should not differentiate between differential and derivative?
> That's different.
> 

Essentially yes.  (Ignores irony). But then, I'm not only a 
mathematician, I'm also a linguistics freak....

> 
> Google hits prove commonness of usage on the World Wide Web.
> All else is derivative - an important difference.
>

<grin>
As I intimated, "common usage" is to be distrusted.  After all, the 
planet's population is now so large that virtually any human-behavioural 
parameter, via the central limit theorem, gets modelled as obeying a 
Gaussian distribution, whose *central* area dominates the sample results.
I call the universal welcome currently accorded to this situation "the 
cult of mediocrity" and it is an example of positive feedback.
Examples abound.  Think about it.

The linguistics scene has "descriptive grammarians" (currently in the 
ascendant) versus "prescriptive grammarians" (started declining maybe 50 
years ago).  That's why I regularly find books, and even learned papers, 
which confuse "throes" with "throws", "pour" with "pore", and many more, 
since schools ceased to bother students with (horror!) rules, 
substantive examinations etc.

To pull things together:  my derivative/differential fusion is based 
upon a return to fundamentals (mathematical and linguistic).  I find 
that this approach is superior to all others I've tried.
YMMV.

Geoff.