The Cyber-Spy.Com Usenet Archive Feeds Directly
From The Open And Publicly Available Newsgroup
This Group And Thousands Of Others Are Available
On Most IS NNTP News Servers On Port 119.
Cyber-Spy.Com Is NOT Responsible For Any Topic,
Opinions Or Content Posted To This Or Any Other
Newsgroup. This Web Archive Of The Newsgroup And
Posts Are For Informational Purposes Only.
From: Chuck Simmons
Organization: You jest.
X-Mailer: Mozilla 4.61 [en] (X11; U; Linux 2.0.33 i586)
Subject: Re: Historical question: negative feedback and the op amp
Date: Wed, 13 Nov 2002 14:19:45 GMT
NNTP-Posting-Date: Wed, 13 Nov 2002 06:19:45 PST
> From: "Kevin Aylward"
> Subject: Re: Historical question: negative feedback and the op amp
> > > I do not think OpAmps come from the negative feedback idea, nearly
> > every
> > > electronic circuit has been using feedback since the beginning of the
> > tube
> > > aera.
> > >
> > > Rather the main idea is to combine a functional block in one unit,
> > which
> > > requires some kind of integration.
> > >
> > > Feedback in itself is more a cure for the unlinearity and other
> > > shortcomings, nowadays it is no more needed,
> > Jesus wept dude. How you can claim this is pretty amazing really.
> > Feedback is most certainly a very usefull device, and imo, will *always*
> > be used.
> I'm an analog designer such as you are and I know very well it will be used,
> but not anymore to improve the circuit linearity.
> digital circuits work with saturated transistors, no feedback there. A FIR
> filter does not use feedback.
I beg your pardon, in an FIR filter, what is the output term(s) doing in
the equation if not feedback?
> An amplification can be done with a comparator, triangle generator (A/D, no
> feedback involved) and 2 MOSFETs to higher voltage and a passive filter
> (D/A, no feedback either).
> Somewhere a control circuit will have to be used with feedback to keep the
> voltages constant but the feedback is not in the signal chain. And also the
> digital amp will have a offset stabilisation etc.
I am sorry but a SAR type A/D definitely uses feedback. Many D/As use
feedback in order to convert from current to voltage.
SAR type A/Ds are going to be around a bit longer because ramp types are
two slow and flash types consume real estate at an alarming rate (worse
than memory and people throw rocks at me for asking for more memory).
> Control systems per se work *by principle* with feedback. A balance is a
> good example.
> But if a mathematician multiplies by 100 he cannot improve his result by
> dividing it by 100 and compare to the input. He will just waste his time and
I don't know how you design digital control systems but when I design
them, the feedback is explicit. The input number from the A/D is
compared to the goal and a command to the motor is computed using an FIR
compensator and sent to a D/A. I also use an integrator in control
loops. Except for rounding consequences, it doesn't matter where I put
the integrator in the loop. The integrator provides feedback to remove
tracking offsets and to improve low frequency performance. Sorry about
the offsets. I can't help it. Real power amps and honest to god motors.
Even a speed loop needs an integrator because exogenous inputs can
produce steady state errors even though the speed loop is fully digital
and has no offset. Again, the feedback is explicit. The time measurement
to determine speed is compared with a goal calculated from external
data. The error is then feed to an FIR filter (compensator) and then
goes to a D/A. Lots of feedback in that.
> > >it
> > > wasts power and a *lot* of speed. It is just a crutch not something we
> > > should cherish.
> > I don't why you say this. Sure using analogue feedback has some
> > negatives, but in most cases the advantages easily outweigh the
> > disadvantages. Try and make a reliable fixed gain of 100 without
> > feedback. This whole concept is not even debatable. Feedback is one of
> > the most important pieces of science around.
> Ok we needed the crutch, and it is helpful if you have deficiencies. But now
> almost everything is done digitally. And the real progress is done by
> inventing and refining algorithms.
> Sure there are people like you who will give the IT guy the chips he needs
> to implement his algorithms.
> And the elderly analog specialists for the sensor interface.
> But as you have seen, analog wise there is hardly any progress, not like in
> the 70s and 80s where everyday new ICs were produced.
How do you propose to get rid of the vagaries of sensors and motors?
Suppose you use a CCD sensor to replace a split photodiode in a control
system. You are still faced with gain variation, leakage variation and
other little problems that make the system less than perfect.
BTW, for shaft positioning, there is an essentially perfect sensor
available (these have been around for maybe 50 years). It has high
absolute accuracy and no offset. It is called an absolute brush type
shaft encoder. These are astronomically expensive. They achieve near
perfection by being almost purely mechanical.
... The times have been,
That, when the brains were out,
the man would die. ... Macbeth
Chuck Simmons firstname.lastname@example.org
Go Back To The Cyber-Spy.Com
Usenet Web Archive Index Of
The sci.electronics.design Newsgroup