Hi guys, I've a project on amplifier stability and compensation. I'm
about half way through it now, anyboody got any info or anything
interesting. It's a fairly simple project, what i have to do is complie
notes that will be used to teach 3rd year electronics students about
amplifier stability.
The next stage is where i create four one-hour lab session to
demonstrate how to;
calculate the loop-gain,
show how the loop gain varies with frequency,
show how a amplifier becomes unstable,
demonstrate how to stabilise it.
I haven't completely decided on these labs so if anybody thinks they'd
do it differently please let me know.
I've already decided that i'm going to use simple op-amp circuits in
these labs. What i need to do is create these circuits. Does anyboddy
have any examples / suggestions/ ideas on what circuits would suit the
labs?
I would be inclined not to start out with amplifiers or op-amps. They are
designed to be stable in the first place.
Why spend time taking something that is stable and then break it to prove a
point?
What you, and they, should be more interested in is 'feedback systems'. The
whole point of stability is to take an open loop system and then close a
loop around it to achieve a stable system whose performance is predictable.
The op-amp is a tool used to close such a loop. It's response, as an error
amplifier, is tailored according to the gain response of the original system
to achieve this.
But I wouldn't start there.....
The first thing I compensated was a phase locked loop, CD4046, op-amps not
included. I monkeyed about with the piece of shit trying to make it work,
even read AoE but did so superficially.
Did it want to play..... no it did not.
For the first time in my life I sat down and read through the words
properly. Then I did the sums. Then I applied the results and it worked.
I was, sort of, taught about loop stability on my Applied Science degree.
They used op-amps and bode plots and wein bridge oscillators but it didn't
sink in. Well perhaps some of it did subconciously.
Things started to come together when I had to do it myself.
So, that's your first lab.
Two weeks before.
Tell them they are going to be using a 74HC4046. Give them a copy of the
relevent section from AoE (I won't tell Win or Paul if you don't..... just
mention where it came from).
Tell them they are going to multiply the frequency of an input signal by a
number between 1 and 16.
Ask them to deliver a written report one week before the lab that says how
they are going to make it work and how they would change it so it doesn't.
Say why it does and why it doesn't.
Tell them they can cheat and talk to each other about it. The reports can be
singular or a collaboration, list the contributors.
Make sure that they understand it's a 'hands up if you didn't understand'
situation.
Any questions?
Tell them that four days in there will be a meeting for
questions/clarification.
Do the meeting.
Take the report(s) at one week and set another meeting for three days.
Read what they have written. Learn from 'your' mistakes.
Take the next meeting and tidy up loose ends, whatever they may be...
Do the lab.
Ask for reports one week later.
Read them, where did you go wrong?
Meet again for a revue.
DNA