want help learning practical circuit design

On a sunny day (Thu, 21 Feb 2008 08:19:04 -0800) it happened John Larkin wrote in :

We are, after all, neural nets. Theory, or in this case understanding of basics, is very important however. I do not 100% agree with 'random solutions', but I do think that we, in our neural net, have individual structures of how we think things work, and what comes out given a certain input, is the sum of all those structures, previous experiences, that is why we see so many people here sometimes with different solutions to the same problem. Interesting is that 'neural net' is not exact numerically exact in it's output, I will go so far as to say that the math hardware (hardwired) that grew in our brain during the learning process is only a subset, personal architecture, and not exact either. I really pissed of a well known mathematician when I pointed that out some time ago.

So from this is seems: The more people, the more different solutions. We, are, neural net is, 'intuition'.

Jan,

You keep going on about being a neural net. It seems that your training algorithm omitted any of the ~2500 year long discussion (still going strong) about the mind-brain problem. It really isn't as simple as that.

Cheers,

Phil Hobbs

On a sunny day (Thu, 21 Feb 2008 14:41:26 -0500) it happened Phil Hobbs wrote in :

Hi, Phill, we *are* neural nets. Philosophy will not change that, ever. I understand that perhaps for _some_ the fact that we are not more then that, is as intimidating as the fact that we are not at the centre of the universe.

That what you call 'you', or 'me', is just a small subsection (some neurons), that build some world picture of you (you think you are an engineer perhaps, and that you are good at this, and that, and perhaps 'own' a lot of things. And some time you spend maintaining and enhancing that 'ego' part. In essence, when you sleep for example, where is that 'you'? People like to assign a higher 'something' to consciousness, however even a simple solar sensor, that controls a shade, is _conscious_ of light. We are but a machine. It is possible there is some universal 'something' that pervades everything, and while we are alive makes us tick, but is essence that is hard to point to. We like (us humans), to call anything beyond what we really have an explanation for, 'God'. Like the medicine man in the tribe in the 4th world country dances to please the rain gods, he assigns powers to the rain god, and likely uses it to prop up his position in the tribe, _we_ know about clouds, thunder, electricity, water vapour, have weather radar. For him the 'why' of it is beyond his horizon, he assigns the label 'God' to it. For us we use science to explain it. And yet this same species goes and prays for the storms to quiet down, the floods to stop. Human species, in all its glory. So it is not years long discussions that count, it is YOUR understanding, and if you want to understand 'you', the bigger picture, not just the ego part , you will need to look at 'you'. I recommend doing meditation, as that will make you look inside. What you find will be *your* understanding, it cannot be put onto you as normal lecturing. It is the real challenge, as it is about reality, not discussions and philosophy.

Interesting subject, and I can go on about this for a very long time, many things in this world are exactly the opposite as people think. In my view it is of the most importance, before studying any other subject, to study your own neural net, as, to make a link back to somebody 2000 years ago, the 'Kingdom of Heaven is Within'.

Yeah, but she's always doing some abbreviated and marginal hack of the interface.

Maybe you are. I'm not.

John

explanation for, 'God'.

the rain gods,

in the tribe,

it.

floods to stop.

if

will

normal lecturing.

philosophy.

things

ago,

I wasn't making a religious point, just a philosophical one. Plato knew all about the mind-brain problem.

Cheers,

Phil Hobbs

Well, it doesn't have step-by-step soldering, but it's got practically everything else. :-)

I've also gotten some use out of a couple of Don Lancaster's cookbooks. :-)

Cheers! Rich

Of course, you'll be needing a proofreader/transcriber, right? ;-)

Cheers! Rich

Those who recommended 'hands on' learning and some good books provided useful replies. Another to learn important basics was the "CMOS Cookbook". But the first thing to do is build a power supply. For example, a transformer, diodes, capacitor and the 7805 power supply chip is a perfect first project. Circuit will be available in the 7805 power supply chip's data sheet and in application notes for that IC. A simple project, easy to debug, that leaves you with a tool for other circuits.

I never did, but would like to. I did a couple of phone interviews for a company that makes welders. I was kinda looking forward to taking their classes. Didn't work out though...

Ouch!

His Active Filter Cookbook is handy when you need a basic filter. Williams for the more serious stuff.

John

to

A few others already pointed out the benefits of simulation, and you already know from software the importance of mastering primitives, one at a time, so that you can build your arsenal for synthesizing systems, and you probably realize that "premature optimization is the root of all evil" - Donald Knuth - so it is ok to not always find the most minimally perfect circuit. :)

However, one thing that helped me greatly was to stop thinking of electronics in terms of formulas. By this I mean, stop thinking about signals and systems, small and large-signal models. Maxwell's equations, althogth they are intuitive at one level in their own right, can even be the culprit in blocking intuition. The equations are true, of course, but the stickyness of what you learn depends on intuition, and IMHO, that can only be gain by throwing away the math and actually thinking about the physical processes.

The most important change for me was to reassess from a charge distribution point of view, not state variables. Charge distribution is is introduced in electrostatics and electrodynamics, but only to solve field problems. I learned that if you abstain from reachnig for the formuals but asking instead, "What is my charge distribution?" for the entire system, repeatedly, you slowly gain more insight than you would simply from relying on the math. I think it is unfortunate that basic EE courses are not taught from a chage distribution point of view. And by this, I do not mean field problems. I mean thinking about charges as they move, and how they interact.

For example, a voltage V exists across as resistor R with specified- but-weird dimensions. What is the charge distribution inside and outside the resistor? Many EE's can write down I=3DV/R, but when you start asking "What is my charge distribution?", it gets very interesting very quickly depending on the geometry. You can apply eletrostatics, but if you take multiple resistors of different dimentions and sandwich them together, it gets even more interesting. If you think long enough, you eventually arrive at the border of quantum physics. How far you go then depends on how much time you have. I have not very much, so I stop at comfortable abstraction barrier and not look beyond it.

Thinking about charge distribution in a PN junction or MOS structcture also helps quite a bit.

You can do the same with a radiating antenna. "What is my charge distribution in the antenna?"

If you ask this question about every component, and then fidget endlessly until you are confident of the answer, you'll gradually develop the same intuiton that you've developed in programming. Of course, experts in this group already have that intuition, which they gained from either raw ability, decades of practice, or both. If you are a young person and would like to cheat a bit, the charge distribution point of view might help.

On a more pratical note, I found oscillators in my early career, using both discretes and OP AMPs to be a lot of fun.

I remember the joy of developing my first tone generator from a cascade of OP AMPS with various features such as volume control, frequency, ramping, warble...I sat looking at the circuit for over and hour scratching my head trying to figure out why it made no sound. I checked my math over and over...poles on j-omega axis...gain stage looks right...buffer is right...no clipping..argggg!!!...what's wrong with my circuit!!!????....

Only when my girlfriend came to look over my shoulder and i started showing her how it worked did I realize the problem. I said, "...and finally, the purified signal comes out the speaker here, which is connected by these wires....these wires...that...are not connected to the rest of the circuit.:("

Those moments are priceless, and if you do it without a cookbook, you will never forget.

-Le Chaud Lapin-

I dunno... from his usenet posts, I think Larkin is that anomalous creature that's confused anthropologists for years: the mythical engineer who can write.

A modern day missing link, if you will.

:-)