yes!

If you parameterize using the LC values, I believe that. It’s very difficult to tune a high-order filter unless you start out pretty close.

Parameterizing f_0 and Q for each section works much much better.

Cheers

Phil Hobbs

Some intuition might make us rebel against the idea that a suitably programmed classical deterministic computer could be conscious. But we don't know what consciousness is, which makes it rather difficult to formulate a reasoned argument against its existence in such a computer.

Sylvia.

I only got a couple of pages in AoE3, but I did better in the X-chapters. I made the preface (with Phil Hobbs) and am in the index at the end, and I think I'm named about 22 times between. I don't recall seeing your name.

What difference does the design process make, if the result works?

I've always annoyed PhD academic types who resent people who are creative and have instincts.

The H+H books are deliberately call The ART of Eletronics. I think higher education, especially the PhD process, beats the creativity out of people. I recently had to fire a PhD; she thought that being a PhD made her right, which it didn't.

Neo-Darwinian evolution is crazy inefficient. Why wouldn't we evolve a better way for evolution to work? The critters that did ate the critters that didn't.

Active filters are easier, where the sections don't interact.

Sounds about right. Convergence depends on the existence of even a tiny average gradient.

But if one keeps at it, eventually a lucky jump will occur. It may take a million years to get anywhere useful.

Joe Gwinn

Yeah, you just pick the poles and zeros. Of course the GBW has to be several times f_0*Q.

Cheers

Phil Hobbs

Your stuff sells under your name. Mine sells too, but not under my name.

Thought-out designs work better and cost less.

Academics don't resent people who are creative. They aren't fond of people who don't talk about how they get to their results, because academics are in the teaching business - academies are where people learn stuff - and if they can get hold of approaches that work, they can pass them on to their students.

Instincts aren't teachable and you can't pass them on - sometimes your kids inherit them. There are quite a few people who don't like passing on their skills, and consequently claim that they operate by instinct, so that they won't have to educate potential competitors.

There are academics who study all kinds of creative arts - painting, sculpture and music all have their academies.

Knowing what you are doing, and how it fits in with what other people have done, doesn't seem to stifle creativity.

Getting a Ph.D. isn't any kind of training in being creative, and most of the ones I've known haven't been all that creative. I got together with my wife before she did her Ph.D. training, and it didn't beat any creativity out of her - she was an original thinker before she started (despite having aready got a Master's degree) and she stayed creative thereafter (and wrote lots of highly cited papers along the way).

Language probably is that better way. We can now talk about individual defects in particular genomes, and use CRISPR to correct some of them.

It's taken more than three billion years to get this far, which would be crazy inefficient if we'd known where we were going.

Efficiency in this context, would the ratio of the time a perfect system would take to the time our actual system has taken. Since we don't know what a perfect system would be it's an imaginary number.

The Nuhertz software designs amazing LC filters using standard-value Ls and Cs. Fast. I don't understand how that is even possible.

Nuhertz is a product of Ozen Engineering, who ain't talking directly about this. But it seems to be that the speedups come from better choice of what exactly to optimize, and especially the equivalent of better mesh generation (for finite-element design).

I'll mull it over.

Joe Gwinn

...

Which pages?

Around 360.

Ok since posting the question I discovered that you're mentioned on pages xxx, 294, 360, 524

Bill Sloman should probably not read page 360.

His whining centers on my inability to explain how I design electronics, or where ideas come from.

Sorry, I don't know. It just happens. If invention happened from definable algorithms, everything would be invented all at once.

Not sure I agree but I don't speak for him.

I think that's a bit like saying that a program which implements the quadratic formula solves all quadratics all at once.

One of the things I have learned is to stay confused for a while, stagger around in the solution space for a few days at least.

Too many engineers dislike uncertainty, so they lock down a design, preferably a textbook-sanctioned design, asap so they can implement.

Sounds reasonable to me. And it applies to any form of art. Here's a silly example.

Back when I used to read magazines like this one

I preferred Popular Electronics myself.

Winfield Hill used to post here a lot, and is probably aware of John Larkin's appetite for flattery."To which we've added a few decorations" dilutes the endorsement.

"Whining"? That isn't what I'm complaining about - John Larkin doesn't explain what his circuits are intended to do or what problems their - presumably unique - features are intended to deal with,

Design is all about using what you can get to do what you need to do, and a useful conversation about circuit design has to be specific about both the problems being dealt with and the way the approach adopted solves them.

So he just stumbles across his solutions, and doesn't know why they actually work. That isn't design.

I have put circuits together that worked better than I expected, but I then put a lot of effort into finding out what was actually going on, so it didn't stop working in the middle of demonstration.