PnP WoW

You might go easy on the "inventing" aspect of electronics. People tend to think that invented circuits are sufficiently novel to be worth patenting, and the process of turning a novel idea into a patent is all lawyers and has very little to do with electronics.

Working out what kind of circuit your are going to design should be a process of finding an existing circuit that you can modify for your application. That way you can be much more confident that the circuit will work. Having to inventing something is always a desperation move.

EMI Central Research invented a lot of stuff but didn't have any formal procedures for doing it that I ever ran into.

ASML is Dutch and a Philips spin-off, and you can study "policy" as a university subject in the Netherlands, so they have people whose job it is to write policies. They are as useless as their English language equivalents.

The idea of reviewing designs at the conceptual stage, and then again when you have detailed design that you could get laid out as printed circuit board (so you can catch the obvious defects) is one that got more popular during my career. Some people are much better at it than others, and some managers thought they were saving time by skipping the review meeting (and wasted a lot more time than they could ever have saved).

Electronics can be complicated. The process of going from a simple scheme for putting a signal or two into sensors and deriving a bunch of outputs from the sensor outputs isn't always straightforward either.

The one thing you can rely is that somebody who claims to know a single right way of doing it doesn't know what they are talking about.

The business is much too diverse for that to be possible.

On top (or in lieu) of my companies PnP's, I have my own, documented, Personal Engineering Practices (PEP) that I follow for all my projects.

Actual dialog:

"Where do circuits come from?"

"Books"

"Where do the circuits in books come from?"

"Other books."

But inventing things is fun.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Yes. There are so many ways to make expensive mistakes, it's worthwhile to have a procedure and some checklists. And reviews.

A mistake in software can be fixed in minutes. A mistake in hardware, months.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Mostly true.

I've done it, and I've known people who have done it more often. The actual moment of invention may be fun, but the rest is a chore.

I fixed one in half an hour. It wasn't a complete fix, but it was enough to show that the problem was fixable, and got the machine working.

We should have had a review which should have found the error when we released the design for printed circuit layout, but relevant manager saved a couple of hours by skipping the review. It took about a month to work how the hardware wasn't working right, and once the software guys told me what was going wrong I knew exactly what needed fixing. Irritating.

There's a ton of academic and industry information out there about this topic. The fact of your saying there is none is due to 1) not even looking for it, and/or 2) having difficulty putting higher-level abstraction into practice.

This is an example of one source. The author knows quite a bit about the subject, to look at his resume:

Concurrent Engineering and Design for Manufacture of Electronics Products

And don't forget the requisite certifications making the product legal.

Substitute "app notes" and "data sheets" for "books" and that's the level I'm at.

Sometimes you can pull circuits from the instrument literature, but a lot of that isn't too well thought out. The American Review of Scientific Instruments has published some bad examples. The British journal Measurement Technology (originally the Journal of Scientific Instruments) wasn't as bad. I refereed stuff for them (not all that often) and what they got offered could be pretty horrible.

One way to get past that point is to assume that all the app note circuits are junk, and try to get figure out everything that’s wrong with them.

Some of them work fine, of course, but you can usually find stuff like input protection, supply rejection issues, slew limiting, or poor noise performance.

And then there’s the BOM to consider. Does the circuit use gold plated parts because they’re made by the same company?

Is there a simpler or cheaper way to do the job?

You can learn a lot of design that way.

Cheers

Phil Hobbs

That definitely happens. I wrote a couple of app notes for TI codecs and was required to use certain op-amps because they were being promoted at that time. They were perfectly suitable and not expensive, but I would probably have made different choices if given a free hand. John

We had a friend that had a nice little business making little PC boards that were entirely copies of example circuits in data sheets, or re-dos of eval boards.

Unfortunately, he started day-trading too. Lost his house, lost his wife.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

If a design is public, anybody can do it. So lacking some attractive specs, the business becomes a race to the bottom, minimum price.

I was considering making some motor drivers, but found some nice looking units online for $15.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

If they understand what has been published.

Minimum price involves production in very large volume - the rule of thumb is that producing in ten times the volume lets you halve the price. Germany and China both chose to do exactly that with solar cells, and Germany hogged the market for a couple of years after spending big to manufacture in high volume. It's not clear that the market is big enough to let anybody trump the Chinese.

If it's a routine application and somebody has mass-produced an assembly that you can use, that's what you expect to happen.

What did you learn from your talk?

You are actually Sloman, right?

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

I met a few people who have startup ideas and may want some help with electronics.

One guy is from India and just got his masters degree and is going back to India to start something up. I told him to contact me if he needs help.

And a lady has a startup that invoves a safety device for disabled people. Ditto.

There are so many people with startup ideas. Few will be sucessful. And few know much about electronics.

The maker meetups are fun. Typically good beer and mediocre pizza.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics

Wrong. There are other people out there who are aware that you are prone to sound off about subjects you know little about.

John Larkin thinks that CE stands for "can't enforce".

And, obviously, there are other people who begin technical discussions with lame insults.

That's what my brit friends told me in Milton Keynes. They just slap stickers on things and are not yet in prison.

John Larkin Highland Tech Glen Canyon Design Center Lunatic Fringe Electronics