The stupidity of slimulations

On a sunny day (Sat, 22 Mar 2014 16:53:22 GMT) it happened Jan Panteltje wrote in :

Actually no, then there is a 10k / 10k divider bringing the collector to about 5.7V ? Oh what fun, will the base-emitter junction zener? This is getting complicated fast :-) +10 | 10k | a k GND---- 10k -- b PNP k a | ///

LOL :-)

And instinct. People need to feel the current flowing, get an idea of what a circuit is doing by inspection. There is probably a single-semister hands-on course that could be taught, preferably when a kid is 10-12 years old, that creates basic electrical instincts. That is the proper baseline for the mathematical theory that could be learned later, if the kid is so inclined.

Kids would enjoy it: batteries, resistors, LEDs, audio effects, stuff like that. Let them see and hear what electronics does. Enliven it with occasional smoking parts, loud bangs from capacitor discharges (see how the BANG noise goes as voltage squared?) stuff like that.

Every middle-school kid should get some electrical training, and some machining, soldering, welding, and woodworking experience. Few people will ever use algebra in their adult life, much less calculus.

It's impossible in US public schools, of course. There's a private high school nearby ($40K per year) that does this sort of thing; welding is a mandatory course.

Universities don't do this any more because labs are expensive. Let the kids run free simulators on their laptops.

On a sunny day (Sat, 22 Mar 2014 10:01:01 -0700) it happened John Larkin wrote in :

Yep, I had welding, had to design steam engines and steel bridges, what not. No that I remember too much about steam engines...

Vc = +10 for an NPN, which was the intent of my question.

The PNP situation is more complex, because it depends on inverse beta. +2 or 3 volts, possibly. If an applicant mentioned that, I'd hire them instantly.

The applicant did ask if it was NPN, and I told her that it was. But that's as much progress as she made. I think she was better at remembering than understanding. She was friendly and intelligent, but whatever courses she was taking allowed her to pass without understanding much.

I forgot that entropy/enthalpy/steam table stuff as soon as I could. Some of the thermo is worth remembering. Classical physics is definitely useful in my business.

On a sunny day (Sat, 22 Mar 2014 10:15:12 -0700) it happened John Larkin wrote in :

You are right, here my 'simulation':

Its a BC557B

Never too old to learn :-)

I didn't know you were a Stanford alum. Great class.

Don't know that I follow. Any examples come to mind?

Cheers

Phil Hobbs

Your problem is that you're too abrasive. When I tell people they're wrong, they like it. ;)

Cheers

Phil Hobbs

Full marks. Of course, I think John meant it to be an NPN and not an inverted PNP, but that's his lookout. ;)

Cheers

Phil Hobbs

It was a whiteboard question. She asked if it was an NPN, and I said yes. It was downhill from there.

A PNP with zero inverse beta would give about 5.3ish at the collector. A real PNP might be 2 or 3 volts, very roughly.

This is a great whiteboard quiz question:

What is the base voltage? Emitter voltage? Collector voltage? Base current? Anything else to say?

You'd be astonished at the answers that I get from self-claimed circuit designers with EE degrees.

Can I take lessons ?>:-}

All I said was, "No, that's not correct, it would look like this..."

And I could instantly see in his eyes that all was lost :-(

Their chip will probably fail, so I'll get the last laugh. ...Jim Thompson

It's called "The Internet".

Same goes for digital design of any complexity. There are things that can't reasonably be tested in "reality", as well. Modern microprocessors simply wouldn't exist without logic simulators.

When a set of masks, alone, cost $2M, or more, it makes sense to throw a bone to the developers. Never saw it happen, though. ;-)

Sure. Not to mention checking error conditions.

The problem is that most device models for board-level simulations are crap. Chip designers don't have that problem.

Yes, simulations can only be as good as the models.

NPN or PNP? ;-)

I dunno -- about 3.3V.

This is a PNP you're talking about, right?

It couldn't be an NPN -- that would be too trivial.

I found the transformer equations

V1 = L1.dI1/dt + M. dI2/dt

V2 = M.dI1/dt + L2.dI2/dt

where L1, L2 and M are the inductance of the first and second coils and the ir mutual inductance respectively

in a Siemens (now EPCOS)application note on using their ferrite cores. I ne ver found it in any academic text-book. From that time on, I didn't have an y trouble understanding transformers (though I still had all the usual prob lems of getting them to do what I needed them to do).

E.C.Snelling's application notes for Mullard (which became Philips then Fer roxcube) were much more voluminous and comprehensive, but much less informa tive, mainly because he built his models around "perfect transformers" wit h tacked-on imperfections.

in

not.

of the thermo is worth remembering. Classical physics is definitely useful in my business.

Quantum physics too. If you'd bothered to pay more attention in those class es, you might be less sceptical about anthropogenic global warming, which i s all about infra-red absorbtion and re-radiation, which does depend on fai rly low level quantum physics, as do lasers.