The stupidity of slimulations

What I usually say is "a simulator is a useful thing but it won't make you a design engineer, just like a pocket calculator won't make you a mathematician".

Overall I share your attitude to things going way too virtual. As a guy said in a (non-English) forum the other day, "you don't have to make a product nowadays, you just need to draw a 3D, worst case to simulate it". It appears to pay off, looking at the vast numbers of useless clickers... :D .

Dimiter

------------------------------------------------------ Dimiter Popoff, TGI

------------------------------------------------------

I doubt very much if anybody would see a simulation as a product.

Simulations are useful - up to a point - but any mathematical model is a si mplification of reality, and until you've built something that works, you c an't tell if it is an over-simplification.

Which is a pity, because simulations are a lot quicker and cheaper to build and test than the real thing.

Used right, they can let you get away with fewer prototypes, and can let yo u learn more from the prototypes you do build, but you've always got to bui ld something.

--
Bill Sloman, Sydney

Hey Jan, surely you could have tossed AGW in there to get the pot boiling!

I wonder how you get to fit so many projects into your waking hours. It's often an interesting read anyway.

Sure, a simulator won't make you a design engineer, but for complex analog asic design, without a simulator, you are certainly not a design engineer. Large transistor count analog ic design is simply impossible to do without extensive simulation. The models are usually extremely good, and providing that the correct simulations are actually performed, physical silicon typically comes out nuts on, no ifs or buts. There is often an incentive to do this, like $50k bonuses for 1st pass silicon!

Pissing about on a bread board only checks out a one off. The only realistic way to design for all component variances is by doing simulations. The full equations of even a one transistor circuit are simply intractable.

The fundamental issue s that you need to know how to drive the simulator correctly. If you don't add in board parasitics, sure you wont get the right answers, but why would a competent designer not add them in from day one? If inductances and capacitances vary all over place, put those 1000% variations in the simulations. Dah...

Kevin Aylward

s,

ou

g

r.

Complex analog asic design is a rather well-specified field.

t g

to

Which has meant that the industry has done a great deal work characterising precisely what it makes.

tic

ll

Which didn't stop people from designing useful circuits.

ght

The usual answer is that they don't know them all that precisely. If you ar e designing with microstrip transmission lines over a ground plane or a pow er plane, you can come pretty close, but lower frequency components aren't engineered to have predictable parasitics, in part because they don't matte r as much at low frequencies, until they rear up and bite you.

Easier said than done. It's usually a better idea to minimise the variation by being picky about the way the circuit is going to get put together. Pro duction always want something that they can slap together, but they also wa nt something that almost always works.

Not a problem with complex analog asic design, but the rest of the world ha s to fit those asics onto printed circuit boards, and plug the boards into racks or housings.

--
Bill Sloman, Sydney

Simulated woman, sex?

Have you read his book? You should.

Right. Too many people use simulation because they don't understand electronics.

We've been interviewing lately. It's shocking to interview someone in college, who just passed a course on transistors, who has no idea how a transistor works. Did that yesterday.

+10V | | | 10K | | c gnd-----10K--------b e | | | gnd

What's the collector voltage? Applicant began mumbling about electron-hole pairs and drawing charactures of vaguely remembered (saturation?) curves. She could probably answer the question if I let her Spice it.

I'd really like to interview her professor, see if he can tell me the collector voltage.

Face it, the majority of EEs aren't very good. Simulation is the only way they might possibly get something to sort of work.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation

Let me guess, applicant was from Berkeley, or was it Tulane ?:-)

Actually, I'm troubled by all universities and their curriculum. All they're teaching now is CAD... no clue as to understanding circuits, fundamentals, or creativity. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142   Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

Wrong and wrong. You are usually wrong lately, and I don't think you're going to get any better.

--

John Larkin                  Highland Technology Inc 
www.highlandtechnology.com   jlarkin at highlandtechnology dot com    

Precision electronic instrumentation

Hear, hear.

Even back when...the areas least taught are the areas I MOST use now! Statistics, Numerical Analyses, Noise, Fourier, and the like, Transforms [luckily was in Ron Bracewell's class]

And the one that bugs me the most, Magnetics all those 'basics' are just misleading. yes, all those 'tripe formulas' work on macroscopic circuits, but cause people to erroneously predict expected performance when extrapolating. It's like dragging baggage.

My wife took one semester of EE years ago in a fit of boredom. Her circuits prof taught circuit analysis as if you were going to program a computer to solve the circuits. On the final exam, there was a problem with a voltage source with two resistors in parallel to it. Using inspection and Ohm's law, she solved it in about 3 lines. She was worried that the prof would have dropped such an easy question on the final and took a peek over to the girl sitting next to her, who was starting her second page of matrix algebra. Interestingly enough, one of that prof's co-op students was doing a work term under me, so I had a chance to talk to the fellow. It was difficult to keep my mind on the co-op student's performance and not call him out for his useless teaching style.

Great discussion (so far...). Thanks Jan for stirring the pot.

The group I'm hand-holding right now is of the cut-and-paste crowd. Every circuit needs a switched-cap OTA :-(

No clue at the device levels... all school-book, CMOS Level=1 equations, and when I chide, they get hostile.

But a fun project starts on Wednesday... and I call all the shots >:-} ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142   Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

Yep. Sad :-(

I lost a project last year... during the initial show-and-tell I told a PhD he was wrong... NOT the thing to do... my wife chides me weekly about that one ;-) ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142   Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.

On a sunny day (Sat, 22 Mar 2014 08:07:43 -0700) it happened John Larkin wrote in :

No I have not, except for some pages he put online IIRC,. Nothing wrong with that.

I did some job interviews, and one question I usually asked was 'draw a relay driver'. Just to see if they forgot the diode....

Well, teachers, sometimes they come from industry with a lot of practical experience, sometimes they are old and have been out of the real world for a long time. I remember one stdent asking in electrionics class back then 'Sir what exactly what is a complementary pair?" he got almost kicked out because the teacher though it was a sex joke. It took us (the whole class) a lot of convincing him that that was a legitimate question. OK in those days transistors were kind of new, and that was a tube guy, but really, these days it is likely the same..

I look for dedication, I do not mind if somebody makes a mistake, I do that myself all the time too, but persistence to solve a problem, really wanting to solve it, makes the right person. Persistence.

.6 volt or thereabouts sorry, no education

Applicant began mumbling about

Excuse me, image pan off screen didn't see ground.

10 volts

On a sunny day (Sat, 22 Mar 2014 16:39:16 GMT) it happened Wayne Chirnside wrote in :

Vce sat is lower than Vbe, this is important to know. It seems a bit un-logical, but Vce sat is usually a few hundred mV for a Si transistor (say 200mV or so).

As I've opined before, a simulator is just a checking device to cope with huge designs.

However, circuit design by use of brain-power must come first.

I will point out (aka brag :-) that I designed chips for ~17 years without aid of any simulator... ~50 total chip designs... and they all worked.

And I didn't own my own tools until 1987 (25 years into the chip business), when I spent $6500 for a 386/387 plus $8000 for a copy of PSpice... nasty amounts of money back then, but certainly a great return on investment over the years. ...Jim Thompson

--
| James E.Thompson                                 |    mens     | 
| Analog Innovations                               |     et      | 
| Analog/Mixed-Signal ASIC's and Discrete Systems  |    manus    | 
| San Tan Valley, AZ 85142   Skype: Contacts Only  |             | 
| Voice:(480)460-2350  Fax: Available upon request |  Brass Rat  | 
| E-mail Icon at http://www.analog-innovations.com |    1962     | 
              
I love to cook with wine.     Sometimes I even put it in the food.