RIP Bob Pease

[RedHouse]

Another great mind has passed.

http://www.elektor.com/news/in-memoriam-bob-pease.1869803.lynkx?

[amptramp]

I would like to see National Semiconductor put some of his original application notes back up on their site.  He was the applications guru for an entire industry.  He definitely had heretical views such as insisting that a designer should understand how something works rather than just running simulations, to the point where maybe simulations would not be needed.  His attitude to the design process was as important as his expertise.  He will be missed.

[The Tone God]

That is a shame. I enjoyed reading his material. His book "Troubleshooting Analog Circuits" should be required reading IMHO.

R.I.P. Bob

Andrew

[Morocotopo]

It´s sad to read about the passing of such knowledgeable people.
Andrew, that sounds like an interesting book to read. Is it a commercially edited book or an application paper from National?

[The Tone God]

It´s sad to read about the passing of such knowledgeable people.
Andrew, that sounds like an interesting book to read. Is it a commercially edited book or an application paper from National?

Its a commercial book. ISBN 0-7506-9499-8

Andrew

[electrosonic]

Reading the obiturary, he died leaving a service for Jim Williams (http://en.wikipedia.org/wiki/Jim_Williams_(analog_designer) ). Kind of a shock to have two analog electronics pioneers die with in days of each other.

Andrew.

[Jdansti]

It's been a little over a year since Bob Pease passed away. Here's a neat tribute to him:

[Cliff Schecht]

While I dig the old school mentality and love to sit down and build a circuit, I've always disagreed with the old school mentality about simulations. It seems like a lot of the distrust for simulators comes out of not understanding (or not wanting to understand) how they work. People don't understand that they are working with models of physical devices which are crude in some aspects (and perfectly fine in others). Your simulations are as good or bad as your models and knowing how these models and your simulator works is the key to getting accurate sims.

Lately I've been using NI's Multisim for my simulations. They've got more included models than any other software I've used and they have a unique feature (AFAIK) that lets you add virtual instruments (oscopes, function generators, bode plotter, spec an, logic analyzer, etc..) and tweak them on the fly. Just the other day I built the MS-20 filter in simulation and could tweak both the cutoff input (setting a voltage with a function gen) and the resonance pot and watch the waveforms change on the virtual oscope. There's even a speaker that lets you listen to samples of whatever you put into it. This is about as close as one can get to breadboarding virtually and it lets me test circuit ideas wherever I have my laptop. Pretty dang cool!

[Jdansti]

While I dig the old school mentality and love to sit down and build a circuit, I've always disagreed with the old school mentality about simulations. It seems like a lot of the distrust for simulators comes out of not understanding (or not wanting to understand) how they work. People don't understand that they are working with models of physical devices which are crude in some aspects (and perfectly fine in others). Your simulations are as good or bad as your models and knowing how these models and your simulator works is the key to getting accurate sims.

Lately I've been using NI's Multisim for my simulations. They've got more included models than any other software I've used and they have a unique feature (AFAIK) that lets you add virtual instruments (oscopes, function generators, bode plotter, spec an, logic analyzer, etc..) and tweak them on the fly. Just the other day I built the MS-20 filter in simulation and could tweak both the cutoff input (setting a voltage with a function gen) and the resonance pot and watch the waveforms change on the virtual oscope. There's even a speaker that lets you listen to samples of whatever you put into it. This is about as close as one can get to breadboarding virtually and it lets me test circuit ideas wherever I have my laptop. Pretty dang cool!

As a newish hobbiest, I've never used a simulator, but I'm sure they have their place and are quite good at doing certain types of modeling.  Otherwise, the companies that create and sell them wouldn't stay in business very long. My guess is that like any tool, they meant to be used for certain applications and are not perfect for all applications. I've seen people drive a nail with a wrench, but a hammer would have been a better tool. 

[Paul Marossy]

I see these simulators as being approximations of how it works in the real world.

[Jazznoise]

The limitations can be problematic, however, and certain abuse of feedback/differential amps can have pretty inconsistent behaviour. In my quests to build a Ring Modulator I used a FET in the feedback loop of a TL071 which simulated perfectly.

It didn't work so good.

The problem with Multisim is it guesses what you want to do, or at least it seems to. Once it see "Ah, you're building a filter!" it'll do all your filter calcs perfectly. Likewise when it goes "Ah, that's a push/pull pair" or "That's a 4 bit counter". I've crashed it by building a precision rectifier, boosting the rectified signal and feeding it into one side of a diffierential amp with the original signal (cheeky attempt at building an octavia type unit sans XFrmr).