microelectronics design

Not since Jim Thompson died. He was the only person who posted here who actually designed integrated circuits. I can't say I was all that fond of the parts he had designed.

People like Barry Gilbert and Bob Widlar devised much nicer parts. Some of us do know a bit about what goes on inside integrated circuits, and there are people who lurk here who may know more.

Fire away! It cannot possibly be worse than some in here.

Jeroen

OK, thank you for the answer. I try to post my question anyway in case somebody is deep in the matter. My apologies if the question may seems a bit obscure.

I just wonder if somebody makes use (or plans to do it) of the Murmann's and Jesper's method of MOS sizing (gm/Id method). This is not a new method but it is getting more and more popular among analog designers, mainly now that quadratic law fails with deep sub-micrometer devices.

I am interested in any kind of feedback (both postive, negative and neutrals). Recently I tried to use it in a couple of design but I found this method not much practical. Finally I had to adapt it to my needs and it took some effort from my side. I just am interested to listen to real practical design experience from other colleagues.

Thanks to anybody could advise.

Do you mean design of a semiconductor chip itself? This is mostly a political/virus/feud forum, with a very few people who can design at board level, with available chips.

But ask.

You could try asking in this slack:

It seems like a large fraction of the internet's hobbyist/student chip designers are in there, lured by the promise of a free tape-out. Perhaps you are already reading it.

Rather more than John Larkin thinks. Anybody who doesn't praise his efforts at this kind of work gets written off as inexpert, which does leave out quite a few people who are more competent at it than he is (though less good at making money out of it).

Good advice,

Thank you, I didn't know. I will surely try.

I cannot read (yet) the original reply, so I answer here. My humble question was in the first reply to Bill Sloman: shortly I was looking for feelings from somebody who makes use of the gm/Id method for MOS sizing in analog design. I start to think that it looks like a sort of "scam", kind of "microelectronics design marketing" since it seems to gain a certain popularity. But I would be glad I'm wrong; perhaps it depends only me and I would figure out if and where my opinion fails.

orts at this kind of work gets written off as inexpert, which does leave ou t quite a few people who are more competent at it than he is (though less g ood at making money out of it).

I've passed the question on to the chair of electron devices chapter in the NSW branch of the IEEE. She worked for some years at IMEC in Belgium, and did known what the question was about. She passed it on to the secretary of the chapter, who might know enough about it to give some kind of answer. N o response from there yet.

ctronics design marketing" since it seems to gain a certain popularity.

Things that work can get popular too.

gure out if and where my opinion fails.

Asking questions is usually a good idea. The chances of getting a useful an swer aren't all that high, but sometimes you can get a response that lets y ou ask a more useful question, so it can open up new avenues of enquiry, o r even put you in a place to ask a better question - sometimes one you can answer for yourself.

Science and technology is all about talking about and dissecting problems. Sometimes you can put together an answer, but most of it is getting a clea rer idea of the problem you ought to be tackling. which isn't always quite what you thought it was.

Just to be more precise about this point, it was not my intention to be spiteful about the method: it do works! My concern was about some other aspects such as: a long preliminary job to put it in place and too many degrees of freedom (that could be good sometimes, but it may not drive you to an optimal solution or, in the best case, you must always cope with the unavoidable design compromises, cluelessly, as in the standard design flow). Finally, when I was talking about "design marketing", the fact that sometimes in a few papers it is boldly presented as "the ultimate solution to the analog design issues". Well, to be honest, in the end I find it not so different to the old classical methods which make use of charts, graphs and tables. Only the set of design parameters (or figures of merit) is different. So, I do not understand the aforementioned claim. But, perhaps it's me...

ectronics design marketing" since it seems to gain a certain popularity.

imes in a few papers it is boldly presented as "the ultimate solution to th e analog design issues".

There's always a tendency to talk up a method that works for you. It may w ork better for some kinds of work than others, and if you haven't had to us e it cases where it's bit time-consuming you won't know about that.

ssical methods which make use of charts, graphs and tables.

Anything you can automate can get done a lot more quickly by a machine. Cha rts, graph ans tables are just ways of presenting numerical information so that you can process it fast. If you can work out how you are processing th e data, and get a computer to do it for you, the computer is usually a lot quicker.

o, I do not understand the aforementioned claim. But, perhaps it's me...

It isn't always easy to work exactly what kind of processing you are doing in your head. If you can formalise that, you can usually get a machine to d o it faster.