有点长，google另外一个group analog IC design一个家伙的提问引发的一些讨论

4 views

Skip to first unread message

oliv...@gmail.com

unread,

Nov 30, 2006, 4:30:18 AM11/30/06

to 可以谈工作，也可以八卦

Hello,

I am an electrical engineer (BSEE + some graduate work in EE)exploring
ANALOG IC DESIGN as a career. The image I have from this filed is
very attractive to me and I would like to ask individuals to offer
their perspective on this filed.

My entire background can be summerized as follows:

-DSP software development for communications (FAX, Voice, Telephony)
-Software testing for communications
-QA for modmes including hardware and software qualifications

Through out school and my career I have always been drawn to designs
involving transistors, amplifiers, RF circuits, etc. There just seems
to be something very magical about all these components coming
together and creating electronic magic. I think part of this goes
back to when I was in the 6th grade when I put together my first radio
kit- I did not know how it worked,but I was facinated by it. I
understand the difference between IC and discrete design, I just
wanted to give the reader a picture of why I am attracted to analog
circuit design.

The questions I have are:

1. What is it like to be an anlog IC designer? what is a work
day/week/month like?
2. What part of this job is frustrating? negative ? etc. ?
3. How can someone with my background getinto this filed?

Thanks a lot for guiding me.

Martin

On 20 Jun 2003 20:09:21 -0700, marto...@yahoo.com (Martin) wrote:

>Hello,

>I am an electrical engineer (BSEE + some graduate work in EE)exploring
>ANALOG IC DESIGN as a career. The image I have from this filed is
>very attractive to me and I would like to ask individuals to offer
>their perspective on this filed.

>My entire background can be summerized as follows:

>-DSP software development for communications (FAX, Voice, Telephony)
>-Software testing for communications
>-QA for modmes including hardware and software qualifications

My personal opinion is that the above skills are *software*, not
*analog*.

If all this strikes you as "magical" you are actually off to a good
start. Inquisitiveness is essential to being an analog designer.

(Please note that most of the lurkers here aren't *designers*...
they're *tinkerers*.)

If you want to pursue analog my suggestion is to *build something*!
Maybe you're into audio... build an RIAA preamp... a design you can
find most anywhere on the web.

BUT!! Analyse it, using Algebra, no simulation. If you can
understand *why* it works, then you're on your way. If you can't
understand it, go into sales ;-)

As for your questions:

(1) It's marvelous... I work 60+ hours per week because I'm having
fun!

(2) Customers cause frustration, they never really know what they
want... it's part of the job description ;-)

(3) BUILD SOMETHING and UNDERSTAND HOW/WHY IT WORKS!!

For proper E-mail replies SWAP "-" and "_"

Get Lolita Out of Debt... Add Three Inches to Your Mortgage!

Reply Rate this post: Text for clearing space

- Hide quoted text -
- Show quoted text -
Jim Thompson wrote:

> On 20 Jun 2003 20:09:21 -0700, marto...@yahoo.com (Martin) wrote:

> >Hello,

> >I am an electrical engineer (BSEE + some graduate work in EE)exploring
> >ANALOG IC DESIGN as a career. The image I have from this filed is
> >very attractive to me and I would like to ask individuals to offer
> >their perspective on this filed.

> >My entire background can be summerized as follows:

> >-DSP software development for communications (FAX, Voice, Telephony)
> >-Software testing for communications
> >-QA for modmes including hardware and software qualifications

> My personal opinion is that the above skills are *software*, not
> *analog*.

I'm going to back you up on that. I used to do (20 years ago and more)
quite a bit of analog design. Now I write software which does what used
to be analog functions. My analog experience gives me intuition into
how
to write my software and test it but it is software all the same.

> >The questions I have are:

> >1. What is it like to be an anlog IC designer? what is a work
> >day/week/month like?
> >2. What part of this job is frustrating? negative ? etc. ?
> >3. How can someone with my background getinto this filed?

> If you want to pursue analog my suggestion is to *build something*!
> Maybe you're into audio... build an RIAA preamp... a design you can
> find most anywhere on the web.

> BUT!! Analyse it, using Algebra, no simulation. If you can
> understand *why* it works, then you're on your way. If you can't
> understand it, go into sales ;-)

There is great truth in what you say. There is a lot of leaning on
simulation. Analyzing a design the hard way gives intuition. Intuition
is part of inspiration. The circuit to simulate must come before
simulation. It is getting that initial circuit that stumps some
perfectly intelligent engineers.

Many years ago, a customer engineer from a well known company asked me
how I started a design. I told him I started with a blank sheet of
paper
and then drew a line. Enigmatic as that is, it is precisely what I do.
It is also a fact that, back in those days, I always had a Pink Pearl
Eraser in my pocket.

A good analog designer starts intuitively or inspirationally. The
intuition comes from analyzing and building circuits. The inspiration
is
from within.

Chuck
--
... The times have been,
That, when the brains were out,
the man would die. ... Macbeth
Chuck Simmons chrl...@webaccess.net

Reply Rate this post: Text for clearing space

- Hide quoted text -
- Show quoted text -
Chuck Simmons wrote:
> Jim Thompson wrote:

>> On 20 Jun 2003 20:09:21 -0700, marto...@yahoo.com (Martin) wrote:

>>> Hello,

>>> I am an electrical engineer (BSEE + some graduate work in
>>> EE)exploring ANALOG IC DESIGN as a career. The image I have from
>>> this filed is very attractive to me and I would like to ask
>>> individuals to offer their perspective on this filed.

>>> My entire background can be summerized as follows:

>>> -DSP software development for communications (FAX, Voice, Telephony)
>>> -Software testing for communications
>>> -QA for modmes including hardware and software qualifications

>> My personal opinion is that the above skills are *software*, not
>> *analog*.

> I'm going to back you up on that. I used to do (20 years ago and more)
> quite a bit of analog design. Now I write software which does what
> used to be analog functions. My analog experience gives me intuition
> into how to write my software and test it but it is software all the
> same.

>>> The questions I have are:

>>> 1. What is it like to be an anlog IC designer? what is a work
>>> day/week/month like?
>>> 2. What part of this job is frustrating? negative ? etc. ?
>>> 3. How can someone with my background getinto this filed?

>> If you want to pursue analog my suggestion is to *build something*!
>> Maybe you're into audio... build an RIAA preamp... a design you can
>> find most anywhere on the web.

>> BUT!! Analyse it, using Algebra, no simulation. If you can
>> understand *why* it works, then you're on your way. If you can't
>> understand it, go into sales ;-)

> There is great truth in what you say. There is a lot of leaning on
> simulation. Analyzing a design the hard way gives intuition.

This is true, however,... *some* "intuition" is also only available by
in depth simulation. The world is far to complicated to understand it
by
pencil and paper. The equations are simply far too complicated for any
sort of hand calculations. The classic example, is designing nuclear
bombs or colliding black holes. Sure, an overview might be possible,
but
in *practise* you can only gain true understanding by doing the
details,
extensivly, no matter how much we would like to con ourselves and think
otherwise.

> Intuition
> is part of inspiration. The circuit to simulate must come before
> simulation. It is getting that initial circuit that stumps some
> perfectly intelligent engineers.

Althouh getting the componet values is a little easier now
http://www.anasoft.co.uk/DeviceDesigner.html :-)

> Many years ago, a customer engineer from a well known company asked me
> how I started a design. I told him I started with a blank sheet of
> paper and then drew a line. Enigmatic as that is, it is precisely
> what I do. It is also a fact that, back in those days, I always had a
> Pink Pearl Eraser in my pocket.

> A good analog designer starts intuitively or inspirationally. The
> intuition comes from analyzing and building circuits. The inspiration
> is from within.

Yes and no. Yes, the "intuition" is actually experience and knowledge.
The *true* inspirational bit, when you *really* analyse it in *depth*,
is actually random chance. The starting point for any design is an
evaluation of the existing ones you already know about.

Kevin Aylward
salesEXTR...@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Reply Rate this post: Text for clearing space

> The questions I have are:

> 1. What is it like to be an anlog IC designer? what is a work
> day/week/month like?

This will depend on where you are working. Some companies treat IC
designers
as a commodity, others (the successful ones) treat them as a resource
for
the company's growth. These companies generally have lots of product
ideas,
but not enough designers to implement them. Analog IC design is the
gating
resource. They know this, so productive designers are treated well.

IC design work that I am familiar with (I am an apps engineer that
works
closely with analog IC designers) involves a lot of simulation, some
lab
work, and a lot of rigor. In the old days. it was OK if a chip didn't
work
the first time, now it is expected that no "preventable" mistakes will
be
made. Many chips still need a 2nd pass, but the reason should be good.
Having said that, it is still a very rewarding creative career.

> 2. What part of this job is frustrating? negative ? etc. ?

Schedules. They are much more critical now than in the past. You are
asked
how much time something will take, and many cogs start turning based on
your
answer. You might not like getting leaned on by your boss to make a
deadline.

> 3. How can someone with my background getinto this filed?

Most analog IC designers that are hired these days have had some grad
school
IC classes. If you don't have that, it will be harder but not
impossible.
You might consider starting in applications which involves circuit (not
IC)
design and customer support. This gets you in the door. From there you
can
talk to IC designers and managers and sell yourself as a design
candidate,
but be prepared for at least a year of apps. You'll have to prove
yourself
as a "live one" before you will have credibility. Taking IC design
classes
at night of through extension programs will help a lot too.

I final comment is that you will ALWAYS be employable if you choose
this
career. There are a lot of people out of work now, but none of them are
analog IC designers.

- Hide quoted text -
- Show quoted text -

> Thanks a lot for guiding me.

> Martin

Reply Rate this post: Text for clearing space

A good book to read is "The Art and Science of Analog Circuit Design"
(edited by Jim Williams). Highly recommended!

Reply Rate this post: Text for clearing space

Jim Thompson wrote:

> BUT!! Analyse it, using Algebra, no simulation. If you can
> understand *why* it works, then you're on your way. If you can't
> understand it, go into sales ;-)

Or get a copy of AoE....

Reply Rate this post: Text for clearing space

Jim Thompson wrote:

> (Please note that most of the lurkers here aren't *designers*...
> they're *tinkerers*.)

You mean "tinker-bells"...

Reply Rate this post: Text for clearing space

On Sat, 21 Jun 2003 12:03:13 GMT, Fred Bloggs <nos...@nospam.com>
wrote:

>Jim Thompson wrote:

>> BUT!! Analyse it, using Algebra, no simulation. If you can
>> understand *why* it works, then you're on your way. If you can't
>> understand it, go into sales ;-)

>Or get a copy of AoE....

AND hand analyse as many of those circuits as possible.

For proper E-mail replies SWAP "-" and "_"

Get Lolita Out of Debt... Add Three Inches to Your Mortgage!

Reply Rate this post: Text for clearing space

On Sat, 21 Jun 2003 13:23:09 GMT, Fred Bloggs <nos...@nospam.com>
wrote:

>Jim Thompson wrote:

>> (Please note that most of the lurkers here aren't *designers*...
>> they're *tinkerers*.)

>You mean "tinker-bells"...

Now Fred I wasn't going to say anything personal about you ;-)

For proper E-mail replies SWAP "-" and "_"

Get Lolita Out of Debt... Add Three Inches to Your Mortgage!

Reply Rate this post: Text for clearing space

On Fri, 20 Jun 2003 20:47:32 -0700, Jim Thompson wrote:
> If you want to pursue analog my suggestion is to *build something*!
> Maybe you're into audio... build an RIAA preamp... a design you can
> find most anywhere on the web.

> BUT!! Analyse it, using Algebra, no simulation. If you can
> understand *why* it works, then you're on your way. If you can't
> understand it, go into sales ;-)

Understanding is the key. Whenever I hear someone talk about analog
design
as "black magic," I know that they aren't an analog designer. There's
no
magic to it - it might look like magic to someone who doesn't
understand
it, but when an analog circuit is well designed and works well, it's
not
because of magic.

> (2) Customers cause frustration, they never really know what they
> want... it's part of the job description ;-)

If only management knew that the success of their engineering team is
largely due to ignoring the oft-cited "The Customer is Always Right"
slogan. Customers have lots of ideas, from the architecture to the
design
of individual circuits. Some of them may be fine, and others may not
be. If
you can't go to your customer and say, "I don't think this is going to
work," then you're going to build lots of bricks.

-- Mike --

Reply Rate this post: Text for clearing space

Mike <m...@nospam.com> wrote in message

news:m0v9xnolc7e0$.1gh7525cikhoh$.dlg@40tude.net...

- Hide quoted text -
- Show quoted text -
> On Fri, 20 Jun 2003 20:47:32 -0700, Jim Thompson wrote:

> > If you want to pursue analog my suggestion is to *build something*!
> > Maybe you're into audio... build an RIAA preamp... a design you can
> > find most anywhere on the web.

> > BUT!! Analyse it, using Algebra, no simulation. If you can
> > understand *why* it works, then you're on your way. If you can't
> > understand it, go into sales ;-)

> Understanding is the key. Whenever I hear someone talk about analog design
> as "black magic," I know that they aren't an analog designer. There's no
> magic to it - it might look like magic to someone who doesn't understand
> it, but when an analog circuit is well designed and works well, it's not
> because of magic.

> > (2) Customers cause frustration, they never really know what they
> > want... it's part of the job description ;-)

> If only management knew that the success of their engineering team is
> largely due to ignoring the oft-cited "The Customer is Always Right"
> slogan. Customers have lots of ideas, from the architecture to the design
> of individual circuits. Some of them may be fine, and others may not be.
If
> you can't go to your customer and say, "I don't think this is going to
> work," then you're going to build lots of bricks.

> -- Mike --

Digital design maybe. A computer can do it. Analogue design NO.
I'd defy anyone to sit down and *design* a SMPU (for example) that by
dint
of pure cold design knowledge, circuit analysis and long experienced
ability
will be assured of meeting particular pre-assigned limits for conducted
and
radiated emissions. (Brute force, solid Silver casings and enormous
filters
not allowed).
The physics always apply but until we have fast enough personal
computers
that can keep track of the whole 3D product matrix of individual
electrons
at the quantum level then analogue design can only ever be 'Black
magic'. It
not blank math(s?),it's not manipulation of 1's and 0's, it's an
artform. We
can only hope to aspire to a solution.
The skilled workers are more liable to produce a better result than
those
less skilled. They are simply the better artists or better able at
keeping
the evil spirits under control.
I seem to remember that one contributor has his book aptly titled "The
Art
of ..."
regards
john

Reply Rate this post: Text for clearing space

On Sat, 21 Jun 2003 19:37:28 +0100, "John Jardine"

- Hide quoted text -
- Show quoted text -
<j...@jjdesigns.fsnet.co.uk> wrote:

>Mike <m...@nospam.com> wrote in message
>news:m0v9xnolc7e0$.1gh7525cikhoh$.dlg@40tude.net...
[snip]

>> If only management knew that the success of their engineering team is
>> largely due to ignoring the oft-cited "The Customer is Always Right"
>> slogan. Customers have lots of ideas, from the architecture to the design
>> of individual circuits. Some of them may be fine, and others may not be.
>If
>> you can't go to your customer and say, "I don't think this is going to
>> work," then you're going to build lots of bricks.

>> -- Mike --

>Digital design maybe. A computer can do it. Analogue design NO.
>I'd defy anyone to sit down and *design* a SMPU (for example) that by dint
>of pure cold design knowledge, circuit analysis and long experienced ability
>will be assured of meeting particular pre-assigned limits for conducted and
>radiated emissions. (Brute force, solid Silver casings and enormous filters
>not allowed).
>The physics always apply but until we have fast enough personal computers
>that can keep track of the whole 3D product matrix of individual electrons
>at the quantum level then analogue design can only ever be 'Black magic'. It
>not blank math(s?),it's not manipulation of 1's and 0's, it's an artform. We
>can only hope to aspire to a solution.
>The skilled workers are more liable to produce a better result than those
>less skilled. They are simply the better artists or better able at keeping
>the evil spirits under control.
>I seem to remember that one contributor has his book aptly titled "The Art
>of ..."
>regards
>john

I guess I should have added that a requisite skill is the ability to
throw away designs that have grown due to too many patches, and start
completely afresh.

For proper E-mail replies SWAP "-" and "_"

Get Lolita Out of Debt... Add Three Inches to Your Mortgage!

Reply Rate this post: Text for clearing space

- Hide quoted text -
- Show quoted text -
Jim Thompson wrote:
> On Sat, 21 Jun 2003 12:03:13 GMT, Fred Bloggs <nos...@nospam.com>
> wrote:

>> Jim Thompson wrote:

>>> BUT!! Analyse it, using Algebra, no simulation. If you can
>>> understand *why* it works, then you're on your way. If you can't
>>> understand it, go into sales ;-)

>> Or get a copy of AoE....

> AND hand analyse as many of those circuits as possible.

I would agree with this. It gets you a real feel for what's going on.
The detailed hand calculations gets you to being able to look at the
1st
and some 2nd order approximations from inspection.

Kevin Aylward
salesEXTR...@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Reply Rate this post: Text for clearing space

- Hide quoted text -
- Show quoted text -
John Jardine wrote:
> Mike <m...@nospam.com> wrote in message
> news:m0v9xnolc7e0$.1gh7525cikhoh$.dlg@40tude.net...
>> On Fri, 20 Jun 2003 20:47:32 -0700, Jim Thompson wrote:

>>> If you want to pursue analog my suggestion is to *build something*!
>>> Maybe you're into audio... build an RIAA preamp... a design you can
>>> find most anywhere on the web.

>>> BUT!! Analyse it, using Algebra, no simulation. If you can
>>> understand *why* it works, then you're on your way. If you can't
>>> understand it, go into sales ;-)

>> Understanding is the key. Whenever I hear someone talk about analog
>> design as "black magic," I know that they aren't an analog designer.
>> There's no magic to it - it might look like magic to someone who
>> doesn't understand it, but when an analog circuit is well designed
>> and works well, it's not because of magic.

>>> (2) Customers cause frustration, they never really know what they
>>> want... it's part of the job description ;-)

>> If only management knew that the success of their engineering team is
>> largely due to ignoring the oft-cited "The Customer is Always Right"
>> slogan. Customers have lots of ideas, from the architecture to the
>> design of individual circuits. Some of them may be fine, and others
>> may not be. If you can't go to your customer and say, "I don't think
>> this is going to work," then you're going to build lots of bricks.

>> -- Mike --

> Digital design maybe. A computer can do it. Analogue design NO.
> I'd defy anyone to sit down and *design* a SMPU (for example) that by
> dint of pure cold design knowledge, circuit analysis and long
> experienced ability will be assured of meeting particular
> pre-assigned limits for conducted and radiated emissions.

Many can do this, and this is because they have done many before.
Suppose, the last 5 smps you had designed were 10A jobs, and you new
one
was a 100ma, you can be pretty sure you know how to get it within emc
specs.

(Brute

> force, solid Silver casings and enormous filters not allowed).
> The physics always apply but until we have fast enough personal
> computers that can keep track of the whole 3D product matrix of
> individual electrons at the quantum level then analogue design can
> only ever be 'Black magic'.

Not at all. I think your confusing the fact that not everything is
known, does not mean that nothing is known. For the most part, the bulk
of any analogue design is prior knowledge and experience. Many, people
have had shippable, first time passes of quite complicated i.c.s This
would not be possible if it was "Black Magic". In addition, you don't
need to know everything in order to ship product. Its certainly true,
that in general, there are always mistakes, and unknowns cropping up,
but that does not imply that you need magic to account from them.

Kevin Aylward
salesEXTR...@anasoft.co.uk
http://www.anasoft.co.uk
SuperSpice, a very affordable Mixed-Mode
Windows Simulator with Schematic Capture,
Waveform Display, FFT's and Filter Design.

Reply Rate this post: Text for clearing space

On Sat, 21 Jun 2003 19:37:28 +0100, John Jardine wrote:
> The skilled workers are more liable to produce a better result than those
> less skilled. They are simply the better artists or better able at keeping
> the evil spirits under control.
> I seem to remember that one contributor has his book aptly titled "The Art
> of ..."

Indeed. Not "The Black Magic of..."

-- Mike --

Reply Rate this post: Text for clearing space

I am very grateful for all input and would appreciate more from input.
I have received some indication that work opportunities in this filed
are scarce. What are the major metroplotian areas that have a lot of
Analog IC Design firms?

Thanks,

Martin

- Hide quoted text -
- Show quoted text -
Mike <m...@nospam.com> wrote in message
<news:m0v9xnolc7e0$.1gh7525cikhoh$.d...@40tude.net>...
> On Fri, 20 Jun 2003 20:47:32 -0700, Jim Thompson wrote:

> > If you want to pursue analog my suggestion is to *build something*!
> > Maybe you're into audio... build an RIAA preamp... a design you can
> > find most anywhere on the web.

> > BUT!! Analyse it, using Algebra, no simulation. If you can
> > understand *why* it works, then you're on your way. If you can't
> > understand it, go into sales ;-)

> > (2) Customers cause frustration, they never really know what they
> > want... it's part of the job description ;-)

> If only management knew that the success of their engineering team is
> largely due to ignoring the oft-cited "The Customer is Always Right"
> slogan. Customers have lots of ideas, from the architecture to the design
> of individual circuits. Some of them may be fine, and others may not be. If
> you can't go to your customer and say, "I don't think this is going to
> work," then you're going to build lots of bricks.

> -- Mike --

Reply Rate this post: Text for clearing space

On 12 Jul 2003 16:24:32 -0700, marto...@yahoo.com (Martin) wrote:

>I am very grateful for all input and would appreciate more from input.
> I have received some indication that work opportunities in this filed
>are scarce. What are the major metroplotian areas that have a lot of
>Analog IC Design firms?

>Thanks,

>Martin

[snip]

The Bay Area probably has the most. But custom design is spread all
over the US.