what are the methods of analog circuit design?

[Mehdi12]

Hi all
I learned digital circuit design. First I should get truth table and then get the equations of the circuit,
but I didn't find the methods of analog circuit design like I did in digital circuit.

 

[jiripolivka]

Hi all
I learned digital circuit design. First I should get truth table and then get the equations of the circuit,
but I didn't find the methods of analog circuit design like I did in digital circuit.

All my life I was an analog system designer, quite successful. I do not remember I needed more than Ohm¨s Law for my designs. What analog circuits do you plan to design?

 

[andre_luis]

For a static analysis of analog signals of a more complex circuit having linear components, you could use for example Kirchoff law (KCL), Nodal Analysis (MNA), etc...

 

[schmitt trigger]

Also, you must know the frequency dependency of reactive components: Xl = 2*pi*f*L Xc = 1/2*pi*f*C
And that those reactances, the inductor has an angle of +90 degrees (or j) and the capacitor an angle of -90 degrees (or -j).

Also calculate time constants: tau = L/R, tau = C*R

 

[dick_freebird]

There are methods within methods and what you do
depends on what specialty you're working in and the
product.

Learn by doing is a way to go. Worked for me. I went
to analog by way of interface products that gave me
a series of bite-sized challenges / experiences. Still
go at it by gut more than head. There's room in this
world for equation-mongers and shade tree mechanics
both.

 

[CataM]

Get experience by practicing is the only way. If you want to be good at something, then, just practice that something.

I have a question related with this for you guys...

If you go through a lot of circuits then you get used to the way circuits are made and their topology (the way elements are placed)... E.g. a small signal amplifier has the very well known structure that is first taught when starts studying electronics...

My question for an analog designer is: how much should one build circuits by copying already made ones ? I mean, for an experienced designer who has seen a lot of circuits, when he wants to build up something he just adapts an already existing one ? or he come up with a very new one ?

 

[jiripolivka]

Get experience by practicing is the only way. If you want to be good at something, then, just practice that something.

I have a question related with this for you guys...

If you go through a lot of circuits then you get used to the way circuits are made and their topology (the way elements are placed)... E.g. a small signal amplifier has the very well known structure that is first taught when starts studying electronics...

My question for an analog designer is: how much should one build circuits by copying already made ones ? I mean, for an experienced designer who has seen a lot of circuits, when he wants to build up something he just adapts an already existing one ? or he come up with a very new one ?

I did not see what circuits you need to analyze or design.
In the analog world, most reference designs are offered by manufacturers, like opamps and other devices. I used available data to design needed systems, often better than specified parameters.
Again- what exactly do you need?

 

[Romansh]

For me the best way to learn was to simulate basic circuits from the books and than to try changing some parameters and to observe what is going on on different nodes.

 

[andre_luis]

when he wants to build up something he just adapts an already existing one ? or he come up with a very new one ?

The more one knows about something, more rigorous becomes that person in regard to the performance of his design in all the aspects that he's able to assess. For many cases without much constraints, someone could even copy/paste a working part of another design, but just an experienced (skilled) engineer is able to decide if have to make changes or not, aswell to tune the design to the limits of what the current technology/topology is able to achieve.

 

[dick_freebird]

In most cases, especially anything complex, you will
borrow what you can to leave yourself time to design
what you must.

 

[crutschow]

All my life I was an analog system designer, quite successful. I do not remember I needed more than Ohm¨s Law for my designs. ..............

So all your analog designs were only DC, no AC?

 

[chuckey]

I think most analogue circuits have already been invented. if you are after any circuit just Google it. The real problem are your constraints, is the IC too expensive?, does the amplifier run off odd supply lines? Can we make it from two $.2 FETs rather then a $10 IC?, have we the SMT technology?
In the end you will have to build a prototype and this is when you really start to learn. Fiddling around on simulators should give you a theoretical start, but in the end you have to build it.
Frank

 

[Dan Mills]

Almost all (industrial) design can be summed up as copy what you can, design what you must.

Now the place that tends to separate the men from the boys is when you are doing low noise/high Z/high frequency or non linear analogue (A 'Scope vertical input amplifier makes a fun design project).

Analogue devices sell a wonderful log amp, but can you get away with using a two transistor array and a spare opamp (And maybe saving $5 off the BOM)?
The answer is maybe, use the non linear Vbe/Ib relation from the Ebbers Moll eqn, in a feedback loop, with the second transistor compensating for temperature....

Take a look at some of the old articles by Bob Pease (And Jim Williams) to see the art in all its glory.

I would warn about simulation, it is useful for sanity checking, and usually if the sim says no, then the thing will not work, but the reverse is NOT true, a simulation showing something working is NOT in general good enough to prove an idea.

Regards, Dan.

 

[jiripolivka]

So all your analog designs were only DC, no AC?

Hello Crutschow: I have worked from DC to over 100 GHz, and Ohm¨s Law was quite good in all my designs. Do you think it does not work for AC?
I even designed some simple digital control circuits I needed, but not micro controllers.
Most analog circuits were possibly already invented. I used test circuits to verify my designs, and often my designs exceeded the specifications of manufacturers¨components, mainly MMICs and opamps. I can recommend such procedure with prototyping to all designers.

 

[andre_luis]

I guess he wished refer to the transient analysis rather than a steady state AC analysis. In fact, for all other situations, we can compute the impedances of each element and apply Ohm's Law.

In shorts, looks like both you are correct on each viewpoint.

 

[schmitt trigger]

The way I learned a lot of analog tricks, was by reading the "Linear Applications Databook" by National Semiconductor. Literally a goldmine of circuits and tricks, and written in a way that you doesn't require a PhD to understand them.

Another favorites of mine was Thomas Fredericksen's "Intuitive IC Opamps" and all of Walt Jung's cookbooks