Application note component values

[Tomarino]

When reading application notes from many different manufacturers component values are left off of the schematics provided. How do I derive these values?
Thank you for your time and patience.

 

[mister_rf]

I think they starts with some general transfer function of the components, and later for a specific design you can calculate according to application details.

 

[LvW]

When reading application notes from many different manufacturers component values are left off of the schematics provided. How do I derive these values?
Thank you for your time and patience.

You should consider that there are many, many (in theory: infinite) different circuit alternatives, design strategies and dimensioning ranges to meet a specific design goal.
With other words: Circuit synthesis has many different solutions - opposite to circuit analysis with only one single solution.
And it is a typical task of a good engineer to find the best trade-off for a specific application.
That means: It does not make much sense to present readily designed circuits.

 

[Tomarino]

While I certainly understand that there are infinite solutions this does not really help me.:-D Should I set up the circuit in a cad/cam program without
any values attached to the individual components and go back and plug values in according to my needed output and hope for the best? Would the cad program warn me about incorrect or dangerous decisions?

 

[LvW]

While I certainly understand that there are infinite solutions this does not really help me.:-D Should I set up the circuit in a cad/cam program without
any values attached to the individual components and go back and plug values in according to my needed output and hope for the best? Would the cad program warn me about incorrect or dangerous decisions?

No, I don't think you will get a warning. The program assumes (no, it doesn't assume - it takes for granted) that you know what you are doing. That means: You have, of course, to understand the function of the circuit as well as some principles for a good and correct dimensioning. No program can design a circuit for you. It only can visualize the consequences (good or bad) of your design strategy.

 

[Tomarino]

Ok then how do I go about finding those component values? Circuit analysis? Transfer functions? What are transfer functions and how are they used in this context? Could the cad/cam approach be used to check correctness in conjunction with other tool?
Let us look at this another way; your boss hands you an application note and says build this circuit. What would your first step be after reviewing the circuit?

 

[alexxx]

What are transfer functions and how are they used in this context?

Transfer function - Wikipedia, the free encyclopedia

Ok then how do I go about finding those component values? Circuit analysis? Transfer functions?

Circuit analysis is defenitely what you need above all. Transfer functions are usefull for sure, but even if you don't know the transfer function, you should be able to extract this transfer function on your own.
Take the inverting amplifier as an example. If you want to design a circuit like that, you use the transfer function to find out the proper resistor values for your needs. But even if you don't know or remember the transfer function (still you can google it to find out), you would come up to the same result if you did a circuit analysis.

your boss hands you an application note and says build this circuit. What would your first step be after reviewing the circuit?

Google search for what you need, find articles and start design the circuit step by step. Maybe you will get stack on the way, but still you can google again to get more information about the point you got stack. When your prototype is ready you test it, see what you designed wrong, update hardware (replace resistors, cut and create signal paths etc) and get the final result.

Could the cad/cam approach be used to check correctness in conjunction with other tool?

My design software will throw an error if I short circuit two outputs together. But no warning will be thrown if for example I connect a LED anode to 5V and ground the cathode through an 1 MOhm resistor. This LED will never be lit, but from the software point of view this is a legal connection. Maybe there are softwares that gives you such warnings who knows? But either way, this cannot replace your knowledge and understanding on a circuit. Circuit simulators are also nice tools, but as LvW pointed out:

No program can design a circuit for you. It only can visualize the consequences (good or bad) of your design strategy.

Hope that helped.

 

[Tomarino]

Thanks a lot Alex. I really appreciate the time you took for your reply. Which circuit design s/w do you use? I already have a circuit but it has no values for the components surrounding the chip they are trying to sell me. Can you point me to a text or regular book to get me started with circuit analysis?

 

[LvW]

Google, google, google....Alexxx, I cannot agree with you.
I have tried to tell houly that the main point is UNDERSTANDING. And this is something more than only to "google".

Take this example (quote alexxx):
Take the inverting amplifier as an example. If you want to design a circuit like that, you use the transfer function to find out the proper resistor values for your needs. But even if you don't know or remember the transfer function (still you can google it to find out), you would come up to the same result if you did a circuit analysis.

This doesn't help at all to solve houly's problem: No transfer function gives you an information on proper resistor values.
The simplified gain (transfer function) is -R2/R1 for an inverting amplifier. Does this really help?
No, of course not. Instead, you must know the following - and I doubt if you are succesful by googling:
The resistors R1 and R2 must be chosen according to the following rules:
* The ratio must satisfy your gain requirements,
* R1 must be small if compared with the finite opamp inverting input impedance (some kohms max. for BJT inputs);
* On the other hand, R1 must be large if compared with the signal source resistance;
* R2 must be large if compared with the finite opamp output impedance.
* But R2 must not be to large because of unwanted (parasitic) effects (dc voltage, parasitic capacitance).

As you can see (and as I have mentioned before) even such a simple task to design an inverter stage requires a trade-off.

 

[alexxx]

Which circuit design s/w do you use?

I use Pulsonix. It is not freeware.

Google, google, google....Alexxx, I cannot agree with you.
I have tried to tell houly that the main point is UNDERSTANDING. And this is something more than only to "google".

Take this example (quote alexxx):
Take the inverting amplifier as an example. If you want to design a circuit like that, you use the transfer function to find out the proper resistor values for your needs. But even if you don't know or remember the transfer function (still you can google it to find out), you would come up to the same result if you did a circuit analysis.

This doesn't help at all to solve houly's problem: No transfer function gives you an information on proper resistor values.
The simplified gain (transfer function) is -R2/R1 for an inverting amplifier. Does this really help?
No, of course not. Instead, you must know the following - and I doubt if you are succesful by googling:
The resistors R1 and R2 must be chosen according to the following rules:
* The ratio must satisfy your gain requirements,
* R1 must be small if compared with the finite opamp inverting input impedance (some kohms max. for BJT inputs);
* On the other hand, R1 must be large if compared with the signal source resistance;
* R2 must be large if compared with the finite opamp output impedance.
* But R2 must not be to large because of unwanted (parasitic) effects (dc voltage, parasitic capacitance).

None said that this is wrong, but what is the problem searching all these on internet? When someone understands the transfer function, then he will face the challenge of selecting resistors. He will defenitely search for those too. All these questions that you asked, could found answers on google why not? Will be a difference if you read it in a book or an article? The result will be the same. There are scientific articles on the web that are really good. UNDERSTANDING as you mentioned does not come from a book or from an article. It comes directly from the workbench. To seat on the bench, you need a prototype. To build a prototype you first need to do some reading. That reading includes some theory (if you are entering new areas of knowledge) and datasheets and application notes for new parts you will use. Does the source of that reading (paper or PC monitor) really matters? If you know what you want, you know what to search. Even when you design a board for the first time, you need to do some reading. I am not an expert like you I suppose (and this is not ironic, I have seen many answers from you), but when I need something I search for it and give a solution. I can't see why my point of view excludes understanding.

 

[LvW]

...........but when I need something I search for it and give a solution. I can't see why my point of view excludes understanding.

Yes alexxx, I agree with you since such an approach most probably is the right one for YOU!
But remember, tomarino has asked how to "derive" component values. It looks as he is a beginner and he even does not know that component values can (and in most, if not in all, cases: must) be selected based on some practical rules.
That is the reason I emphasize reading, thinking and understanding before searching for a solution via google.
Regards
LvW

 

[Orson Cart]

Gentlemen, it appears that the answer is that one can only design a circuit (i.e. derive useful component values) if one has had the necessary training, i.e. a technical qualification or degree followed by a period of work and technical instruction by a knowledgeable engineer. A novice trying to design circuits without this background may find themselves extremely frustrated by a lack of fundamental concept understanding - or would have to be extremely bright and well read to educate themselves in what is a failry difficult field to do well. Regards, Orson Cart.

 

[Tomarino]

So I should go and get a four year degree then to find these values? Then go work for a certain number of years then I will find the necessary values for this app note.:roll:

 

[Orson Cart]

This is what all others have had to do who can successfully calculate these values in a useful and meaningful way, however, you can read the right text books - and even online guides that will help you to get a grasp of what is required in component calculation - a lot of judgement is involved that can only be gained from experience and guidance... if it was easy - everybody would be doing it... Regards, Orson Cart.

 

[LvW]

So I should go and get a four year degree then to find these values? Then go work for a certain number of years then I will find the necessary values for this app note.:roll:

I am afraid - for some specific tasks - yes!.
To clarify the point let's take a simple example: A resistive voltage divider.

Such a two-resistor circuit is used to fix the operating point of a BJT amplifier stage.
You have the choice: 144k/72k or 36k/18k or 3.6k/1.8k or .....?
The decision requires a trade-off between power consumption and operating point stability.
But this decision cannot be made without sufficient understanding WHY such a trade-off is necessary
(why stability properties are concerned?).

 

[Orson Cart]

In this case the trade off would be between power consumption and noise - tending to use the higher values - also the min gain of the transistor at the min temp of operation comes into it, Regards Orson Cart.

 

[LvW]

Orson Cart, what about the trade-off between (a) voltage control and (b) current control, respectively, of the BJT bias condition (or in practice: somewhere in between) - and it's different behaviour associated with temperature changes and parameter tolerances?

 

[Orson Cart]

Hello LvW, yours is such an open ended question and so application specific that there is no single answer, for systems where you are using current as the controlled and measured variable then a design engineer would be interested in the gain over temp and device spread to create a system that compensated for these to provide calculable results for the various input(s) to this end. Where voltage is the variable of interest a different approach would be used in the design of BJT circuitry to give suitably compensated results over temp and device spread - often a designer will use principles of current driven operation in one part of a design and voltage in another to achieve an overall result that meets design requirements of accuracy, drift, noise. Such design is usually the preserve of experienced (10 years +) design engineers who have knowledge of a wide range of techniques and realise their applicability to the situation at hand. Regards, Orson Cart.

 

[LvW]

Hello, Orson Cart - there is nothing to add, I think. For my opinion, your reply describes very good one aspect of the background for the necessity to carefully select the parts for such a simple circuit like a resistive voltage divider - not only with respect to power consumption and noise. And exactly this was the intention of my example.
Regards
LvW