Reading schematics

[keninverse]

I never uderstood this.
almost all schematics that I've seen don't include any current measurements.  so my question is how do figure out how much voltage is dropping across a resistor if you don't know how much current is available?  I guess this is dependant on the circuit but is this an experimental sorta thing where you have to measure it or is it purely empirical.

anybody got any links?

TIA.

[Hal]

ususally it doesnt matter.  If you really wanna know, wire it up and use a meter.  But including information like that in a schematic would be rediculus.

[smashinator]

It seems to me that you could figure that out if you know the voltage and the resistance using Ohm's law.

...which, being a horrible person, I don't have memorized.   Is it Voltage=Current * Resistance?

If so, Voltage/Resistance=Current.

[niftydog]

INTRODUCING...
THE MAGIC TRIANGLE!


Now you'll remember it forever!

It's fairly intuitive, but just in case;

V=IR
I=V/R
R=V/I

The same can be done for the power equation P=IV

[niftydog]

oh yeah... most electronic devices are "voltage controlled". Meaning, there operation depends on the voltage present at the terminals. Less often you'll find "current controlled" devices.

So usually you want to know the voltage at a point. An even then, DC voltages won't necessarily tell you how a circuit is working.

[Satch12879]

::soapbox mode on::

Is this a good question? I mean shouldn't we know how to calculate current, voltages, etc. in the circuits we're designing without having to prototype and measure by hand? Isn't that the point of design? Or is it just my Master's Degree in Engineering talking?

Listen you'll never hear a structural engineer say "Well, let's build it first and then we'll see if it can handle the load."  That gets people killed.

::soapbox mode off::

I'd say invest in a good circuit analysis text, learn the methods, and I think you'll be on the way to an answer.

[smashinator]

Listen you'll never hear a structural engineer say "Well, let's build it first and then we'll see if it can handle the load."  

Structural engineers are Pansies!   :lol:

[niftydog]

"Well, let's build it first and then we'll see if it can handle the load."

no, but they too make a lot of assumptions when making new and unique designs. They also rely a lot on past experiences of their engineering forefathers. Same deal here. Show me a structural engineer that doesn't have sweaty palms on the day of the grand opening of a new stucture and I'll show you a completely confident electronic engineer!

If I were building something that might potentially kill someone, then yes, I'd be far more thorough. (In fact, medical electronic equipment is subject to incredibly stringent design rules) But, particularly with guitar electronics, you're dealing with miniscule currents and very tolerant components. I mean, throw a signal at an op amp and it just works; No need to pick it apart!

[Boofhead]

There's two main types of voltages on schematics.   Power supplies and measurement points.   The power supply voltage obviously lets you build the schematic as intended.  The measurement points are used to help diagnose problems in the circuit due to a fault, or, to verify it is working within parameters.

The reason you don't often see currents for problem diagnosis is because to measure current you have to insert a meter into the circuit.   Such intrusive meausrements are undesirable.  Moreover with the correct voltage measurement points it is possible to work out the current anyway - so the less intrusive voltage measurement is more popular.

As far as knowing the currents in a circuit is concerned: this can be done through circuit analysis, by design, or by measurement.    Just because they don't appear on the schematic doesn't mean they have to found through empirical means.  Some currents in a circuit have wide tolerances, due to part tolerances. So if the same circuit is built many times the currents (and usually the voltages) will vary from circuit to circuit.  An example of this is the transistor base current it (usually) varies a lot because the transistor gain varies a lot.

[Hal]

when i was saying "it doesn't matter" that was refering to 'reading a schematic.'

Obviously when designing, you need to know whats going on.  But lets say I take an LFO stage from something, and plop it in something else.  I don't really need to know whats going on, in terms of voltage and current, just the final output.  And if I really wanted to know, I could measure.  And thats still "circuit design," kinda.  To a point.  

Thats the point i was trying to make...

[R.G.]

Is this a good question? I mean shouldn't we know how to calculate current, voltages, etc. in the circuits we're designing without having to prototype and measure by hand?

Thank you Lord - someone is sniffing around the real nature of the concept of "design".

Design is NOT:
- subbing in parts of different values til it sounds good
- taking some already canned circuit and removing/adding this and that until you find your own beautiful reality
- copying some schematic and then hooking other junk up to it.
- imagining some function and how the box would look pretty, then asking someone to just go build one since all the hard work's already done.

Design IS:
- a rational process of defining some function to be performed
- selecting a circuit basis for doing something similar
- using accumulated knowledge of Mother Nature's laws and previous designs, setting up a circuit to match the pre-defined function to be performed
- selecting things that just must be to get what you want to happen, then calculating parts values to get that to happen, and knowing when the numbers are telling you that it won't happen, and you have to start with another departure point
- knowing from experience what thing just must be to get what you want to happen

The simple answer to the quoted question is "Well, duuhhh, yes. We should insist on being able to calculate the desired voltages, currents, power dissipations, frequency and phase responses for the full set of parts we have selected beforehand, without having to build a prototype.

The prototype is best viewed as a consultation with Mother Nature, asking Her to please point out where you may have forgotten to consider something.

Reverse engineering is a great way to learn design. Take a circuit, with the parts values specified and the power supplies know. Work your way through the circuit, and through logic and calculation, define the voltages and currents at each pin of each part. You may use the datasheets of the parts involved, a calculator, and as much time as you need.

When you can get through with a good reverse engineering on three transistor (or two opamp) circuits, you are now in a position to begin thinking about designing something forwards.

Sorry - the casual nature of the use of the word "design" by people who don't read schematics gets on my nerves sometimes. At least **try** to snatch the pebble.

[niftydog]

R&D = Rip-off & Duplicate.  

[Hal]

R.G. - yes, that is design for the EE's out there.  But the EE's should already know ohm's law, so whats the point of the question...?

I know its circular reasoning...but...I think you should see my point.

[mlabbee]

R.G. - overall, I agree with your point. I think that we should be able to go into a design with a general understanding of the circuit we wish to build.  However, when it comes to audio, the end results are so subjective that a certain amount of experimentation is essential.  Even with something as simple as a TS clone, I spent hours playing around with input and output buffers and clipping diode arrangements until I had something I liked.

[keninverse]

sorry my first post was a little cryptic.  But I was specifically referring to circuit design and/or re-engineering.  (BTW the and/ors are great)

V = X
I = Y
R = given number on schematic.

two variables = no answer

calculus can yield a range of numbers, but if current measurements are avail.  then we can calulate resistance...

[Paul Perry (Frostwave)]

Design is NOT:
- subbing in parts of different values til it sounds good
- taking some already canned circuit and removing/adding this and that until you find your own beautiful reality
- copying some schematic and then hooking other junk up to it.
- imagining some function and how the box would look pretty, then asking someone to just go build one since all the hard work's already done.

Forgive me Father, for I have sinned :wink:

Seriously, I have done ALL the 'nots' at one time or another, as well as all the 'dos'. Let's not forget that the first fuzz (so far as I know) came about from accidentally putting the 'wrong' resistor in a standard textbook amplifier of the time.
Looking at the circuits posted by peopel from this board, I can see that ALL the techniques have been used at various times, and there are some that the RG approach would never have led to. Of course, that isn't to take away from any of RG's designs, which would never have appeared spontaneously from a pile of scrap parts and a spliff...

"From each according to his ability, to each according to his need."
And now I'm inspired to do the whole lot of them again!! WOOHOO!

[R.G.]

Paul says:

Forgive me Father, for I have sinned  
Seriously, I have done ALL the 'nots' at one time or another

So have I. But it's not design.

The 'nots' cover a wide range of things, from raw creativity through clueless wandering. And wandering around, you do occasionally discover a gold nugget.

But once you've found a nugget, the very next thing you will do is to try to figure out how to get another one, and you have to decide how to do that.

Do you pocket the nugget and start wandering around again?

Or do you think about how and wherefrom the gold nugget came to be where you found it, and how you might trace it back to the source, whether by tracing the streambed back up stream, or digging near the location of the nugget? Tracing and/or digging are much more likely to turn up another nugget than doing more wandering.

You perhaps more than most of all the inhabitants here understand the difference between easter-egging and design.

And keninverse says:

sorry my first post was a little cryptic. But I was specifically referring to circuit design and/or re-engineering.

So was I. It is possible, given only the datasheets for chips, transistors, resistors, and caps, the schematic and power supply voltages plus some modicum of background knowledge to figure out all of the voltages and currents.

First, the obvious - if you have a resistance and a voltage, you know the current.

If you have a silicon transistor, you know the base-emitter voltage within close bounds.

That means you know the emitter voltage if you know the base voltage, and vice versa.

If you know the current gain even approximately  (like, say... it's greater than 100) you know that the external resistors determine entirely what the transistor does at DC.

From there, you typically have the bias point (from the base bias network) and the static voltages and currents, even if that is fully saturated, or completely off.

If it's an amplifier, you also know it's typically somewhere between saturated and full off, so you can check your work for sanity.

Opamp circuits are usually even easier. Check the bias, review where the + input is sitting, determine whether it is possible for a non-damaged opamp to have its - input at that voltage given the feedback network, and you have voltages and currents for all the resistors.

This is not a classic N-equations-in-N-unknowns problem. You do have to know things like how opamps and transistors work, what junction voltages are, limits on current gain, transconductance, etc. Those things supply the missing variables.

Often you'll have to guess at an operating point. Once you see what that guess does, you can then recalculate things to see how close you were. That lets you make a better guess. This iterative process usually converges within a couple of iterations. Iterative approximation is a technique well beloved of the programming simulation people, hallowed as (among other things) Newton-Raphson integration.

It's not a calculus problem. It's an insight and logical inference problem.

[keninverse]

RG Wrote:

It's an insight and logical inference problem.

And there it is...
Thanks for the help RG.  I'll start thinking more instead of just building.