Electronics is easy!

[Phaedrus]

I was having a discussion with my (non-EE) friend about electrical engineering. He didn't understand why it's so easy for him to understand simple transistor/resistor/diode/capacitor circuits, but so hard to understand anything I say when I start on about SMPS. I told him this:

Electronics is super easy -- as long as you stay inside of these parameters:

<100V
<100W
<10A
<10MHz
<1000 lines of code

Inside of those parameters, electronics engineering is a piece of cake. Unfortunately, just about everything useful inside those parameters has already been done. Which means if you want to do something useful/interesting, you're basically forced to venture into territory where key information has to be wheedled out of graybeards and 10-year-old blog posts, or derived by maths, or discovered through trial and error, or is only accessible in a datasheet locked behind a paywall. Or there is no datasheet. Then it isn't easy.

[c4757p]

<100V
<100W
<10A
<10MHz
<1000 lines of code

>5%

[skipjackrc4]

<100V
<100W
<10A
<10MHz
<1000 lines of code

>5%

>1uV
>1nA
>1mOhm

[coppice]

I was having a discussion with my (non-EE) friend about electrical engineering. He didn't understand why it's so easy for him to understand simple transistor/resistor/diode/capacitor circuits, but so hard to understand anything I say when I start on about SMPS. I told him this:

Electronics is super easy -- as long as you stay inside of these parameters:

<100V
<100W
<10A
<10MHz
<1000 lines of code

Inside of those parameters, electronics engineering is a piece of cake. Unfortunately, just about everything useful inside those parameters has already been done. Which means if you want to do something useful/interesting, you're basically forced to venture into territory where key information has to be wheedled out of graybeards and 10-year-old blog posts, or derived by maths, or discovered through trial and error, or is only accessible in a datasheet locked behind a paywall. Or there is no datasheet. Then it isn't easy.

Watch a good documentary about something deep, and it all makes sense. Try to contribute to something deep and it gets really hard, really quickly.

[rs20]

I was having a discussion with my (non-EE) friend about electrical engineering. He didn't understand why it's so easy for him to understand simple transistor/resistor/diode/capacitor circuits, but so hard to understand anything I say when I start on about SMPS. I told him this:

Electronics is super easy -- as long as you stay inside of these parameters:

<100V
<100W
<10A
<10MHz
<1000 lines of code

Inside of those parameters, electronics engineering is a piece of cake. Unfortunately, just about everything useful inside those parameters has already been done. Which means if you want to do something useful/interesting, you're basically forced to venture into territory where key information has to be wheedled out of graybeards and 10-year-old blog posts, or derived by maths, or discovered through trial and error, or is only accessible in a datasheet locked behind a paywall. Or there is no datasheet. Then it isn't easy.

Interesting discussion points, but <pedantic>this doesn't even come close to answering your friend's question about why he can't understand an SMPS -- an SMPS is often < 100V, <100W etc etc.</pedantic>  Maybe he just doesn't have a good intuitive understanding of an inductor?

[AlfBaz]

to drag out an old tired cliché, the more you learn the more you realise how little you know.

[hamster_nz]

  Maybe he just doesn't have a good intuitive understanding of an inductor?

I don't think anybody has an intuitive understanding of induction! How can the energy flowing through a conductor to create a magnetic field that stores the energy, and that can then case energy to flow through another nearby conductor?  it doesn't happen with water in a pipes, so why does it happen with electricity?

The closest analogy to induction that I can think of is a turbine with a flywheel on it, where the energy in the flowing through it converts to kenetic energy stored in the flywheel. If you suddenly block the outflow, the energy stored in the turbine keeps trying to push the fluid, and pressure builds up far higher than the original flow. A transformer is much the same, but instead of the flywheel there is a pump that can move a secondary flow around at a different pressure or rate of flow.

Is the universe filled with little turbines? not really, and so the analogy falls apart pretty quickly.

[c4757p]

It's a lot easier if you ditch the awful water analogy.

[BravoV]

It's a lot easier if you ditch the awful water analogy.

Its not easy to find better alternatives other than water analogy when it comes to explain the basic to a complete noob. 

[coppice]

to drag out an old tired cliché, the more you learn the more you realise how little you know.

Or the more elegantly worded:

"The fundamental cause of trouble in the world today is that the stupid are cocksure while the intelligent are full of doubt" - Bertrand Russell

[rs20]

  Maybe he just doesn't have a good intuitive understanding of an inductor?

I don't think anybody has an intuitive understanding of induction! How can the energy flowing through a conductor to create a magnetic field that stores the energy, and that can then case energy to flow through another nearby conductor?  it doesn't happen with water in a pipes, so why does it happen with electricity?

The closest analogy to induction that I can think of is a turbine with a flywheel on it, where the energy in the flowing through it converts to kenetic energy stored in the flywheel. If you suddenly block the outflow, the energy stored in the turbine keeps trying to push the fluid, and pressure builds up far higher than the original flow. A transformer is much the same, but instead of the flywheel there is a pump that can move a secondary flow around at a different pressure or rate of flow.

It does happen with water in pipes, the inertia of moving water behaves in a corresponding way -- look up the water hammer effect. And yes, a turbine (specifically, the one in hydraulic analog illustrations that completely blocks flow until the wheel is turning) is a perfect analogy of an inductor.  A hydraulic ram is a perfect component-for-component analogy for a DC-DC boost converter, etc etc.  Transformers are much more difficult, but the correct response there is to ditch the water analogy when working with transformers, not to declare the water analogy broken under all conditions*.

Is the universe filled with little turbines? not really, and so the analogy falls apart pretty quickly.

The analogy is from inductor to turbine, energy in magnetic field to kinetic energy, and the equations and behaviour line up perfectly, so you have all you need. If you want a deeper understanding of how inductors work, then obviously you need to ditch the water analogy, but the point of the analogy is to understand the behaviour and interactions of the components, not their underlying workings -- even the analogy for a simple wire vs pipe is awful for that purpose.

It's a lot easier if you ditch the awful water analogy.

You don't like the water analogy, but a lot of people do find it very helpful, especially for starting concepts like current and voltage. I suspect a lot of people dislike it because I see poor analogies suggested often -- open buckets (that disobey KCL) instead of springed-piston-in-pipe for capacitors, etc. But I'm sure there's already a hundred other water analogy debate threads on this forum... * I do support weaning off the water analogy as your confidence grows, since a natural intuitive understanding is always going to be better.

[dannyf]

to drag out an old tired cliché, the more you learn the more you realise how little you know.

Absolutely.

That's why I always smile when someone tells this "this is settled science",

[schopi68]

sounds to me like an "Everything that can be Invented has been Invented"   

ehm... what has been the topic's question?

[Tac Eht Xilef]

to drag out an old tired cliché, the more you learn the more you realise how little you know.

Absolutely.

That's why I always smile when someone tells this "this is settled science",

You'd be surprised how often actual scientists never say that. Everytime I've heard it it's been from a first-year student, an engineer or some other techie-type, or just some nerd who hasn't gotten over Heinlein or Ayn Rand yet...

[dannyf]

You'd be surprised how often actual scientists never say that.

Agreed. That's my way of telling a real scientist from a climate "scientist",

[Bored@Work]

You'd be surprised how often actual scientists never say that.

Agreed. That's my way of telling a real scientist from a climate "scientist",

Guys, I have this overunit device for sale. It is driven by the ghost of Tesla and magnets. It is not settled science, so must work.

[SeanB]

The only thing that you can be absolutely sure of is that everything you think you know is invariably at some time going to have an edge case where you suddenly realise that what you thought you know id just an edge case of a larger more general theory. The things you know are just an approximation when the measurement is smaller than a certain value of something.

Best analogy is Newtonian mechanics, which works for all cases where v is very much less than 0.001C, after which the correction factors become noticeable to sensitive enough measurement. Above 0.01C it becomes visible to the gross senses of humans, and above 0.1C very visible and a large source of error from the equations of Newton. Then again Einsteinian physics is a subset of another branch.

Most scientists are very settled in the branch they deal with, and often you really have to dig to get them out of the inertial field that they keep to. You need new people with the unfettered thinking that says "Mmmmmm, why did it do that now?", and then you have that Quantum Leap that gave us that funny theory that fitted the observations of the known universe, so dismissed by Hoyle with the throwaway comment of "The big bang".

[jlmoon]

I was having a discussion with my (non-EE) friend about electrical engineering. He didn't understand why it's so easy for him to understand simple transistor/resistor/diode/capacitor circuits, but so hard to understand anything I say when I start on about SMPS. I told him this:

Electronics is super easy -- as long as you stay inside of these parameters:

<100V
<100W
<10A
<10MHz
<1000 lines of code

Inside of those parameters, electronics engineering is a piece of cake. Unfortunately, just about everything useful inside those parameters has already been done. Which means if you want to do something useful/interesting, you're basically forced to venture into territory where key information has to be wheedled out of graybeards and 10-year-old blog posts, or derived by maths, or discovered through trial and error, or is only accessible in a datasheet locked behind a paywall. Or there is no datasheet. Then it isn't easy.

Interesting discussion points, but <pedantic>this doesn't even come close to answering your friend's question about why he can't understand an SMPS -- an SMPS is often < 100V, <100W etc etc.</pedantic>  Maybe he just doesn't have a good intuitive understanding of an inductor?

SMPS are where all the magic smoke hides!

[Syntax_Error]

I was having a discussion with my (non-EE) friend about electrical engineering. He didn't understand why it's so easy for him to understand simple transistor/resistor/diode/capacitor circuits, but so hard to understand anything I say when I start on about SMPS. I told him this:

Electronics is super easy -- as long as you stay inside of these parameters:

<100V
<100W
<10A
<10MHz
<1000 lines of code

Inside of those parameters, electronics engineering is a piece of cake. Unfortunately, just about everything useful inside those parameters has already been done. Which means if you want to do something useful/interesting, you're basically forced to venture into territory where key information has to be wheedled out of graybeards and 10-year-old blog posts, or derived by maths, or discovered through trial and error, or is only accessible in a datasheet locked behind a paywall. Or there is no datasheet. Then it isn't easy.

Why is this so? You have my curiosity. I imagine these may be soft constraints, but I am curious about the differences between low power, low frequency, small small code (HDL or executable? Both?) and the higher power/voltage/size/speed variants.

In all honesty, I can sort of guess why, but I suspect my reasons are incomplete. I know high frequency work has a reputation for difficulty. I won't pretend to know first-hand, but it seems to be due to distributed parameters, nonideal components, and having to consider E&M phenomena that are not players in the slow game.

But I can't make similar claims for high amperage, voltage, power, or code size, so I'm genuinely curious what gives these systems such a leap in difficulty, other than for code size, where bigger is more error prone and simply just takes longer to code, debug, or analyze outright. Perhaps it has do to with the rise and fall time constraints contraindicating the component capacitances with the voltages involved? I can only speculate.

[ajb]

There's also a point where you go from being able to get by with basic arithmetic and a couple of formulas to analyze a circuit to suddenly needing differential calculus and complex numbers to analyze transfer functions and frequency dependency.  Considering how most non-EE people I know feel about basic algebra, it's no surprise that EE goes from science to arcane sorcery around the same point.

I was having a discussion with my (non-EE) friend about electrical engineering. He didn't understand why it's so easy for him to understand simple transistor/resistor/diode/capacitor circuits, but so hard to understand anything I say when I start on about SMPS. I told him this:

Electronics is super easy -- as long as you stay inside of these parameters:

<100V
<100W
<10A
<10MHz
<1000 lines of code

Inside of those parameters, electronics engineering is a piece of cake. Unfortunately, just about everything useful inside those parameters has already been done. Which means if you want to do something useful/interesting, you're basically forced to venture into territory where key information has to be wheedled out of graybeards and 10-year-old blog posts, or derived by maths, or discovered through trial and error, or is only accessible in a datasheet locked behind a paywall. Or there is no datasheet. Then it isn't easy.

Why is this so? You have my curiosity. I imagine these may be soft constraints, but I am curious about the differences between low power, low frequency, small small code (HDL or executable? Both?) and the higher power/voltage/size/speed variants.

In all honesty, I can sort of guess why, but I suspect my reasons are incomplete. I know high frequency work has a reputation for difficulty. I won't pretend to know first-hand, but it seems to be due to distributed parameters, nonideal components, and having to consider E&M phenomena that are not players in the slow game.

But I can't make similar claims for high amperage, voltage, power, or code size, so I'm genuinely curious what gives these systems such a leap in difficulty, other than for code size, where bigger is more error prone and simply just takes longer to code, debug, or analyze outright. Perhaps it has do to with the rise and fall time constraints contraindicating the component capacitances with the voltages involved? I can only speculate.

With AC systems (including SMPS and RF signals) a lot of the difficulty can be related to slew rates.   At very high dV/dt or dI/dt (where higher frequency->smaller dt) a lot of otherwise inconsequential parasitics (capacitance and inductance) come into play. 

In DC systems with high currents even small voltage drops add up to substantial power losses--especially resistive losses which increase with the square of current.   With high voltages you have to be increasingly concerned with the voltages your components (including the PCB layout) can withstand, and even small leakage currents result in substantial power loss. 

A switching power supply is sort of the trifecta of these effects, where there's an impetus towards high switching speed (to minimize switching losses) and high switching frequency (to minimize inductor size and control ripple frequency) both of which involve high slew rates.  When very high voltage differentials or output currents are required, the problems all compound.

[tom66]

Interesting discussion points, but <pedantic>this doesn't even come close to answering your friend's question about why he can't understand an SMPS -- an SMPS is often < 100V, <100W etc etc.</pedantic>  Maybe he just doesn't have a good intuitive understanding of an inductor?

When I was first learning the design of buck converters and boost converters, I used two analogies to help me learn.

A buck converter is basically an LC filter, even with a diode, you can consider the diode to be a switch, so the LC filter is driven by a square wave. The "filter" should ideally remove as much of the ac ripple, leaving only dc output.

A  boost converter is basically taking the flyback pulse you'd get from a relay coil, or other coil, and using that energy usefully. It's "charging" an inductor on one cycle, and discharging it into the output capacitance through a diode, or a second switch for a synchronous type.

Of course, once you get over the basic analogy, and actually learn WHY these things happen, it's far easier.

[Syntax_Error]

There's also a point where you go from being able to get by with basic arithmetic and a couple of formulas to analyze a circuit to suddenly needing differential calculus and complex numbers to analyze transfer functions and frequency dependency.  Considering how most non-EE people I know feel about basic algebra, it's no surprise that EE goes from science to arcane sorcery around the same point.

I can see how this is true, but couldn't the same be said for *all* engineering disciplines, as well as physics work? All employ calculus and differential equations extensively. In fact, not using these tools in their education is considered laughable. Not to imply that such math is a trivial matter, because it is not. EE uses complex numbers and linear algebra, but ME uses probability and statistics. In your experience, are MEs  averse to their math, while EEs are not?

With AC systems (including SMPS and RF signals) a lot of the difficulty can be related to slew rates.   At very high dV/dt or dI/dt (where higher frequency->smaller dt) a lot of otherwise inconsequential parasitics (capacitance and inductance) come into play. 

In DC systems with high currents even small voltage drops add up to substantial power losses--especially resistive losses which increase with the square of current.   With high voltages you have to be increasingly concerned with the voltages your components (including the PCB layout) can withstand, and even small leakage currents result in substantial power loss. 

A switching power supply is sort of the trifecta of these effects, where there's an impetus towards high switching speed (to minimize switching losses) and high switching frequency (to minimize inductor size and control ripple frequency) both of which involve high slew rates.  When very high voltage differentials or output currents are required, the problems all compound.

Thanks for this. This makes sense, and is more than I thought anyone would say. I thought someone would mention EMC issues with the switching speeds employed, but I hadn't considered these things. I kind of always figured your average Joe Engineer was very comfortable with multivariable calculus and systems of PDEs, etc. I have heard for years now how any EET can cobble together chips and route their connections per somebody's app note, but an EE can use mathematics to analyze a problem and creatively derive and deploy a solution, and write the app notes for it in the process.

[Tac Eht Xilef]

You'd be surprised how often actual scientists never say that.

Agreed. That's my way of telling a real scientist from a climate "scientist",

As long as you apply the same level of critical evaluation to the ones you agree with...

There's also a point where you go from being able to get by with basic arithmetic and a couple of formulas to analyze a circuit to suddenly needing differential calculus and complex numbers to analyze transfer functions and frequency dependency.  Considering how most non-EE people I know feel about basic algebra, it's no surprise that EE goes from science to arcane sorcery around the same point.

A purpose of good analogies is to get you a bit further past that hump so you can grasp a few fundamentals of the 'arcane sorcery' and see why learning them helps.

I've always been partial to the 'compressed spring / rubber sheet' version of the water analogy for inductors - electricity is the force that compresses the spring / distorts the rubber sheet, and that also affects things that are coupled / close to it; when that force is removed the spring / sheet pushes back in the opposite direction. You can even extrapolate that analogy a bit with other real-world knowledge and come up with correlations parallels to things like slew rate & hysteresis.

Of course, that skates over the whole 'electricity <-> magnetism' relationship (as well as a few other things) - but that's why it's only an analogy...

[ajb]

There's also a point where you go from being able to get by with basic arithmetic and a couple of formulas to analyze a circuit to suddenly needing differential calculus and complex numbers to analyze transfer functions and frequency dependency.  Considering how most non-EE people I know feel about basic algebra, it's no surprise that EE goes from science to arcane sorcery around the same point.

I can see how this is true, but couldn't the same be said for *all* engineering disciplines, as well as physics work? All employ calculus and differential equations extensively. In fact, not using these tools in their education is considered laughable. Not to imply that such math is a trivial matter, because it is not. EE uses complex numbers and linear algebra, but ME uses probability and statistics. In your experience, are MEs  averse to their math, while EEs are not?

Well, I wasn't actually thinking of other engineers when I wrote that so much as the general population, including up to a certain level of hobbyist/tinkerer.  I think you're right that in any engineering discipline, beyond some threshold you can't just plug some numbers into a case formula and get a solution anymore.  That said, I suspect that there are a substantial number of engineers (of all types) out there who mostly do just that on a daily basis.  That's not necessarily because they aren't capable of or willing to do the higher level work, it's because such a wide range of real world problems fall well within the domain where the higher level stuff isn't needed.  It's the same reason that things like Arduino have been so successful--people have found that they can do cool stuff with electronics with little or no formal knowledge, and without any hairy mathematics.  Of course on the other hand you don't see those sorts of people producing much high precision or high power stuff, because at a certain point that knowledge (and the hairy math!) become crucial to achieving a working design.

[Galenbo]

<100V
<100W
<10A
<10MHz
<1000 lines of code

>5%

>1uV
>1nA
>1mOhm

I can add:

Without earth and ground loop problems
With good documentation
With my browser next to me
With modern components
In a good lab
Without having to take account about CE or other medical/army/secret regulations
Without Dilbert's boss