Learning electronics

[svaha]

Hi All,


Having just read through the long thread on the diy US antifouling kit it's pretty obvious there are some pretty well informed people on the forum so...

I'm wondering if it's possible to learn enough about electronics to have an idea of what's going on in a fairly simple circuit (eg perhaps the diy kit mentioned above) from any available books or do I need to go back to uni to really learn anything useful. Could anyone make any book recommendations?

Cheers Martin

 

[nigelmercier]

... long thread on the diy US antifouling kit ... Could anyone make any book recommendations?

Do you mean US antifouling kit, or the kit from Jaycar based on the Australian magazine "Silicon Chip"?. Although this is a fairly simple circuit, the main power is in the firmware in the PIC microcontroller.

Practical Electronics Handbook is excellent, I used to recommend it to my OU students.

If you want to know more about PIC microcontrollers, then a good start is 50 PIC Microcontroller Projects: For Beginners & Experts, although the programming language used is not necessarily the way forward.

Another good PIC book, this time dealing with machine language is The PIC Microcontroller: Your Personal Introductory Course. It's worth understanding some of this before moving on.

Once you have got the basics of machine language, you can move onto a real-life programming language named "C". Programming 8-bit PIC Microcontrollers in C: with Interactive Hardware Simulation is my choice of starting point.

 

[electrosys]

A very good book for people of all abilities is:
The Art of Electronics by Horowitz and Hill.

I believe there's a copy of the 2nd edition floating around on file-sharing sites. The first few chapters should be enough for your purposes - don't get bogged down by the maths - just grasp the concepts at first. You can always revisit the maths later.

 

[prv]

Do you mean US antifouling kit, or the kit from Jaycar based on the Australian magazine "Silicon Chip"?.

I think he meant "US" as in "ultrasonic", not as in "United States".

Pete

 

[Sybarite]

Didn't Nigel Calder do a basic initiation book?

 

[st599]

Do you mean US antifouling kit, or the kit from Jaycar based on the Australian magazine "Silicon Chip"?. Although this is a fairly simple circuit, the main power is in the firmware in the PIC microcontroller.

Practical Electronics Handbook is excellent, I used to recommend it to my OU students.

If you want to know more about PIC microcontrollers, then a good start is 50 PIC Microcontroller Projects: For Beginners & Experts, although the programming language used is not necessarily the way forward.

Another good PIC book, this time dealing with machine language is The PIC Microcontroller: Your Personal Introductory Course. It's worth understanding some of this before moving on.

Once you have got the basics of machine language, you can move onto a real-life programming language named "C". Programming 8-bit PIC Microcontrollers in C: with Interactive Hardware Simulation is my choice of starting point.

Or buy an arduino and have a play

 

[Landale]

Instead of PIC you might consider Arduino http://arduino.cc/en/ as it might be possible to use one of the boards without having to get to grips with the infrastructure the PIC needs - however that is really just several capacitors, a crystal or resonator and a 5V voltage regulator.

 

[svaha]

I think he meant "US" as in "ultrasonic", not as in "United States".

Pete

Yeah I did, sorry for the confusion.

Thanks for the suggestions



Cheers m.

 

[jdc]

H&H

A very good book for people of all abilities is:
The Art of Electronics by Horowitz and Hill.

I believe there's a copy of the 2nd edition floating around on file-sharing sites. The first few chapters should be enough for your purposes - don't get bogged down by the maths - just grasp the concepts at first. You can always revisit the maths later.

+1 for Horowitz and Hill. It is used by professionals as well as hobbyists. Old editions (from the mid 80s) are still good.


I think it an excellent idea to not get bogged down by the maths on the first read, but DO come back to it; it's not difficult, and if you try to do electronics without at least Ohms law and simple AC circuit analysis you're doomed to failure. The level 1 MOSFET or Ebbers-Moll bipolar equations are trivial also (no function more complicated than a square, square-root or exponential: O level stuff, so don't be frightened).

Some of my 'rules':
- You should know why the value of every component is as it is / what it should be
- For that reason there should never be a 'select on test' or trimmable component (aka variable resistor) unless it's for a variable (eg volume control) or to calibrate a some true unknown such as a sensor response (and even here they're better avoided if possible)
- All circuit design should include a calculation of the allowable tolerances (you can normally use a spreadsheet)
- No circuit should rely on any particular value of transistor characteristics, eg hfe or Vth or diode Voltage drop, only that its in a range
- All circuit designs should work first time (in contrast to software which usually does need debugging)

 

[Landale]

+2 for Horowitz and Hill. - the latest version is even better

 

[nigelmercier]

I think he meant "US" as in "ultrasonic", not as in "United States".

D'OH!

 

[nigelmercier]

A very good book for people of all abilities is:
The Art of Electronics by Horowitz and Hill.

Yes, The Art of Electronics is excellent, but quite expensive. It is also quite dated.

 

[electrosys]

One other suggestion is to acquire a basic simulation program, such as ElectronicWorkBench, so you can see how circuits function in practice, and what effect the substitution of components has.

EWB comes with it's own in-built multimeter, 'scope, signal generator etc, and apart from the serious business of learning a subject, it's also great fun to simply play around with ... !

EWB versions 5 to 8 are freely available on the web. One source (of many) is: http://electronics-workbench-ewb5.0.fyxm.net/
EWB will run on any Windows platform from Win9X upwards.

 

[prv]

if you try to do electronics without at least Ohms law and simple AC circuit analysis you're doomed to failure. The level 1 MOSFET or Ebbers-Moll bipolar equations are trivial also (no function more complicated than a square, square-root or exponential: O level stuff, so don't be frightened).

Some of my 'rules':
- You should know why the value of every component is as it is / what it should be
- For that reason there should never be a 'select on test' or trimmable component (aka variable resistor) unless it's for a variable (eg volume control) or to calibrate a some true unknown such as a sensor response (and even here they're better avoided if possible)
- All circuit design should include a calculation of the allowable tolerances (you can normally use a spreadsheet)
- No circuit should rely on any particular value of transistor characteristics, eg hfe or Vth or diode Voltage drop, only that its in a range
- All circuit designs should work first time (in contrast to software which usually does need debugging)

Would you apply the same rules to digital stuff that consists mostly of a microprocessor and a few supporting components? As a professional software engineer, if I were to dabble in hardware that's probably the approach I'd choose. I admit to being somewhat intimidated by analogue electronics, and wary even of digital stuff made of discrete components

Pete

 

[grumpy_o_g]

Would you apply the same rules to digital stuff that consists mostly of a microprocessor and a few supporting components? As a professional software engineer, if I were to dabble in hardware that's probably the approach I'd choose. I admit to being somewhat intimidated by analogue electronics, and wary even of digital stuff made of discrete components

Pete

Probably yes if the digital stuff was an encoded analogue signal - although it affects the result differently, variation in tolerances still alter the output from the intended result, induction and capacitance at high frequencies being an especially tricky problem. Don't think bunging a high-frequency crystal-tuned oscillator in there will stabilise the entire circuit for example.

If the circuit, or a discrete part, is genuinely digital throughout (e.g. +5V at one point, 0V on another relative a common rail causes it output +12V at another point to switch a relay) then not quite but remember that you have make sure that all components are sufficiently in tolerance or you may end up with 2V on the 0V rail and only 3V on the 5V rail at which point you're probably out of luck.

I'd disagree slightly with jdc though. I have no problem with calibration components (not select on test but a genuinely variable component) but we aren't talking about a little trimming pot or anything like that. If you can introduce any kind of negative feedback (PLL) or shove a Zener in (or use anything with a good constant voltage across a junction) and use that to provide a reference voltage.

If you're a software engineer I'd actually play with any of the languages (doesn't need to be arduino though you'll get lots of help with that) but use pre-built assemblies for some stuff at first at least. Buy a power supply for example - if the demands aren't too precise or severe then they're easy to build (though they are buggars if you need a a very precise smooth voltage over a wide current draw!!). But, if you buy one, you can get on with another bit of the circuit that has to be bespoke and you'll save a lot of time - which means you see results earlier. Actually buying the PS has another advantage in that it means you can stop worrying about the high voltage side of things too, cause it's all neatly tucked away. Some people do sneer at this but even so called discrete components are manufactured for you, it's just a question of where to draw the line.

Other big point, like any software coding that's more than a few lines, make sure it's a structured approach to a structured product.

Requirements (set in stone)
Spec and design (can be varied as you proceed but this is functionality, size, scalability, etc)
Modular layout (design discrete sections of the circuit such as i/o modules, tuners, amps, drivers, etc.)
Signal flow diagrams, power and common rail diagrams, etc.
Circuit design and layout (where you actually draw the circuit diagram)
Physical design and layout (which can significantly affect the behaviour of components - depending on what you're designing you might start doing some heat and airflow modelling and thermal calculations at this point).


The whole thing is iterative and is very similar to designing, speccing and coding an app, especially in that the whole thing quickly becomes a bag of worms if you try and change the requirements halfway through.

 

[nigelmercier]

I admit to being somewhat intimidated by analogue electronics, and wary even of digital stuff made of discrete components

I quite agree, I like my transistors either on or off