How do you learn about everything?

[JackP]

Hi all,
  I thought this should go into General Chat, since there is no specific answer to my query. Since I started in electronics with Arduino, I still don't really know anything about using external ICs, with a uC or even just a peripheral (e.g. an oled or camera module). I have a rudimentary understanding of basic building block components, but coming to slightly more complex ICs, I am well out of my depth. I have found that this is starting to limit my capacity in creating projects, so I would like to know how you become knowledgeable about a great deal of chips, or even just how to find out about one when it is needed (obviously no one knows every chip out there in any detail).

  An example of this is a recent project idea: An AVR controlled camera, using a cheap camera module and an OLED. It seemed like a great idea, and I probably could have hashed something together on breakout boards. However, when I thought about turning it into an actual 'product' (an exercise in DFM perhaps) I realised that I had no idea about hooking up an OLED to a chip that allows I2C, which is already done on the breakout board. I encountered similar problems when thinking about the camera module, as well as storage; RAM, Flash, EEPROM? I had no idea.

  So I am asking you, how do I gain the ability to find out about these things? I should know about the memory aspect, but what about the ICs used? Is it even necessary to use the separate chip, can I control a peripheral directly? Any advice would be greatly appreciated.
Regards,
Jack

EDIT: What about protocols? Can anyone recommend a good source of information for the protocols in common use? Months of searching have left me no better off.

[IanB]

This is an interesting question.

Building a body of knowledge as a bit like building a framework. At first, each beam and strut is standing by itself without much support and the whole edifice is a bit wobbly. But over time, as you add more pieces to the framework they start to support each other and the whole thing becomes firmer. Once much of the framework is in place you can even see where the gaps are, and can start to predict what is needed to fill in those gaps.

If you attend formal education, such as obtaining an engineering degree, the program is designed to give you all the main pieces of that framework. There will be many missing details and it won't look like a finished house, but it will be a very good starting point to build on.

If you are starting from scratch and educating yourself, it becomes harder. You can still build your framework, but you won't have a guide or a plan to follow. The best thing to do will be to take it project by project, and for each new project figure out what can be applied from previous projects and what is new in the current project. Then research the gaps, take good note of what you find, and try to remember it for the future.

When it comes to chips, you should reach a point where you can predict what chip should be needed to fill that hole and then go look for it. If you are lucky other people have the same hole to fill and the chip you need will be out there. If you don't know what to call the chip, or what terms to search for, then ask in a forum like this. If you can describe what you are looking for, people should be able to tell you what it's called and point you to examples.

[dannyf]

how do I gain the ability to find out about these things?

Two separate aspects of the same question:

1) how do you approach a problem? I typically try to figure out what I want to do in the end, and then build functional blocks to take them bit by bit. For example, in your case, I would figure out first if I need a display; what kind of user input I need; what functionality the mcu may need to have to interface with my display + user interface...

2) how do you utilize a specific chip in your design? read the datasheet. There really is no way out. You have to read about the chips you use and the datasheet is the best document you can ever hope to have.

As a beginner, you will understand not much in the beginning but after the 10th or the 20th reading of the datasheet, more of it will start to make sense to you.

[JackP]

  Thanks for the detailed answer,
I am currently attempting to teach myself, and I have noticed that it is difficult to learn about something if you have no need for it at the moment; as a teenager, I am not always able to complete a project due to financial problems. That issue is obviously, as you mentioned, solved by having a set curriculum to follow. One question I do have is where would you recommend to start for someone with limited knowledge, who has none of the chips. Should I start by looking at memory solutions, with application specific ICs following when a project arises, or is it better to look at the various protocols before even looking at an IC. Something I have wondered about is how you can communicate with many different chips using I2C or SPI, all of which have separate functions.

In regard to dannyf, I try to approach projects in that way, but often my inexperience results in me dismissing the original notion on the basis of it being to difficult; I hope as my knowledge grows, so does my confidence to try something new. I also really need to learn how to read datasheets as well, like you mentioned.

Thanks again,
Ben

[dannyf]

Analog chips are generally much tougher to use than digital chips.

I would suggest that you start with some digital modules, like a rtc, a spi/i2c dac, etc. Take them one at a time and just play around with it to train yourself.

There are usually lots of code pieces floating around and compare them vs. the datasheet to get a sense how others have done it. Ask for help when you do run into issues.

After a few rounds, you will become an expert.

[IanB]

If you say that you are a teenager, then still at school and not yet attending college or university, it may be that you are attempting projects that are too challenging for your level of learning. Though, unfortunately, the more "accessible" projects that would have been standard fare 30 or 40 years ago are no longer so relevant. Times have moved on. Even so, you might want to look out for less complex projects to work on. There will be more satisfaction in finishing things rather than having unfinished projects on the go.

[Galenbo]

EDIT: What about protocols? Can anyone recommend a good source of information for the protocols in common use? Months of searching have left me no better off.

I can recommend you to put 100 hrs in understanding the OSI layer model, guided by a book or serious online series.

Then, you choose 1 protocol. Can be rs485, spi, UDP, whatever.
Then, you actively spit up the whole protocol, build something, test, from voltages to memory allocations on bit level. Takes minimum another 100 hrs.

Now you are ready to learn whatever protocol very fast. You will quickly see 'through' it.

 

[JackP]

Thanks again for these great answers,
I did start with playing around with simple digital modules, but I was spending money on the modules instead of the chips themselves. I'll try to get some different digital ICs, now that I have been given some new dev tools for the AVR platform. IanB, I try and mix easy and hard projects together, so if I get bored with a long running project, I might look at transistor logic for a while, or something of the sort.

  Galenbo, thanks for this, it seems like a great way to learn protocols, getting to know everything down to the bit level. Have you come across any particular books or series that do a good job of explaining the OSI layer model? Thanks,
Jack

[mikeselectricstuff]

  Thanks for the detailed answer,
I am currently attempting to teach myself, and I have noticed that it is difficult to learn about something if you have no need for it at the moment;

It is much better to define a specific goal, e.g. to build a widget that does <x>
That way, when you hit a problem you are more likely to work through it, rather than getting bored and moving to something else.

You will never learn everything. I've been doing electronics for over 35 years, yet it's rare that I don't learn something new during a job.

[JackP]

That's what I mean, if I haven't got a goal (which I have to admit I find hard to come up with) then I will be less inclined to solve the problems in my way.
Jack

[tautech]

Start fixing stuff, anything.
As you hit each hurdle, ask questions, you are already a member of the best place for assistance.
Knowledge will grow and you will have "eureka" moments along the way.
 .

[Rigby]

On a large scale, it's just basically reading, reading, reading.  Absorb through your eyes and ears at every opportunity.

Imagine putting together a jigsaw puzzle.  In order to finish the puzzle, you have to find the corners, the edges, and sort pieces by color or pattern.  It's the same for learning something, except it's a 30,000 piece puzzle, and you can't see the puzzle (you aren't able to quantify what you don't know).  Through listening and watching and reading you learn to see the corner pieces, and the edge pieces, and you begin to see how things fit together slowly over time.  Little patches here or there become clear, and you just fill in the puzzle as you go.

Unfortunately, you'll have to focus on a region of the puzzle and not the whole thing, because a 30,000 piece puzzle is a very large puzzle, and no one person can put it together in their lifetime.  This is why specialties exist in engineering, and even within engineering disciplines.

On a small scale, just absorb any information you can find on the subject(s) you're interested in, any time you can do so.  You'll soon find out whether or not you're really interested in the subject, and if you really are, you'll start to get annoyed by your human body because it can't take information in as fast as you're ready to receive it.

[Howardlong]

EDIT: What about protocols? Can anyone recommend a good source of information for the protocols in common use? Months of searching have left me no better off.

I can recommend you to put 100 hrs in understanding the OSI layer model, guided by a book or serious online series.

Then, you choose 1 protocol. Can be rs485, spi, UDP, whatever.
Then, you actively spit up the whole protocol, build something, test, from voltages to memory allocations on bit level. Takes minimum another 100 hrs.

Now you are ready to learn whatever protocol very fast. You will quickly see 'through' it.

Below is just my opinion on the OSI model, feel free to rip me apart!

I had to do a double take on that bit about spending 100 hours learning the OSI model. The OSI model is little more than an academic tool that lecturers use so they don't have to change their syllabi and exams every year to keep up with technology.

I learned it formally at university 30 years ago and to be honest my life would have been better without it ever existing, it added no value to my understanding, and if anything it confused things later on. Being aware of it, is worthwhile, yes, but learning it parrot fashion like I had to do to pass my exams, no use whatsoever.

Once I asked the lecturer in class how the OSI model fitted in with a real network implementation on ethernet or token ring, or X.25 or anything really (remember TCP/IP wasn't anywhere near as ubiquitous as it is now, IBM's SNA ruled the roost in those days). I was met with a blank stare. He didn't know.

It's a conceptual model, that's all it is, it is not a practical implementation. It wasn't until the mid 90s when I did some real networking work with TCP/IP, and then later on in the early 2000s when the concepts of layer 3 switching became common that I could see some semblance of where the IP stack fitted into the OSI model. In the interim, I'd developed and implemented all sorts of global enterprise networking infrastructure implementations for mega corps without knowing WTF a presentation or session layer was, and my peers were the same. In fact, I still have no idea practically speaking what use it really has other than something that academics use to test their students.

[tggzzz]

How do you learn about everything? There are two simple and correct guides...

Firstly hard work over a long period. Secondly, in the words of Piet Hein:

The road to wisdom? — Well, it's plain
and simple to express:
Err
and err
and err again
but less
and less
and less.

Piet Hein's other scribbles are worth reading too.

[tautech]

other than something that academics use to test their students.

Exactly. 
Just as you have described there is nothing better than practical experience tied with/to the theory.

But which comes first? (the chicken or the egg?)

IMO in order to solve a problem in hand, one the applies the theory. (along with a bit of basic experience/knowledge  )

[G0HZU]

It is much better to define a specific goal, e.g. to build a widget that does <x>

Agreed.

Why not try making something practical like a universal (multi) sensor module that can measure and log lots of things and store the results?

But don't try and make it particularly compact or too 'integrated'

Choose a controller board (eg Arduino) and then find some external chips for temperature, pressure, current sensing, elapsed time, I2C EEPROM memory and FLASH memory and some external DACs and ADCs.

Make something that can be easily expanded with more and more peripherals and it could end up being a useful monitoring/logging workhorse on your workbench as well as a learning tool. i.e. something that can log the sensor data to its own FLASH in standalone/portable mode or something that could log to a PC via USB.

[Howardlong]

I've always found that the best way to learn for me is by doing. Sure read up on stuff, but reading is not doing.

Some people, who I am sure are much smarter than me, can deal with things on an abstract level far easier than I seem to be able to.

Like Mike, it's rare I go through a day without picking up a nugget here and there, although at my age it's probably a squishy balloon effect, one bit of info goes in, out pop a couple of others.

I do think it was easier twenty five years ago or so. At around that time, an individual could realistically have enough knowledge to try their hand at anything in electronics and computers given the right tools and experience.

Since then, the complexity has increased at such a rate that you really do have to pick and choose and specialise to some degree.

[xrunner]

  So I am asking you, how do I gain the ability to find out about these things? I should know about the memory aspect, but what about the ICs used? Is it even necessary to use the separate chip, can I control a peripheral directly? Any advice would be greatly appreciated.
Regards,

Get some working electronic junk, and take it apart. Take apart as much as you can, and then rebuild it. Hopefully (this may sound weird) it doesn't work when you put it back together again. Go back in and see if you can fix it. At first you will be stumped but with some basic skills you will build on your knowledge over time.

[IanB]

So I am asking you, how do I gain the ability to find out about these things?

Age, and experience, and practice. Don't expect it to come quickly, and don't expect it to come easily. There have been times in my education where my head was spinning. It happened a lot when I was a teenager, it happens a lot less now.

Also, "gaining the ability to find out about things" is absolutely the right approach. Just keep at it. I think it can be said that teaching doesn't happen, only learning happens. A teacher cannot put knowledge inside your head, you have to put it there yourself. So keep trying to learn and eventually you will get better at it.

[nctnico]

I bought books about electronics and read those. Back then there was no internet though. Still a book is better than wading through random internet sites because a book provides information along a clear path.

[Stupid Beard]

The main problem I always have whenever I want to plug gaps in my knowledge, or just want to make something, is what the bloody hell to make. Coming up with project ideas really sucks. Here's what works for me:

Find a problem you actually have in your day to day life that you can solve with electronics. There's your project idea. You can pick random things off the net that may interest you, but I find it helps to have an actual use for the thing you're making. That means you care more about the project and you'll stick at it until it's finished. For the random useless ideas, I find that they rarely get past the proof of concept stage. They're good for a weekend of messing about, and you'll still learn things from them, but they don't have longevity.

Another option is to pick a random interesting part in your parts/junk box and research it thoroughly, then build something around that. This also suffers from the longevity issue, but can lead you to unexpected and interesting places.

As an example, when I started running I found myself wanting to track things like time and distance. Carrying a phone or GPS while running sucks so much that I leave them behind, so that left a watch. I hated all the ones I could find online, so I decided to make my own that did only what I wanted it to and sod everyone else.

Still working on that, but in the process I found a pile of piezo transducers in my parts box that I bought over 15 years ago and never did anything with. This led to me hunched over a soldering iron on christmas eve quickly hacking up some firmware to play jingle bells, hooked it up to an LDR and a comparator to trigger it when it's light and shut it up when it's dark. I then buried it under the presents under the christmas tree.

I went out for a run on christmas day and came home to confused parents who couldn't figure out why jingle bells was playing or what was doing it after they'd opened some presents. I called it the Annoyatron because it was really annoying, and continued to annoy me every time I opened the desk drawer for the next week until the CR2032 of questionable vintage finally died.

Come to think of it, pranks are probably a good source of learning material in general

[JackP]

  I agree with you - I never know what to make either! I'll try what you said (find a problem etc.) and see if I can come up with a great idea. I also like to learn by reading books, but I find it easier to follow up and continue to build up my knowledge by making something physical. I think one of the major problems in my way is that I have limited access to new components, so when I think of a design from scratch, I end up quitting because I know it will be a while before I can get it, and that's if it is available on eBay or Amazon. Thanks again for your continued help,
Jack

[Rick Law]

JackP,

Self learning is a good thing but difficult to learn without a structure.  As the saying goes, you don't know what you don't know.  So there may be some knowledge others may consider very basic whereas a very advance self-taught person may not know - simply because he hasn't came across it thus never learned it.  Worst yet, he would not know he has a hole to fill.  It might have been the one knowledge that could have taken him to the next level, but he didn't know he was missing that.

To avoid that pitfall, make use of some free resources - look up curriculum/syllabus of courses taught in colleges/vocational school (for EE).  By seeing the skill/subject they would have taught, you can ensure you as least know of (some of) the stuff you do not yet know, some of the stuff that may be valuable to know, and then you can decide if they are important.

Rick

[rx8pilot]

I am self-taught, beginning in the early 80's.

Fast forward, I have learned about learning along the way. At least for me.

1. Have a mission. Others have said this and it is very true. I read a LOT about components, RC and LC networks, transistors, etc, etc.... None of it made sense until I had a mission - a problem to solve. It could start with a blinking LED (discreet components), dimmer, motor controller, anything simple. I always started at the end "What does this thing need to do". Keep the problem definition in front of you and start guessing your way through it. Every mistake you make is learning, not failing. I know a million things because I made a million mistakes. The logical thought process you develop and use to solve the problem of getting an LED to blink with a handful of discreet components is the same process used when you design something very big. In general, all electronic systems have a lot of repeating themes. Those themes are like paint for an artist. If you only know how to make a battery light up an LED, you have one color to paint with. If you can make it blink, you now have two colors. Learn a million more simple lessons and you can paint anything you want.

2. Discipline. I struggled for a ling time with spotted knowledge because I only learned the minimum needed for my projects. I did not always know why a circuit worked, but I was satisfied enough that it did. This is what classes or structured learning will help with. I started to figure out that I was massively more effective when I started learning the basics after I thought I was "past that stage". I am still plugging holes in my knowledge to get through projects.

[SL4P]

The absolute best solution is to find a more experienced peer that can sit patiently alongside you every now and then (once a week or two) - to answer specific questions, and help you prototype some building blocks.

They can look at your level of understanding - and refocus your attention if you try to jump from AB to E without having a very basic understanding of C&D.  You can jump backward, but not knowing at all can cause untold hours of frustration and cooked parts!

We've all done it... a day of head banging over a missed spec, or using the incorrect pinout!