Learning, learning, learning…
I started to implement my another dream: to become an electronics wizard.
Well, maybe not exactly a wizard, but something like that.
Obviously task is quite complex, but I belive people are living just for the dreams implementations, which is cool, so I’m not afraid and do not hurry very much. Reading first couple of chapters of “The art of electronics” I found, that my old knowledge is somewhat between void and zero, but it does not look very complex so far. Unfortunately there are no exercises solutions in the book, so I only need to guess that I’m right or wrong when solving them, and overall it is very interesting. Also electronic book (about electronics) is not very convenient to use, and there are no russian translation in the shops, so I use djvu.
Did not seriously practice yet, will likely start to implement more interesting things when come close to IC.
Btw, a question: do I need to buy an oscillograph now and which one: quite expensive 100 (or more) MHz of input pass band or just 20-40 is enough, with or without logical analyzer, and so on. I plan to develop mostly IC schemes (and likely in the future using high frequencies too), and do understand that it is definitely needed, although right now it is a bit early, so just would like to gather opinions.
And also, is there an open source Linux software for IC design and emulation?
So, things currently in development: learning linux kernel (and of course others too) hacking, learning (basics of the) electronics, learning math (group theory, so far basics and crypto applications), learning music (likely you do not want to hear my trumpet playing :), learning religion (I read the Koran (very strange book, but I can understand a little bit more about people now) and have plans about Zen just to extend horizons, not to deeply addict on that), (not as frequent as I want) learning sports, (rarely) learning appartment development things.
Learning how to make things interesting for myself :)
POHMELFS and local filesystem. Libionotify. New distributed storage release.
Comments are currently closed.
A friend and I work on http://www.digispeaker.com
We just went through the test equipment decision process. There are three things you need to look at: one shot events, complex digital logic and analog. For one shot events you want a digital storage scope. These scopes digitize and record the signal to RAM. After much research we just bought a Rigol DS1102E for $730 (after a lot of shopping). This is a brand new scope out next week. Rule of thumb 1GSa/sec = 100Mhz, 5GSa/sec = 500Mhz, 10x samples to effective bandwidth. The Openmoku project uses Rigol scopes.
For complex digital logic you need a logic analyzer. Research says the best bang for the buck is http://www.pctestinstruments.com/ 34 channels, 500Mhz USB. But you need to use Windows for the UI. You use these for debugging an SDRAM interface or the PCI bus. We have not bought one of these yet.
For working on power supplies or audio you need a low frequency analog scope, 20Mhz is fine. We’ve bought a couple from Ebay for $100. They work six months to a year and then die. Might be better to spend $300 and buy one new. The 400Mhz+ analog scopes are for working on radio and radar.
So if you are going to build digital stuff start with the Rigol. For power supplies and audio start with an analog from Ebay. Logic analyzer is last piece of equipment you need. When you tire of this equipment you can usually get 50% of your money back on Ebay.
Thanks a lot!
I thought exactly about 100 Mhz RIGOL or 150 Mhz Ez, and incline to RIGOL. Also found theirs models with logic analyzer (although simpler than what is present at http://www.pctestinstruments.com), but when I long ago worked with people who made even quite complex IC (ppc cores and set of DSPs for sound and video processing), they never used LA, so I believe it is for really serious projects and not needed at my current level.
I love electronics, but, i don’t know anything about electronics yet :]
ok ok…. I love to know how things work.
I took a look at amazon and found the book “The art of electronics”, but,
it isn’t a book for beginners, is it ? :(
Some reviewers said that the book is for who knows at lest the basics of electronics,
so, what book would you recommend for totally beginners ?
It is the book for the beginners. At least for me it works excellent: very good language (I read russian translation), quite simple exercises aimed exactly on currently discussed issues, very broad range of coevered topics. And at the same time it shows very serious things for the professionals.
Start with it, and if first chapter(s) are very complex for you, you need to learn more basics, but I’m pretty sure it is what is needed.
По опыту работы могу сказать про осцилографы. Брать надо цифровой, минимум 2 канала (а лучше 4), с памятью и всякими фишками по анализу сигнала. Частота сильно зависит от класса устройств. 100 МГц думаю многовато, это уже для микропроцессорной техники, такое дома на кухне не спаяешь.
Да, и еще. Ни в коем случае не бери софтовый! Это который втыкается в PCI или USB. Железка должна быть автономной.
I don’t use any such software myself, but you might want to take a look at:
http://kicad.sourceforge.net/ and http://bwrc.eecs.berkeley.edu/Classes/IcBook/SPICE/
There is also an old article in linuxjournal: http://www.linuxjournal.com/article/1317
/andrnils
Thanks a lot!
4-х канальный совсем неподъемно, так что только 2.
А компьютерные платы да, это не дело…
The Rigol screen is only 320×240. That’s ok for the scope. When you try to add a Logic Analyzer at that resolution everything gets too hard to read. LA at PC Test Instruments is better than one built into Rigol and half the price. Plus it uses you PC’s screen. Software works under Wine but they need the FTDI library and that hasn’t been ported to Wine.
Digital circuits do a lot of non-repetitive things. An analog scope needs a repetitive signal to be able to display. For example you probably won’t be able to see something as simple as a RS-232 line on an analog scope. Looking at RS-232 is easy on a storage scope.
Note the previous Rigol 100Mhz model was 400MSa/s, they have changed this in the new model to 1GSa/s. It is worth it to get the newer model.
Is it ds1102e model? Looks like it does not have sequence mode, is it needed at first, if I understood correctly, it just fills 1000 frames into its 1m of ram and allows to view it in ‘offline’, and likely ds1102e will just fill the whole memory with its very frequent probes, while ds1102c (4oo sa/s) has a 1000 frame sequence. Or ds1000a series, which has only 10k of memory for two channels, compared to 1m in ds1000 series.
We got the DS1102E. It is not a perfect scope but it is a very good trade off for the price paid. Good DSOs are $8,000-20,000.
Rigol scopes have two types of memory, fast and slow. There is 8KB of fast and 1MB of slow. At full 1GSa/s you can capture 8K points on one channel. For two channels you get 500GSa/s and 4K points each. This is because it needs to be able to write to RAM at 1Ghz. You can’t write to normal DRAM that fast. If you slow down what you are looking at it can use the 1MB of DRAM to hold samples. 4-8K points is enough to see what is happening.
DS1102C is identical to DS1102E except for changing from 400MSa/sec to 1GSa/sec. The difference is that when looking at a 100Mhz wave form you get 10 samples per cycle vs 4 samples. The more samples, the more detail you can see. 400MSa/sec was probably too low for a 100Mhz scope.
All of the competing Agilent/HP/Tek 100Mhz scopes are 1GSa/sec but they are $1200-1500. They have between 5KB and 16Kb of fast memory. The LeCroy can do 1GSa/s on both channels at once but it is $1,600. You could buy two DS1102E for less than price of LeCroy.
A decent rule of thumb is to divide the sample rate by ten to get the working bandwidth.
The comment about wanting other things first is true. If you are working with surface mount you want a temperature controlled iron, rework station, big magnifying glass, etc first.
that’s what I’m working on (without glass though): my old soldering iron is not very convenient, but I used to solder even planar devices without glass. I liked to work with Xytronic 137ESD.
It is way easier to buy microcontroller development boards than it is to make them.
This site is very cost effective: http://www.olimex.com
I agree that is is much easier to buy one than to create a board from scratch, but I want to do it because it is interesting for me, not because I have to easily push out some final product.
What you hook up to the development board is usually more fun than building yet another microcontroller board. We used an off the shelf mpc5200 board from Phytec and built all of our peripherals first. That way we had a stable CPU environment to write software in. Now that the peripheral designs are fair stabilized we are building the CPU board. Our current peripheral designs look nothing like the initial versions. Our CPU design is basically a copy of all the other mpc5200 designs in the world, there’s not a lot of flexibility in the CPU design.
The CPU part of the design is the most boring part. The complex parts in our design are the power supply, power line networking and audio amp.
You might find gnuradio to be more interesting.
http://gnuradio.org/
Software defined radios need both hardware and software design skills.
I tend to agree, but you know how to build such systems, and I do not. So it is interesting for me to implement this IC, and then extend it with different boards. Or just buy development board with all accessible interfaces and connect external systems. And obviously I will not start implementing it right now basically because my knowledge is… well, its level is just miserable, and essentially for sure I will first play with development board.
I’m only starting to learn electronics, I do understand that jumping into complex parts without appropriate background is not the way to go.
Maybe, who knows, my hardware background is rather miserable at this point :)
Check out http://www.diyaudio.com/ and try building an audio amp from discrete components. That would be much easier and a more interesting beginning project than jumping into a CPU design. CPUs run at high frequencies which makes them much harder to debug. You can build amps on a breadboard and you need PCBs for CPUs.
That is exactly what I calculate reading “the art of electronics” right now, although my amplifiers are rather simple and only based on transistors so far :)
I’m about to start learning Ebers-Moll transistor model before moving to the operational amplifiers.
Making PCBs (for CPUs in particular) is another question yet to think about.
Ага, как разведешь четырехслойку под BGA – позови позырить :)
by IC I mean digital circuits mostly. Likely analog schemes are now well hidden inside digitally accessed ‘black boxes’.
It’s possible to do a lot with a little in electronics. Often, you don’t need a full blown oscope, you can flash an LED or sample a logic line with a microcontroller and spew it over serial to a host computer. Or use a soundcard’s ADC. Obviously these tricks only work with low frequencies, but for most microcontroller projects you can get by.
I used lots of scopes and logic analyzers in electrical engineering classes at university. When stocking my home lab recently I spent the first $200 on a nice lab bench power supply and a Weller temp controlled soldering iron. I still don’t have a scope…
But for example how to check a filter or ‘quality’ of the signal? Or check any somewhat frequent signal?
Creating microcontroller circuits is actually what I want to know how to develop, so the only method I have is a LED or voltmeter in the circuit… That’s actually how I developed my first circuits (which did not work though :)
Пацаны на работе сказали брать ригол.
Ога, спасиб. Заказал игрушку…