The Amp Hour Electronics Podcast

A weekly show about the trends in the electronic industry.

You are here: Home / Radio Show / #244 – The Art Of Staying Interested In Electronics – Exponible Electronics Ennui

#244 – The Art Of Staying Interested In Electronics – Exponible Electronics Ennui

Play

Podcast: Play in new window | Download

Subscribe: Apple Podcasts | RSS


7155210629_130bcb4115_o

  • Dave has been having troubles with his SD card not recording video footage. Chris has had the same problem with KiCad crashing over and over again (on copper pours).
  • Chris has a new video setup that he claims beats out Dave’s setup. It uses Xsplit and changing views and pausing makes things a lot easier.
  • Chris has been feeling down about “what’s left” in the electronics industry as things continue to change.
    • This was prompted by Chris playing his new piano (Korg SV-1) which pulls in tons of different features.
    • The Teensy 3.1 shows how much can be packed into a simple platform (and how little else is needed!)
    • Reading The Art of Electronics version 3 seems like a good start but doesn’t feel like it will change much (it likely will).
  • yCombinator demo day startups continue to be silly. Day 1 was worse than Day 2.
  • Dave has been starting a product design with an outside design firm. He won’t say what it is, but we know he’ll be able scratch his own itch with something that has his face on it…
  • Sam asked about when to pull the trigger on moving to the next stage of a design.

https://twitter.com/samrustan/status/585554711060811777

  • Finally…a good use of a 3D printer: Making a pogo pin holder. However, you could still order this from a provider online (just like PCBs), the tradeoff is time.

Thanks to TRF_Mr_Hyde for the snarky picture!

Comments

  1. Chris,

    One of the better episodes this week.

    I was in a lull hobby wise last year and then started to really get into sensors. Especially low power ones that push data to the cloud. I know you and Dave scoff at IoT but I take a much more practical approach to it than all of the hype in the media. I’m having fun now building the follow up version of imp.guru called analog.io. Node-red is a fun toy too.

    So far I’ve done some pretty fun stuff like monitoring the Temp & Humidity of Bee Hives in winter. Monitoring Fuel inventory at a local Gas Station. Now I’m building out a weight scale that will continuously monitor the weight of a Bee Hive as well as small scale Solar power system that monitors MPP, Power Generation, Load Power, things like that. It’s a lot of fun and an area where there is pretty much unlimited potential.

    I’m also VERY excited about ESP8266, with NodeMCU it is an amazingly easy and super cheap way to do WiFi. Its like how when USB was hard, FTDI made it easy. Now WiFi is hard and ESP8266 makes it easy. Electric imp is cool too, don’t get me wrong but I can deploy ESP8266 nodes like 8:1 because of cost.

    As far as micros, you really cannot go wrong with any choice. With that said and for whatever reason, my goto is always MSP430 because it is familiar, I have a ton of launchpads laying around and the docs are easy to read. Also TI has their new cloud version of CCS that makes the whole tool chain setup stupid easy. You can do both Arduino like and C development in the same environment.

    If that becomes too limited (which it likely will), I think that there will be a lot of buzz around MSP432 this year and CC1310 looks very kick ass. For non-TI stuff, the new Atmel SAML21 chip looks pretty cool. It even has a surprising number of analog peripherals. ST stuff is great as always and Freescale is the cheap option. Like Dave said, best to just pick one and never look back otherwise it is a sure way to get into paralysis by analysis and you’ll never get out of your funk.

    -Luke

  2. Sd cards, you have to love them when they decide to do wear levelling and corrupt the FAT. format in camera, use for a short while and then get a new one. I have had quite a few that went flaky, and find that you are best just replacing with new ones. Faster, cheaper and more capacity for the same price.

  4. Chris: You need to challenge yourself. Find something that you know nothing about and delve into it to learn it. Then use it in a product.
    Or, work with someone on a project. Another person will pick up your spirits when you are down. Then you will have to pick up his spirits when he is down.

  7. Anyone doing large projects quickly outgrows the Arduino IDE.

    Best development environment in its place is the Atmel Studio 6 IDE(it’s free).
    It is a derivative of Microsoft Visual Studio IDE and uses GCC at it’s lowest level.

    The Studio 6 supports the AtMega(8-bit) family, ATXMega(8-bit) family, ATTiny(8bit) family, as well as all the various Atmel ARM-Based 32-bit families.

    • Can you still use all the built-in Arduino sketches/libraries with Atmel Studio? I’ve been meaning to port some of the functions over (like Print) to be platform independent.

    • I’ll play Sheldon here…ATMEL Studio 6 is built on Eclipse (awesome product).

      I used Microsoft Visual Studio for embedded systems daily, then I tried Eclipse and immediately deleted VS (true).

  8. Chris, I agree, I don’t like online compilers and project management systems like the mbed website either. But mbed is really cool. I’ve never seen C++ used in such an easy to enhance and easy to use way before for building micrcontontroller apps, and the online system makes it easy to use, too. Fortunately you can donwload the build environment from github and use it with the free ARM GCC compiler offline, as I described here: https://groups.google.com/d/msg/fpgalink-users/rwHlnmUgtZY/fTnftsiJEvEJ

  9. Chris, for something completely different try the PSoC 4 or 5 ARM microcontrollers from Cypress, the tools are good and free, and you can do firmware(software), virtual analog and FPGA, all on one chip, they are really nice. You can go and program the registers but it comes with API’s full documented and come with GUI configuration for the micro peripherals

    This protoboard will be released soon for 10 dolares.
    http://www.cypress.com/?rid=108038

    Note: I don’t have any interest part of the company, I just like it very much.

    Best regards
    Joao

    • I just picked up the PSoC 4 the past few months (bought the BLE Explorer Kit), and it’s a very cool architecture. You can even program custom VHDL/Verilog as Joao mentioned (see http://www.cypress.com/?rID=69773), and Cypress has told me they make available the sources of some of their prebuilt blocks to you (under NDA however so it’s not posted), but if you need to slightly tweak something it’s possible! The analog programmable blocks are all interesting as well. It’s well worth picking up a dev kit and going through some of the tutorials to learn about it, even if you don’t immediately plan on using it. The architecture is unique enough that it’s worthwhile having some familiarity with, as there are some projects that are a shoe-in.

      My more regular go-to is Atmel stuff, was using the SAM3U most recently for USB. But to me the Atmel/Microchip/NXP group is all fairly standard – more along the lines of “pick what features you need, then go figure out which vendor supplies all of them in an appropriate package/cost-point”. Although I do like the Atmel IDE / software framework so lean towards it.

      • Hi, Chris. I’ll reiterate what’s already said here: the PSoC 5LP is a very interesting chip. Programmable analog & digital subsystems married to a Cortex M3. Complete flexibility on mapping the peripheral I/O to any pin on the package. Can do USB, even without an external crystal. Good IDE. Good documentation. Nearly the same price as the ARM used on the Teensy, but you’ve traded the M4 with FPU for the M3 with the additional configurability.

  10. Earlier this week, I was listening to the Software Engineering podcast (IEEE Computer society sponsored) and they were talking about “Technical Debt”. What the hell is that? Well, that’s the crufty-ness left in your software because; you’re not very good at this, you don’t have the time to do it properly, you made a stupid decision in the design, or some other factor where you just have to ship it even though you’re not happy with it.

    Well, while listening about technical debt, one thing kept popping into my head; Microchip PIC processors. To save $0.50 on a part you sell your company and career into a heap of bad decisions and technical debt.

    To get a processor with the features that you might need, Microchip stuffs 20 kg of crap into a 5 kg bag. Since their architecture only has so much room for device registers, they will flip a bit in one register to enable a bunch of other registers. You’re dealing with pages and executing code across page boundaries. It doesn’t have a real stack, so you have to worry about how many subroutines you call and what parameters get used. And it just keeps going and going.

    Developing with their “free” tools is a frustrating hell hole until you figure out that you are, not nearly, alone. Everybody on their support forums are having the same problems. The IDE is developed in India and is second to any that you have ever used. Bugs are plenty and get fixed eventually with a cascade of new ones and everything slows down. An edit/compile/link/burn/run cycle takes about 10 minutes for a 64K program, you can never remember what you were trying to fix by the time it is ready for you. The compiler is written in Australia and is pretty good, but it has to be integrated into the damn IDE, kill me now.

    Great, you get your code going and, now that you figure out what features you actually need for your project, you can pick the proper processor – BUT – you have to rewrite your code because the first processor has version 4 timers and the new one has version 6 timers and they don’t work the same at all. (as Dave would say waa-waa-waaaaaa).

    Wash, rinse, repeat. Eventually you get fed up and get an ARM processor from ST or ATMEL and everything is, comparatively, unicorns and rainbows.

    The architecture of the ARM processors is very similar to the Motorola 68000, a very clean design that doesn’t carry along a bunch of bad decisions (mumble Intel mumble). The processor section of the ARMs are basically all the same, the differentiator is the peripherals. One Cortex M4 is the same as the others, so if your version of GCC can generate code for an M4, it will run on an ATMEL, TI, ST, Freescale, take your pick. The library code from ATMEL, TI, ST, Freescale, will help you access the peripherals that they designed.

    ST has a weird SPI implementation where you have to bit bang your own chip select, but their ADC section with DMA and timers is really cool. ATMEL has some pretty skookum serial ports. TI and Freescale are a pain in the ass and won’t talk to you unless you are the size of Apple.

    We’re using ST’s STM32F407’s. You can get a dev board for cheap from digimouse and it has a debugger module built on. There are demo/code restricted versions of the compilers and their chip setup/code generator app saves weeks (really). And if you go listen to the Embedded.fm podcast talking about RTEMS, yeah, I’ve got that running on a ‘407 with GCC, and it costs nothing.

    Actually, if you’ve got a hankering to write code at the low end, you should probably listen to the embedded.fm podcast. They know as much about embedded systems programming as Chris and Dave know about electronics, and by corollary they know as much about electronics as Chris and Dave know about code.

    • Your comments regarding Microchips products are presumably referring to PIC16/PIC12 and MPLAB v8 era products. If so, and out of respect for those concerned, I politely request that you update your experience (PIC24/32 MPLABX) or file that opinion away under “obsolete”.

      • Oh, I’m sorry, I was referring to the PIC18 range, K series, J’s, the inexpensive stuff. You must be right. Isn’t the PIC32 based on a MIPS R4000 design from 1991?

        Enough pissing match, I had to deal with those dogs for the last 10 years and I dread the day when I get stuck on the next project and have to do it again. Nope. And I also don’t need their sales guys coming around offering to pay us to use their chips in our designs. Pass.

  11. Hi Chris,

    I had to choose a new micro-controller just over a year back, supporting a new device is
    a huge cost for a small company, so when I choose a new one, I try to choose one that
    I’ll be able to use for a broad number of applications.

    In the past, I’ve had a small 8 bit, larger 16 bit and dsp, which covered most my needs,
    I keep good quantities of each, so I don’t get caught with supply issues.

    About a year back (closer to two), I started a new design and all the micro’s I’d been using
    were too old to commit to a new design. I’d been putting off making the choice until the
    hardware was getting close to finished, as I needed it to do the electronics. The product
    is a machine used in the mining industry.

    I’ve been a Motorola customer since the 70’s, so my choice would always be biased toward
    Freescale.

    Not long before the decision had to be made, Freescale announced a ARM M0 (KL25Z) device,
    which at around $3 a device, the cost was well below the 8 bitter I’d used for the last 15 years old.
    The dev/programing board was $10. It used Codewarrior tool set, which I was used to. The power
    required was supper low, at least for my needs. A nice simple 1 wire interface for programing and
    debugging, which I also added one of the SCI ports to, just as the dev board does. Which is sent
    back via the same USB cable as the programing interface. leaving me with two SCI ports for my app.

    Other then some small issues with the Codewarrior generated code, it’ working like a dream.

    The actual part number is: MKL25Z128VLK4

    Cheers,
    Darren

  12. I’m a first year engineering student here (but had a few years of doing it seriously as a hobby). I’ve also been on a bit of a downer recently – we’re sitting through super-boring maths, physics and CS courses, and I have a nagging feeling that I’m going to get precious little time to spend on the actual fun stuff! Also, I just got an ESP8266, and I’m wondering where my skills will be used if every man and their dog can do IOT stuff by spending $3 on a chip, wacking a pair of AA batteries in there and copy-pasting some code in lua to run on the nice and shiny 80MHz processor! I feel like it’s getting too easy 😛
    However, I have learnt one really cool trick for staying interested. By diving into a new field (botany, neuroscience, oceanography, whatever) you’re suddenly presented with a whole new set of problems, and as engineers we have a whole bunch of tools other people don’t have, so we get to do great work and learn new things with a whole different set of constraints and some interesting new challenges to have a go at.

    • This is great advice. It might be that I’m focusing too much on the field (electronics) and not the areas of application where the field can be used (botany, neuro, etc).

  13. Chris,

    My first micro was a teensy, which I used with the standard teensyduino IDE. After this I got a Freescale kl25z (frdm dev board, Cortex M0), so the first micro I properly worked with. I used Eclipse with CDT and the GNU ARM plugin (http://gnuarmeclipse.livius.net/).

    Comming from a software background I really value having the nice IDE and compiler combination. I later also programmed an STM-32 series chip with this setup.

    I think going with a 32 bit ARM micro is a good choice for contextual electronics, since the knowledge of programming/debugging for ARM targets is a bit more widely applicable then PIC/Atmel specifics. I also do not really see a compelling reason to use 8 bit micros, for something where power draw and price are not hugely limited.

    I really enjoyed the lower level things I learned in moving away from arduino, like general insights into how peripirals work and are controlled and I especially enjoyed using DMA, which was an interesting way to merge hardware into my software. I think moving from the arduino ide to something like this is a manageable and interesting step for people to take.

  14. Hey Chris. Sorry to hear you are in a rut. I get like that sometimes. I find it more helpful to do something else for a while. Cut the grass, do some house work. Eventually I’ll find something that can be improved in some way, or something I want to monitor and go from there.

    I don’t know if excited is the right word, but I am becoming more and more interested in this new intel edison board. I like the idea that its just a very small pentium computer. This means I can code it up any way I would have coded on a computer. I am so over arduino. I feel like the arduino environment spoon feeds everyone the same crap. Sure some people have done some neat things, but ugh.

    As for Dave’s SD card issue, I am not sure if they have these over there, but they make these SD cards with built in wifi. Here’s the website: http://www.eyefi.com/ It would save him the trouble of having to take the card out of the camera. It would change his work flow a bit, and in doing so, I think it would mitigate the possibility of using windows accidentally to modify the contents on the card. I’m not sure of the wifi speed of this thing. If its too slow, maybe someone can put together a project where you take an SD card, and transmit its contents automatically to a network storage location.

    Finally, I just wanted to say, I really like the show. I have been looking for something more geeky. I used to do this kind of thing a lot, but after I graduated from school…11 years ago now, I kind of fell out of it. This show has definitely got me interested in getting back into it.

  15. Hey guys,
    This is my answer to the “why I’m interested in electronics” question from the podcast.

    I’m an experienced software engineer with almost zero knowledge in electronics.
    When I was a kid (more than 25 years ago) I was super exited about computers because you were able to create whatever you wanted… the imagination was the limit.

    And that same feeling is happening to me again, but with electronics. Components are affordable, you have user friendly IDEs to program micro-controllers, affordable embedded computers, affordable 3d printers, etc. Basically you can create whatever you want… the imagination is the limit (and the budget… since electronics is not as cheap as software, but it is still affordable).

    Do you want to create a robot for your kid? You can do it all by yourself.
    Do you want to create home-automation devices? You can do it all by yourself.
    Are you an artist and want to experiment with lights, networking, moving parts?… you can do it all by yourself.

    Ah… the my other motivation to learn electronics is that I like playing with old computers, computers from the 80s. And I must learn how to repair them because nobody repairs them anymore.

    Well, that’s why I’m interested in electronics! 🙂

  16. It’s funny, I was just noticing that I hadn’t watch the new product posts from Sparkfun or Adafruit in almost 6 months.. So I watched the. Boooooooring. I’m with you. We’ve hit a plateau in innovation. This podcast was perfectly timed to make me feel like I’m in good company.