Member Since: October 17, 2012

Country: United States

  • Nice smart mirror project. I probably would have gone with an OLED display because the light-leaks from the LCD backlight make it easier to see the electronics. But the board edges could also be hidden with some strategically placed RGB LED strips or some other creative solution.

    You’re filling out the Fellowship room nicely.

  • Regarding trying to buy in large quantities not really working well for a company the size of SparkFun (or Adafruit, Pololu, Seeed, etc) is not only as you say “… the amount of items in our “large quantity” and a consumer good large quantity are orders of magnitude in difference.” but also the amount of items in your “large quantity” and a consumer good “SMALL quantity” are also orders of magnitude in difference.

    Hobby engineering (I first said electronics, but robotics has quite a bit of mechanical engineering involved…) is still a niche market. Thank you for continuing to service this little niche market.

  • Ok. That sounds fair enough.

  • Do you have any links regarding tinkersat? I keep coming up empty on Google…

  • Good question. Can it work as a BLE host, or only as a BLE device?

  • Yes, there is always a trade-off between code complexity and external support components.

    The nice thing about the 75HC595 is the latch pin is rising-edge triggered. I have a project that uses one on SPI and I have CS on the latch pin. Granted the 595 is constantly “listening” to the MOSI line, it only pushes it’s buffer to the output pins when the latch pin goes high. Because CS is actually a low-active signal, deasserting the CS provides the positive edge to push the last 8 bits that the 595 saw on the MOSI line. For the 75HC165 one has to invert the CS signal to function as expected on the SH/LD pin. Throw in a small transistor and a couple resistors and voila the signal is inverted. (If you don’t have a spare gate in a hex inverter available from other circuitry…)

    Yes, I’ve recently been thinking about these two chips used in this fashion. My own yak-shaving for a project for my local maker group… (Trying to read 13 inputs, and control 117 RGB LEDs on an inexpensive (and spare from my parts bin) Pro Mini…)

  • I’m thinking that with clever use of shift buffers (both serial to parallel and parallel to serial) one could reduce the pin count on the uP down to the SPI bus and 3 chip select pins (total of 6 pins). This would allow scanning the buttons every time an LED color is written to provide 3x sampling of the buttons for debouncing for every LED refresh.

    Would SPI on an UNO be fast enough to avoid flicker? Would there be enough time to do anything else useful between LED refreshes and still have a responsive keypad?

  • Not picky, accurate, and with technical things those little technical details are very important. I think she should have done another take to get the corrections in.

    To the script-writer(s): even for revised products, a quick description of what the product does in the video would be good for customers who weren’t familiar with the original product which is what you did. But, a quick description of what the revisions are would be good for customers who are familiar with the original product (or for us who are just curious).

    Just trying to make a good thing better.

  • I just looked at the example code for your thermocouple library and it’s interesting that you provide for using I/O pins to power the sensor. I don’t think I’ve seen that in other libraries. Neet.

    As far as specifying the CS pin per instance, AdaFruit’s provides the same functionality. The second example that I mentioned has the benefit of only needing 2 CS pins (and one thermocouple object in uP memory) for 8 thermocouples. Without that circuit one would need 8 CS pins (plus the MISO and SCK for 10 pins) and 8 unique objects gobbling up memory for 8 thermocouples. Eventually I plan on adding more silkscreen notes to make hooking up easier. (I discovered on my prototype that I didn’t have enough on-board documentation to make hook-ups easy…)

    Is it my imagination, or is this thread starting to wander away from topic? ;-)

    Edit: Head-smack. This thermocouple amplifier doesn’t use MOSI, so I removed it from my pin count, above.

  • I’ve played with AdaFruit’s thermocouple digitizer breakout and published my sketches on GitHub. They use the same chip that you use. While their product has been on the market longer, yours has the benefit of being designed to use a standard mini thermocouple connector.

    There might need to be some modification of code to translate from AdaFruit’s libraries to yours, but I offer them as a jumping block.

    This one is just a single thermocouple, 2x40 LCD display, chronodot, and a sd-card reader (I think I used a shield that has both xBee and sd-card sockets).

    This one is quite a bit more advanced. Using a shift register I multiplex the CS signal of the SPI bus. This gives me up to 8 SPI devices with only two CS pins. Using a 2 to 1 multiplexer chip I provide for daisy chaining for having as many SPI devices as signal integrity will allow, still with only 2 CS pins. The demo sketch reads from 4 thermocouples and outputs to 4 OLED displays. (Again, sorry it was all designed with AdaFruit tech, but the underlying structure should be portable.) This would allow multiple temperature zone monitoring.