Minerva

Member Since: January 6, 2010

Country: United States

  • To be honest, I think this is a pointless product, I can’t see how it offers customers a valuable product.

    You can just go to Digi-Key, or anywhere, and buy your SN74HC165 in a through-hole 16-pin DIP. (They cost 52 cents in single quantity, at Digi-Key.)

    Sure, if you’ve got an IC that is only available as a tiny surface-mount QFN or similar, selling a breakout board product is useful and sensible. But for a basic chip such as a shift register that IS easily available in DIP I can’t see the point of taking the SOIC version of the chip and selling it on a breakout PCB to give you the same footprint as a DIP chip that you can use on a breadboard just like any DIP chip.

    You’re spending the time and effort laying out these boards, getting the PCBs made, getting the headers and SOIC chips and manufacturing them… for no reason, really.

    Yes, Sparkfun doesn’t carry the DIP chips, but Sparkfun could easily do so if you chose to.

    Well, I guess grouping the serial pins and parallel output pins in a grouped way to make it easy to daisy-chain multiple boards by just plugging together modules could be an advantage in some cases, but not every DIP chip in the world needs a Sparkfun-style breakout board with rearranged pin groupings plus markup.

    And the two vertical rows plus horizontal row of pins would seem to make this difficult or impossible to use in a breadboard if you want all pins connected, without shorting out one of the rows of pins. for that application, best to use an ordinary DIP chip.

  • You can get the schematics here. http://repositorio.ipl.pt/bitstream/10400.21/3293/1/Disserta%C3%A7%C3%A3o.pdf

    The design is a little crude, and the results will probably be a lot noiser than they could be. This sort of biomedical instrumentation requires good analog design and layout, appropriate use of filters and appropriate power supply and ground design in order to get good results amplifying and acquiring these small, delicate signals without drowning them in noise. Significant improvements can probably be obtained with nearly negligible increases in existing BOM cost. Here are some idea and suggestions to think about.

    • All analog and digital power supplies and grounds should be separated. The designers have sort of recognised this a little bit, but it could be better.

    • There is not one single decoupling capacitor on the analog power supply in the EDA, ECG and EMG analog modules!!

    • Good filtering and decoupling on the analog Vcc rails to the analog stages should be used, with the separate analog power supply derived from the main power supply rail with LC filtering.

    • These sorts of systems often benefit from a narrow band-reject notch filter at 50Hz or 60Hz depending on country, to remove hum from the power grid.

    • I would probably use an external 12-bit ADC with a good independent voltage reference source, for example an MCP3208 connected to the microcontroller over SPI, rather than relying on the AVR’s internal 10-bit ADC and internal voltage reference.

    • I would use a star ground layout where the analog-side ground and the digital-side ground are only joined at one point, at the ADC.

  • I don’t understand why the PCB layout is spaced out so wide. Basic EE literacy will tell you that a bypass capacitor should be as close to the IC’s power pin as is possible (and it wouldn’t hurt to put the MAX232’s charge pump capacitors closer either) but this is obviously not the case here. There seems to be a lot of unnecessary copper track length.

  • Have you considered changing the boost converter to a more capable chip such as the LT1302, so you can get 500+ mA? It’s easy to hit the 150mA limit, say with an Arduino and an XBee Pro and a GPS and a few LEDs, and if you try and draw excessive current the boost converter voltage will drop off, potentially creating strange results that are hard to diagnose, especially for beginners.

  • This switch appears to have three terminals, from the photos.

    Therefore, I assume that it works just like your common sort of illuminated power switch of this type.

    Two of the terminals are silver, and one of them is gold.

    The gold one connects to ground, the first silver one connects to your 12V supply (or 5-12V, but not higher, because the resistor in series with the LED internally is designed for 12V) and the second silver terminal connects to your load +12V rail.

    Basically, the first two terminals are just connected to the SPST switch contacts, and the LED anode (and resistor) is connected to one side of the switch, with the LED cathode connected to the third terminal which you connect to ground. It’s gold plated so you know this is the “odd” terminal which is not connected to the switch.

  • What are the pros or cons of this device relative to other digital-interface 3-axis accelerometer chips that are available?

  • I would love to see this updated to include comparison of some of the newer product lines such as the MMA8452 accelerometer. :)

  • According to the datasheet, the positive acceleration-measurement is in the direction from the side of the chip with the notch on it, but on your board the silkscreen arrow appears to point the wrong way? Is that right?

  • It should also be made more clear in the description of this product that a triac-based SSR does not work with a DC load, it will only work with an AC load.

  • I’m impressed to note that I’m not the only one who draws block-diagram documentation using the Eagle schematic editor :D

No public wish lists :(