It’s aliiiiive!

Muahahahah! It lives! Managed to solder up the first Chronoduino just before my vacation trip and took it with me. Here are some (non-glamurous) glamour shots…

Sorry about the lint. Couldn’t clean up the board properly (had a crappy scented nail polish remover instead of pure acetone) and had to use a ton of swabs and tissues wrapped around toothpicks. By the way, screwed up pretty badly with the debug LED on the main Chronoduino board. I’ve connected it to the 168’s ADC7 pin. Thought – it’s free and unused by anything in the Arduino environment, so why not. Well, turns out it can’t be set as an output pin, unlike other analogue pins, because there is no corresponding direction bit in the 168’s port direction register. So the LED is basically dead in the water. I can’t even use it as a light detector, since it’s underneath whichever face i put above.

Here are the bottom sides:

Those 0402 capacitors were a ton of fun! Putting them in was actually a mistake while designing the board. Originally meant to have them near the magnetometer, topside. Plenty of space on the bottom, so no real need to go that small.

Thanks to the staggered through holes from the SparkFun EAGLE lib, i can put the ISP and TTL headers in without soldering them and remove them after programming. Which is nice. I’ve had a fun day of trying to figure out why the bootloader flashes just fine, but doesn’t actually work nor respond via FTDI. Finally looked at my schematics and realised i forgot to hook up the 3.3V line from the FTDI to anything. So the board needs to be powered by a battery to be programmed. Will fix that in the next revision.

Anyways, here’s what the whole sandwich looks like:

Ended up a bit too thick. Due to headers, of course. I’ll have to find a low-profile board-to-board solution for the next version. The 2-pin JST for the battery on the bottom will have to go too. Simply too huge. Here’s what it looks like all lit up (long exposure):

Gave all the LEDs and sensors a check. Everything seems to be working, except for the RTC. It’s flaky. It refuses to start at times, but runs just fine once it starts up. The Farnell part page said the Citizen 32.768 kHz crystal i’ve ordered was 6 pF, while Citizen’s spec sheet says 12 pF. Not sure who to trust, but that might be the cause.

Realised today i have no way of knowing how far i’ve let the battery drain. And since 168PA can go very low and still work, it’s possible to kill the Li-Poly battery. So i’ll have to add a voltage monitor in the next board revision. Oh, and one more tiny niggle. I’ve got the HMC5883 axes marked wrong (x and y swapped) on the cFace board. Again, something for the next revision.

Finally, here’s a video in all its blinkenlichty glory:

Writing some libs for it now, will post a little demo when i have everything working.

Tags: , , , , , , , , , , , , , , , , , , , , , , ,

5 Responses to “It’s aliiiiive!”

  1. Ferdinand 04/08/2011 at 10:09 #

    The boards look really great. Can’t wait for the first clock demo.
    You can measure the supply voltage of the AVR if you have connected the AVcc pin to Vcc. Now you can connect AVcc to the ADC as reference and the internal band-gap reference as voltage to measure. The ADC value gives you the ratio between the band-gap voltage and the power supply voltage. As the band-gap voltage is nearly constant you can now calculate Vcc.

    • orcinus 04/08/2011 at 14:44 #

      Both AREF and AVcc are hooked up to Vcc, so it should be doable.
      Thanks a lot for the tip! Will give it a try.

      • orcinus 04/08/2011 at 16:55 #

        It works! :D
        Thanks again!

  2. blalor 04/08/2011 at 11:08 #

    Looks fantastic! Is the little chip on the LED board an LED driver? And is that a USB connector?

    Would it be possible to run it off a CR2032 battery?

    • orcinus 04/08/2011 at 14:42 #

      The chip on the LED board is a magnetometer (HMC5883L) :)
      The USB is strictly for battery charging, there was no room for an FTDI chip. I’ll see if i can find something to replace it, as using an FTDI cable and headers is, admittedly, a bit tedious.

      Re: CR2032, it should work, although the voltage might be a bit too low. There is no voltage regulation on board in this revision. The 168PA will work without a hitch, but i have to check the other components (BMP085, DS1337, HMC5883L).

Leave a Reply