Babele's dashboard

Hi Tod,

I realized I was a bit criptic. Excuse me. What I mean is, I just rewrote the params for the startup script. In fact I did this a lot of time ago, so I just do it again now: I have 4 BlinkM, all connected in parallel (both power and I2C). Two of them exibit the "just flash on power up and nothing more" behaviour.

So I get my Arduino, connect everything, and load this:

#include "Wire.h"

void setup()
{
Wire.begin(); // set up I2C
Wire.beginTransmission(0x00); // join I2C, talk to everybody
Wire.send('B'); // set startup parameters
Wire.send(0x01); // play a script
Wire.send(11); // play script 11
Wire.send(0x00); // repeat forever
Wire.send(0x5); // fade level
Wire.send(50); // time adjust
Wire.endTransmission(); // leave I2C bus
}

void loop()
{
}

I compile and upload and reset the Arduino. Nothing. I disconnect the arduino from USB and reconnect it, and voila! All my 4 BlinkM are now shining in full glory (and sync, at least for a minute or so...) ...(why wasn't the soft reset enough? I am a SW developer. I learned not to ask when something works.)

So, it seems definitely related to fade speed, or time, I don't know. When it first happened, I thought I had burnt my blinkm, then I thought it was impossible, and I noticed the fast "flash". Then I just thought to my old Moog and when it was not doing any sound, and I thought it was broken and instead it was just the sound envelope too fast to be heard... I experimented with some values for fade and time, and I got this sketch hereabove.

solar power: yup.. I am using this board:

http://www.sparkfun.com/commerce/prod...

it has also a voltage multiplier, as you can see. No big capacitors, though... I should have thought it by myself. thx!

More than once, I left the blinkms attached until the rechargeable battery completely discharged (or, at least, there was no power enough to lit the BlinkMs. In fact I noticed the three RGB colors have different minimum voltage levels..). Maybe THAT is a possible reason for the problem?

EEPROM corruption: I have read all the thread. No, I have no way to reprogram it, and I am in Italy and sending you back the BlinkMs would cost a lot of money. In any case, it seems to me that there was NO corruption.. my BlinkMs works perfectly now... That fix should be done to AVOID such "forgetting something" behaviours, right?

Or you mean that the BlinkMs could really "break" - I mean, some meaningful firmware bit could die because of that fuses and brownouts and it could become impossible to "reset" them the way I did just hereabove?

Nonetheless, it IS annoying.. also because I plan to do some sequencing, so I will probably hit the I2C forgetting-address bug myself. I'll keep you informed.

Thank you very much for your interest and message. I was a youngster intrigued with electronics, some 30 years ago.. Now I am trying to understand what happened in the meanwhile (once you used a 555 to get a blinking LED.... now you use a microprocessor and write some code :) )

Hi, I have 4 blinkms, I uset them with some solar-powered rechargeable power source, so it is easy they have been under-powered, and I experimented the same problem many times on more of them. Being sure they could have not died, I just tried the same thing - attaching them to Arduino and resetting the default startup sequence. It works, even this is an annoying behaviour. It seems to be related to low-power setup, nearly died batteries, Joule Thiefs and the similar... – Babele, on July 26, 2009 08:32

Hi David, maybe the "support" link from Apple Store should link directly to "http://widetag.com/widenoise/support/" instead of "http://widetag.com/widenoise".. maybe clicking the "support" tab on the Widenoise site is not-so-immediate. Anyway I checked and both the emails there seems to be OK.

Hi @Dan, I am one of the developers... in fact we worked a LOT on stuff like data clustering and heatmaps, but we finally decided to hit the store without that feature because we were not fully satisfied with our results. There are even more technicalities - for example, the Google Maps APIs were not-so-complete at the time, so it was even difficult to understand the zooming factor on the user side... which affects how you cluster your data, obviously. And server-side clustering needs power, which we currently do not have. So we left the stuff for later, also to understand which scale should we keep in count when thinking about clustering WideNoise data. Clustering 10000 reading is something, clustering 10 millions is another... sure, we could have started low, but you know, WideTag is really thinking about the Internet of Things, and we want the OpenSpime protocol to keep on with MILLIONS of small objects sending around small amounts of data...

So, hold on. We will get back on that! :)