OSO Level

WiFi Refrigerator Scale

10/2 Next Steps

Posted by Taylor Bernard on October 2, 2014

- Research high pass/low pass hardware circuits to eliminate unwanted vibration pickup
- Refine software side of vibration detection/threshold to determine running state
- Minimize component size by putting battery switch, piezo amp, and RTC on single board
- Test on bike

Comments

Justin Bernard on October 2, 2014:

thx 
shoot us a pic maybe of current componetns laid out and outline/id parts so can see whats worth it to spend time/money on to make smaller, if we can add or remove inputs such as led's or reset button if come up with ideas for that for testing later on, etc

Taylor Bernard on October 2, 2014:

Justin Bernard on October 2, 2014:

what is the blk/blue board to the far left?

Taylor Bernard on October 2, 2014:

UNO. do all the prototyping on it. easier to move around and plug into. tinyduino is solder on terminals like i have the rtc hooked up now

Justin Bernard on October 2, 2014:

red chipset is what? not redbear labs BTLE?  i see two stacked small blue boards, one real time clock and i thought the other was the arduino main board driving the project.

Taylor Bernard on October 2, 2014:

5 stacked blue boards are tinyduino. core, usb, sd card, btle, and top expansion port card, red one remote is RTC

Taylor Bernard on October 2, 2014:

the two 'boards' arent connected. the breadboard and the uno are the ones running. the others are loose. but will be used when prototyping done on the breadboard/uno

Justin Bernard on October 2, 2014:

so what can be combined.  you ordered a core board that also has RTC on it right?  so next model will have core/rtc + btle + sd

no way to do away with an entire board for usb?  or the expansion board?

mock it up so i can understand what going on

also how did past few hours of testing go?  what got done/issues hit/etc

Taylor Bernard on October 2, 2014:

all the tinyduino stuff is pretty much individual applications per board. core, usb, sd card, btle, rtc, and output pads board. they are all small but when you stack up that many its an odd shape.
best case scenario would be to take the open source files and put ALL components on one board in a custom layout. but like i said earlier, when you do that you'll have to find someway to bootstrap/firmware upload to the board because the chip will be raw.
another option is to do everything BUT the core chip on a board, and then have the same plug they are using on the custom new board, so everything will be new except the core board from tinyduino, and you just pop that on the back of a custom board. this bascially wont add any size to the board, maybe like 3/16" to the height because it will be piggybacked but thats nothing.

Justin Bernard on October 2, 2014:

is there not an option then of prebuilt components but not tinyduino that combines more features.  ie something that is core+btle+usb/etc

Taylor Bernard on October 2, 2014:

another prob with custom board, not prob, but will take a lot of work. is that all this tinyduino stuff is done by machine. the components are stupid small. i got the guy to layout 1206 series size components on the sense stuff.



note the size ref squares are way off so this isnt to scale. but 1206 is on the verge of being able to be applied by hand, tinyduino stuff is using 0603 and 0402 size. like 1/2 size of 1206 which is hard as is.

what this means is that the layout will need to be changed, every component will have to be swapped to something do-able by hand. which will change the layout of the traces, how stuff fits together, ect
will take a decent amount of time for that euro guy to do that. pretty decent amount of work.

Taylor Bernard on October 2, 2014:

no they dont have any combined features. i dont know why rtc isnt on the core board to begin with. but i think they try and seperate everything just for kicks. to be able to only put on what you need.
keep in mind people are using this stuff for like clothing, led bounce balls, ect just simple projects, so using this many boards isnt the norm. most people using like 2.

Justin Bernard on October 2, 2014:

thx.  dont worry wish romanian dude for now then.  lets just get as small as can with off the shelf stuff + the board you said you could make to eliminate the bread board.

what got done in last few hours this afternoon?

Justin Bernard on October 2, 2014:


also what your plan on making up hours for leaving early tomorrow so i can plan, you never said.

Taylor Bernard on October 2, 2014:

i'll just hit earlier am and not take lunch

Justin Bernard on October 2, 2014:

asked a few times on skype and earlier in basecamp, what got done in last 3-4hrs since we chatted

thx

Taylor Bernard on October 2, 2014:

tried to add in a filter circuit to chop off some of the possible vibration handling noise but for some reason wasnt working, think it may be chopping signal down to much.
gonna see if theres any simple FFT methods (software) to do it, if not then maybe just set threshold higher to get something working.

Justin Bernard on October 2, 2014:

that filter circuit is the code you mentioned earlioer right

did you build the circuit you said you could do out of radio shack parts you had on hand?

Taylor Bernard on October 2, 2014:

no, hardware version. testing before started making board because if it worked needed to go inline of the circuit.

Justin Bernard on October 2, 2014:

Hardware filter?

Taylor Bernard on October 2, 2014:

resistor/capacitor network.

Justin Bernard on October 2, 2014:

gotcha

whats next steps being that didnt do the trick

Taylor Bernard on October 3, 2014:

need to just do it software based for now. everything past this is just choosing a better sensor and taking unneeded data out. which i dont even know if that GET unit was doing.
so now we just need to make a array for the sensor to turn on.

im thinking  pull data every 100ms or so, if # greater than X is found, start logging array. if after 10 logs, if array mean is bigger than # turn on led and start logging until array is under X.

Justin Bernard on October 3, 2014:

easy

give me the base code you are using and notes on any specifics you want and i can fix if you need me to

Taylor Bernard on October 3, 2014:

ive been using this one with the best of luck so far. repeats some of the same #'s but seems to react the best.

Justin Bernard on October 3, 2014:

you just now starting work?

Taylor Bernard on October 3, 2014:

no

Taylor Bernard on October 3, 2014:

any luck with coding. eta?

Justin Bernard on October 3, 2014:

im having to work on callidus at moment

what you been working on?

i may not be able to get to code stat for few hrs

Taylor Bernard on October 3, 2014:

tired to boost voltage output range but to no avail. its fine as is. just gets up to around 1.3V on a hard impact so would prefer like 5v but arduino can read it fine

now wiring circuit up from breadboard