A Bit Big for a Bit, But

I’m making the Atomic Clock bit:

Of course, there is not a cesium fountain oscillator in there. But here are some of the attributes of this Atomic Clock bit does have:

  1. WWVB* An ultra low frequency radio connection to a pair of cesium atomic clocks in the Rocky Mountains. The dark gray cylinder with the copper coil is an antenna that picks up WWVB, and decodes the current time. This radio signal is used to set the clock on the bit.

  2. Chronodot (a round CS3231 breakout board) Real Time Clock IC. It keeps track of the hour, minute, and second ( as well as day, week, month, year, and some other stuff.) It’s got an internal, temperature and age correcting, oscillator, so it is very accurate - about one minute per year if uncorrected.

  3. Battery backup. The battery should last at least 8 years according to Adafruit. This bit keeps time in it’s sleep even when it’s not in a circuit. The combination of the WWVB connection. the CS3231mean this device will always be accurate to within a second.

Other features are analog outputs with 8 bit precision** on 3 separate bitSnap connectors. The spacing of the output connectors aligns to the arduino and wireless bits, to facillitate inserting this device into various circuits.

If the clock cannot be set by the radio, then it will try to connect to a time source through the micro usb connector. If it cannot do that then the hours, minutes and seconds can be set by hand.

When the alarm goes off, five, extremely bright (transistorized!) LEDs flash in a clown barf style until switched off. The LEDs can also be used to monitor the WWVB signal.

Here is the schematic:

The front of the bit:

The back:

And a short movie showing a prototype translating the radio signal into minutes and seconds (I only have 2 numbers, but plan on getting a third once I have this running smoothly†. Then I can do some more clock projects - I’m fascinated by time.


.* WWVB is located at a high altitude in the western United States. At night it covers nearly all of North America, Central America, The north coastal areas of South America, a bit of the western coast of Africa, and much of the Pacific and Atlantic oceans. Here is a map of the coverage, animated by time of day (it goes furthest in the wee hours of the morning.)

Similar transmitters are located in the UK, central Europe, and Japan. This bit can be reprogrammed through the usb connection, and the radio antenna can be changed to work with any of the other ULF time signal transmitters. Antenna and receiver chips can be bought from C-Max. (But mine came out of a clock I bought at Target.)


** 8 bit accuracy for a 5 volt littleBits signal gives a resolution of 0 to 255. Time measurements however, are 0 to 24 and 0 to 60. By using 96% of the 8 bit range, the time is represented by 0 to 240, and the conversion math is made MUCH easier. The hours signal has a resolution of a tenth of an hour (in fact it is exactly the same as my earlier Time of Day bit) Minutes have 15 second precision, and seconds could be read down to a quarter second.


† I have various parts of the software running - the radio reads WWVB, and the rtc reports the time. However, this is the most complex bit I have made - the WWVB code and the Chronodot code are not quite compatible - I am currently working on connecting them. There are a number of helpful discussions about this subject at the Arduino forum.

Here is the video of the radio signal running number bits for hours and seconds before I soldered it together:

Any and all discussion is welcome, and appreciated!

2 Likes

Hello Chris @chris101,
CONGRATULATIONS !!
This is really a great Littlebits invention!
You have been working on this silently and now we all can see what you have been doing…
Great !
How are you going to connect the WWVB radio signals with the clock part, I am interested to see how it could be done…
quoting Mick Jagger:
“Time is on your side”…
:grinning: :watch: :grinning:

1 Like

Looks like the perfect companion for the cloud bit! :smiley:

1 Like

@chris101, can you find time to join us at Thursday’s Community Call so we can pick your brain in person about this? If so, come to the Project Buzz room. :smile: … Where we will oooh and ahhh. :smile:

2 Likes

:slight_smile: I will try to be there Jude. Thanks! :+1:

(I have gen-chem lab at 9 (noon in nyc) on Thursday so I’ll be late. Ya know, I kept Wednesday mornings clear this semester so I could attend the comm calls … :wink: )

1 Like

Thanks @blindbento! My interest in a time input bit came from not being able to connect my cloudbit at work. I had been using it as a time marker for the World Time project. So I made a time of day bit to control that, and it just keeps growing on me! :grimacing:

Thank you Alex! (@alexpikkert)

I’ve been super busy with work (in addition to school, I am helping to design a new science building that should be open in two years. The last 3 weeks have been all about how many electrical outlets and faucets in each lab, what kind of chairs in the classroom, where the pencil sharpeners go!, etc. The level of detail in these "ed-specs’ is crazy!) So I’ve not had much time since I posted that teaser picture a month ago. I finished the ed-spec (for now) and spent the entire weekend with a soldering iron.

So, I’m pretty sure I got most of it right, cause the bit turns on and seems to count correctly, but it’s having trouble amplifying the radio signal - I used an NTE928 op amp in my breadboard prototype (the one you see in the video), but a rail-to-rail OPA4342 on this soldered down version. The signal does not seem to be making it through - either I have a resistor in the wrong place, or the 4342 isn’t catching the pulses correctly. I can substitute a pin-for-pin compatible LM324. If I could find an LMV324 in a through hole style, I’d be golden. The other alternative is to run the antenna (and receiver board - that thing on the nylon pedestals - at 5 volts, even though it was designed for 3 volts. It works, but I might have burnt one out that way - not sure.

Right now it uses the 3.3 volt output from the Arduino micro, and runs fine.

As for connecting the radio to the RTC chip - that is what the Arduino is for - the RTC keeps track of the time, the Arduino reads the radio signal every day at 2, 3 or 4 am, and tells the Arduino what the exact time is. It compares that to the RTC time, and if they are less than a second different, does nothing. But if they differ, then the RTC time is reset to the radio time. It should only need to correct the time every few days.

I have wall clocks in every room at work that work this way. The entire mechanism (including all the clock gears, etc) fits in a box about the size of this bit. So I know there is plenty of miniaturization that can be done. Look at this watch antenna at the C-Max site It’s less than an inch long, and covers all of the time signal frequencies, so it will work world wide. I think this device can be made into a bit the same size and shape as a fork bit without too much trouble (use a small antenna, receiver chip, and naked avr328p and cs3231 chips without the breakout boards, and surface mount LEDs, op-amps, and other components. It’ll look more like the picture I made back in July!

So I hope to get more time to work on solving these issues - Jude asked me to bring it to the Community Call Thursday, and I’d love for it to be fully functional by then!

2 Likes

Hi Chris @chris101,
You have put a lot of energy in this clock bit !
I suppose the Littlebits R&D team will assist you in the next step of miniaturization, I really hope (and think) they will accept your RTAC bit…
Indeed the LMV324 seems to be available only in SOIC surface mounting design. My soldering iron is way to cluncky for that design…
But there are transition sockets from surface mounting to Dual In Line Package design:
http://www.globalspec.com/FeaturedProducts/Detail/AdvancedInterconnections/SOIC_to_DIP_Adapters_Lead_Free/116249/0
Maybe for prototyping it could work.
I also read somewhere that the lmv324 has been designed a long time ago (hence the low availability…) maybe there is another opamp rail to rail design floating around on the internet… :scream:
I found this, maybe take a look :smile:
https://www.maximintegrated.com/en/alternatives.cfm/part/LMV324/pk/54

I’ve been acquiring everything I need to start SMT soldering, including itty bitty LV321’s from DigiKey. @Philip_Verbeek even gave me a quick overview of how to get it done. One of these days we should all take the leap and try it!

3 Likes

Hi Jude @JackANDJude,
I also realize that SMD soldering is the future…
For prototyping however I still think breadboard design and more simple soldering will be enough to get results.
I don’t see myself soldering a bit as for example the MP3 player in SMD design… So many parts :cold_sweat: :cold_sweat: :cold_sweat: :cold_sweat:
I suppose that the Littlebits Bitlab team does not expect a fully SMD bit design when submitted and they will take care of this final step during their final review and production, or not ?

1 Like

@chris101 You should be able to find an SMD break out board in your electronics lab. If not there, try Radio Shack. You solder the SMD chip in the middle, then some headers to the outside. A bit of a stretch for one component, but if this would be the final chip used in the bit, it may be worth the extra work. :smile:

2 Likes

Tiny indeed but oh so pretty when it all works :slight_smile:
All you actually need is some tweezers and a good soldering iron with a pointy tip.

1 Like

Thanks @alexpikkert, @JackANDJude, @blindbento, and @Philip_Verbeek for encouraging me to give surface mount components a shot, and of course this is a skill that will be important going into the future as more and more chips are only made in this format. However, soldering an smd to a breakout board doesn’t seem any easier than soldering them directly to a printed circuit board. (My eyes are 62 years old, and I already need a magnifier just to be able to solder regular chips and sockets.)

I did some circuit debugging and found that the op amp I am using (a 5 volt, rail to rail, 14 pin dip, opa4342) is working just fine. I gave it a signal with a much higher frequency than the 1Hz wwvb signal, and it followed it no problem - so that’s not the problem. I did find a switch in the circuit that is not working, This evening I’ll fix that.


[quote=“alexpikkert, post:8, topic:22739, full:true”]
… I suppose the Littlebits R&D team will assist you in the next step of miniaturization, I really hope (and think) they will accept your RTAC bit…
…[/quote]

RTAC - I like that :slight_smile:

I hope so too, but my goal for the bits I make is to come up with something that works for my purposes, and looks good doing so. When I submit devices to the bitLab, I don’t expect 1000 votes. I was delighted, and surprised by the response to my recent module there. I described how to make it in forum posts, so if anyone else wants to make one, it’s all there. If it gets improved and submitted by someone else, I’m all for that - it’s the story of technology advancement. (As an aside, I am using my stepper prototype device in a globe display in the geography department office.)

My reading of the bitLab documentation is that a submitted prototype need only work. I usually go further, and include as much of littleBits’ circuit guidelines as I can conveniently do - input and output especially. This usually means using rail-to-rail op amps and bitSnaps. Then, and only if a bit gets accepted by a review from littleBits, does one need to work on how the actual bit will be made. The documents speak of ‘engineering deliverables’, which include all the details at a component and process level. I don’t think we need to do any of that prior to acceptance littleBits. I would then expect to work with the littleBits staff on the final design.

1 Like

I’ll bring the loupe!

2 Likes

@alexpikkert, I want to learn SMD soldering “just because”. It’s not at all a requirement for bitLab. Still, I wonder if my color sensor might have gotten more votes if it looked like a real bit. It’s certainly easier to demonstrate a compact module’s utility in sample projects with moving parts.

Back on topic, I really dig the way @chris101 designs his pretty circuits. He cares about placement, balance, color coded wires, and photography.

2 Likes

@JackANDJude,
Do you have a shopping list you can share for SMD soldering? :smiley:

OK. I’ve gathered supplies from various places while placing different orders.

  1. Flux - I wanted a pen, but the Shack was all out. Jack immediately stained his shirt with this tub of “Non-Spill Paste” goo, RS 64-022. $7.49
  2. A new soldering tip for my existing iron, RS-2073. Trying to avoid getting a special iron. I figure sharper is good 'nuff, @Philip_Verbeek? $7.xx
  3. $39.95 SMT/SMD 0805 Resistor and Capacitor Book from Adafruit. Why am I especially excited about this?
  4. Copper Tape from Sparkfun, for practice and weird, artistic circuits. $2.95
  5. Printed Circuit Board, RS 276-1499, copper clad & double sided. $4.49 x 2
  6. Etchant Solution, RS 276-1535 $11.49
  7. LMV321 op amps, SOT353 from Digikey.
  8. Not pictured are some cute 25 unit packs of 1206 sized led’s from Sparkfun. Blue, Green, Red, White

3 Likes

Looks like you are good to start your own fab shop Jude!

Great purchases!
Curious what your gonna make with it :slight_smile:

What you build, @chris101, looks fabulous! Great work!

1 Like