Micro Moteino?

[gregcope]

Hi All, esp Felix!!

Felix - any thoughts on creating a micro moteino.  There have been lost of great examples recently and I am sure there is a market!

[Felix]

Hi gregcope,
I think you may refer to the Tino created by forum user joelucid, I don't really want to steal his idea and go do my own thing based on that. Unless he wants some kind of collaboration.
When you say micro, what defines a micro node other than size?

[gregcope]

Yes - was referring to the Tino and others.

I choose "Micro" as a generic term for "little"

I think the main thing that defines a "micro" is coin cell sized.  Could be coin cell or Lipo, but just that Penny sized "thing" and probably only a few GPIO pins exposed.

Not wishing to step on any toes, and I appreciate your concern, however people are making these due to a gap in the market.

Just an idea.

[Felix]

I will let joelucid respond and share his thoughts here about that possibility or his vision of Tino.

[syrinxtech]

I think it's a credit to Felix for designing the current Moteino platform and supporting it in a way that encourages others to develop their own creations.  Collectively, we all benefit when one person builds on another's work.

Congrats to both Felix and joelucid.  There is room and specific applications for everyone.

[joelucid]

I will let joelucid respond and share his thoughts here about that possibility or his vision of Tino.

Sorry - I missed this. Here's the Tino thread for reference: https://lowpowerlab.com/forum/index.php/topic,1269.0.html

I think there are two important use cases for these Tino style boards: long term coin cell operation or energy harvesting for infinite battery life (with Lipo or supercap). I'm not really interested in Lipo by itself because of the high self discharge rate.

Essential for any micro-board is a wireless bootloader: you just cannot use optiboot to upload a sketch because it's not coin cell optimized. The wireless bootloader is also essential so that no headers need to be attached (or heaven forbid pogo jigs be created). And there is just no space for an extra flash chip to support Moteino style OTA updates. See https://lowpowerlab.com/forum/index.php/topic,982.0.html

If you just create a normal arduino app for these nodes you will quickly sit with a dead battery and/or a radio trying to send/receive forever. A lot of additional software is required to enable writing really energy efficient code. E.g:

- Effective sleep code
- Modified RFM69 library
- Listen mode (see https://lowpowerlab.com/forum/index.php/topic,982.0.html)
- Low voltage recovery (sleeping radio and recovering VCC when voltage crashes)
- Minimum voltage reporting for each wake pulse
- Wake pulse length reporting

My vision for Tino is to provide all of this software together with the HW to have a really great out of the box experience: Just unpack your Tino (coin cell already soldered on), attach a separate USB wireless programming adapter to your laptop, compile a sketch using the Arduino IDE, upload wirelessly and run.

HW wise I want to make a couple of additions:
- A RTC or a HW watchdog timer (will reduce power consumption by another order of magnitude).
- And I'm hoping to add some circuitry that will add a fully programmable boost converter to the board to support higher voltage applications (I need this for a sprinkler controller and for a PIR sensor).

The boost converter would be implemented using the 328p to avoid another chip. It would not generally power the Tino or the radio itself to avoid efficiency losses. But it would support powering higher power circuits by charging a cap in short bursts over an extended period of time and the pulsing the joules into the circuit during short wake periods.

If I can fit these components it likely means only one contact header for expansion.

In the end Tino should be usable for most battery use cases (considering coin cells come in up to 1000mAh ). And it will likely be the most power efficient Arduino clone on the market today.

I really have been doing all this just for fun for myself. But I have to say it's so much fun working with Tinos I've been thinking sporadically about making them available to others. I also think there's a lot of misunderstanding in the market wrt coin cells - so many people think what Tom and I have been doing is just not possible.

One of my constraints is that I don't want to manufacture the HW myself - I have way too many left hands for that . So I'd have to partner with someone for that piece.

So what do you say? Is this interesting? Would you want one? Am I missing anything important?

Joe

[gregcope]

One of my constraints is that I don't want to manufacture the HW myself - I have way too many left hands for that . So I'd have to partner with someone for that piece.

Felix might be able to help?

So what do you say? Is this interesting? Would you want one? Am I missing anything important?

I think there would be allot of interest (for a given value of allot in this space).

[joelucid]

Ok, I measure some cautious interest.

Felix might be able to help?

Not sure if this is a really good fit for Felix manufacturing operation. It's certainly seems too early to seriously consider. And selling small self produced batches is not really interesting from my perspective as I have said.

I think I'll just keep on going, letting fun dictate the direction. If something comes together that might appeal to a larger audience I might start a kickstarter campaign to gauge interest in the larger community. There is significant work that would be required to make Tino and the software around it foolproof - and it only makes sense for me to do that if that interest is measurably there in reasonable size.

I think it's a credit to Felix for designing the current Moteino platform and supporting it in a way that encourages others to develop their own creations.  Collectively, we all benefit when one person builds on another's work.

I have to strongly agree with that. Moteino really hits a sweet spot. Thank you Felix!

Joe

[Felix]

Thanks guys for your thoughts.
One particular aspect of my "manufacturing operation" is that I need to keep it as simple and optimized as possible. I cannot spend too much time on a single unit of anything I make. It has to be "designed to manufacture" as often the term is thrown around by the industry. I look at the tiny coin cell nodes as even more artisan creations because they would require quite a bit more manual attention (it looks like).
I am optimized to assemble and test lots of boards in a short amount of time. Hand assembly is where I don't have a lot of capacity, hence part of the reason I sell kits and not ready made products. One of the most significant time burners is soldering the radios to the Moteinos and loading them with test sketches, that probably takes double the time to assemble the board (less reflowing which is a fixed amount of time but only requires me to drop a panel in the oven).

[joelucid]

Hi Felix,

Big picture is clear. However a couple of questions / comments:

because they would require quite a bit more manual attention

In what regard other than support do you think would they require more attention? I know I'm having trouble making them without bridges, but I've so far attributed that to my little experience with reflow. Do you think these problems come necessarily with the small size?

One of the most significant time burners is [...]  loading them with test sketches

Now at least for the test part you could probably save time using the wireless bootloader.

[Felix]

Joe,
Re small pitch/bridges - yes & no. It depends on design, stencils, reflow profile etc.
You have to burn a bootloader and fuses to the atmega and that has to be easy and fast. Then you have to solder the RFM radio, and on top of it the battery holder. That is quite a bit of work, what if someone wants 200 units in 2 days, can you do that? It's a question you have to be prepared to answer. So all is assembled before you have a chance to properly test and see if anything fails, and if so the unit is basically scrap. To reduce that to a minimum you have to have a great design and a great assembly and test procedure.

[kobuki]

I think it's related to the subject. I've never used any of their products, but this thread hints at something that should already available and as low power alternative, directly targeted at battety operation, and looks very interesting. Till now, since a year already, it's all promises and no delivery, sadly. Here.

[joelucid]

Interesting Kobuki, thanks for the pointer. They also use an RTC to get sleep power very low. However the modules with 3.3V LDO regulator are hardly coin cell compatible.

Actually the combination of 3.3V LDO, supercap and radio module looks very weird to me. If it runs with coin cells there must be no 3.3V LDO. If this is optimized for batteries other than coin cells why add a supercap? If the battery is empty and you have a radio module the server can just send a new stamp when the node had its battery replaced.

[kobuki]

Yeah, dunno about the supercap either. Maybe the LDO is there to allow adding components with more stringent power requirements?

[kobuki]

It's not exactly "micro", but when it's going to be available finally, it should rock as a battery-operated, ultra low-power device. It might be interesting in your pursuit for a better LP node - all design choices are documented, bootloader and other code is available, so might be useful to gather ideas. The node itself in its current form looks like a xmas tree with all the bells and whistles (even a Groove socket is included), but that should make life easier for beginners or users that only wish to have a customised remote node for anything.