Tino, the tiny coin cell mote

[kobuki]

Thanks! I'm definitely a noob in PCB prototyping, too. I'm still just learning the basics in Eagle and your Tino looks like a good expertise. I hope someone experienced can help.

[joelucid]

I'm definitely a noob in PCB prototyping, too.

Yeah, then I probably wouldn't recommend to start with Tino. I think it's harder to manufacture than say Moteino and there are also still a few kinks. If you like the form factor wait for the next release.

[jarrods]

Joe, now that it has been a few months, how does it look like the battery is holding up. Do you think you will get months or maybe even a year?

[joelucid]

Joe, now that it has been a few months, how does it look like the battery is holding up. Do you think you will get months or maybe even a year?

You know, I hadn't really looked at it in detail recently. But I just grabbed the data from 9/11 to 11/26 and charted it for one Tino:



Blue is the voltage measured with radio off. Red is the minimum voltage reached during a wake cycle (I measure this interrupt driven in the background to protect against brown out).

You can see that the battery is still going strong - I'd be surprised if it doesn't hold up at least a year. But there are two weird voltage drops which I can't explain. I'm not using acks and I had hoped to have eliminated all potential busy waiting (e.g. canSend in RFM69.cpp). So I don't have a good explanation. This Mote lives in my daughters room in the weekend home and this occurred Friday night. God only knows how she might have abused the poor thing 

There are some battery optimizations I still have on my todo list. Biggest opportunity: this node still sends at 55kbaud and an unnecessarily big ~60 byte packet for each measurement.

[joelucid]

Code: [Select]

But there are two weird voltage drops which I can't explain.

Looking at it in more detail I find that

(a) the voltage drop occurs between the end of the previous wake cycle and the measurement at the beginning of the next wake cycle. So this issue doesn't seem to be caused during the wake cycle.

(b) I also log which time stamp I received in the packet that woke the Mote for the measurement. In the first case this is a 1, in the second a 0. The possible range being [0,2800] I think this cannot be coincidence.

Together it seems as if the radio is kept in rx for a longer period than usual. That leads to it catching very early packets in a burst and it also explains why the voltage is already down when I do the initial measurement. Another listen mode bug?

Tom, how does your data look?

[joelucid]

And here another Tino:



This is from 9/3 till 11/26.

The other Tino was the first one I built - so the battery was already pretty beaten when the series began. This one here was pretty fresh when the series started. Overall this one is behaving much better and more along the lines of many years of battery life.

[TomWS]

Tom, how does your data look?

Depending on your perspective, it looks pretty good 
My Home Server, where the TinyTH motes have been installed, has been unresponsive since l left in early October so I can't gather the battery data.  However, the server is still tweeting the TH measurements from each sensor but one - dunno what happened to that one.

Re your unexplained drops, could it be interference from another transmitter waking up your device?

Tom

[joelucid]

Re your unexplained drops, could it be interference from another transmitter waking up your device?

Possible. Just waking up is not enough though. The radio would have to wake up, switch to rx and ignore the configured timeout, just as we've seen when using address filtering with listen mode. Remember that a burst for another node wakes up a mote and keeps the radio in rx until the burst is over with address filtering.

Very strange. Or maybe it's something with this particular pcb since this Tino also seems to draw more power than the others.

[spambake]

joelucid - In case you never got an answer to your soldering problems, one of the simplest fixes would be to "tent" the vias.

https://www.sparkfun.com/tutorials/115
Scroll down to "Tented"

This will cover the vias with soldermask and prevent most of your problems.  Of course, moving the vias further away from the pads would be a good idea, too.

Great project.  Thanks!

[joelucid]

Scroll down to "Tented"

Thanks! Excellent suggestion. I'm about to do a board again soon, so will try that.

[Lensdigital]

Flux is magic component you need when soldering tiny SMD components
I got the best one I could find for this job and it really works.
Put some on SMD pads, drop of solder on one pad to attach IC. Then I put flux on top of IC leads and start soldering. It really sucks in solder and I almost never get any bridges.

[TomWS]

I got the best one I could find for this job and it really works.

Joe should really like this one!  The datasheet is written in German!

Tom

[Lensdigital]

I got the best one I could find for this job and it really works.

Joe should really like this one!  The datasheet is written in German!

Tom

LOL you should see documentation that comes with this Flux It is imported from Europe.