Some thoughts.

Posted on
  • Love my Pucks, let me say up front that @Gordon has done a fabulous job and I know it's a work in progress. But there are some issues I'm contending with.

    1. Battery out. It's not obvious how often you'll need to do this at first. But you will. The battery is hard to remove. A reset/power down button would be good - or better access to battery.

    2. Flashing. Reverting to Apple/Android smartphones to flash, seems like hard work. We could use a flash routine via the IDE or similar desktop tool, where we are programming the Puck.

    3. Hub software. My eyes glaze over when I see the setup routine for that on a Raspberry Pi. A few have set it up and had no luck running it. I think someone has forked it and had some success. It just seems hard for the sort of person who might buy a Puck - when open to all - commercially. Again I understand this was a stretch goal of the Kickstarter, so the current incarnation isn't what it might look like eventually, but I'm interested in where this project might end up and whether there are other tools available.

    Unrelated but relevant...

    1. Builds. Some, such as ESP8266 build, seem to have broken. I know this is not mainline, but I understood the ESP8266 specifics were pretty stable. I've no idea if Puck additions have impacted the build for ESP8266, but it correlates. We'd all like the latest version of Espruino however we are using it.
    1. Hear you on that. For awhile I was using a pen insert to push with, or sometimes risking a screwdriver when I didn't want to take a pen apart. Finally took 2 thin coffee stirrer sticks and glued them together forming a skinny but thicker, and stiffer, push stick. Works well.
    2. .
    3. Yep. I'm the dude that forked it to make some changes with great results (for me at least). When @Gordon gets back well see if he wants to pull it in. Up to him.

    I'm thinking of sacrificing a puck (well, just its mobility really) and making a puck breakout/breadboard just for development/experimentation. USB power, tactile buttons for power cycle, button click, etc. Breakout of all the I/O. Maybe add a small display. Maybe a small buzzer. And importantly, 3v3 TTL serial so I can have a proper connection to the console that doesn't interfere with the BLE. It's just a thought.

  • I was was thinking of a Puck.js jacket


    1 Attachment

    • Bildschirmfoto 2017-01-04 um 03.54.49.JPG
  • I think the breakout board & jacket ideas are very good, particularly the serial connection - that should be in my original list - monitoring what your program is doing isn't easy since we need to disconnect for any Bluetooth application

  • BTW regarding 1:
    a simple way to "power-cycle" the puck is to get your fingernail under the battery and lift it slightly from the PCB towards the springs of the metal clip.

    A quick way to reboot a stuck puck or get into bootloader mode when pressing the button while doing this...

    For suspending it longer just push in the little transparent plastic tab - but leave it sticking out wide enough so you can just grab it to get it out again.

  • Unrelated but relevant...

    Builds. Some, such as ESP8266 build, seem to have broken. I know this is not mainline, but I understood the ESP8266 specifics were pretty stable. I've no idea if Puck additions have impacted the build for ESP8266, but it correlates. We'd all like the latest version of Espruino however we are using it.

    Actual ESP build is using sdk 2.0 with boot_v1.6.bin which will cause infinit reboots.
    Using boot_v1.5.bin will fix this. No idea what is wrong with boot_v1.6.bin.

  • @Ollie I am with you on your comments and suggestions. It seems there is a lot to do to get the Pucks to be useful to the average maker.

    Standalone, Puck to Puck and web IDE it appears to be reasonably stable, but right from the beginning of the Kickstarter project I was hoping for some form of bridge to the real world of IoT via the Internet.

    I have so far made good progress with node-red despite it's set-up requirements, but it is still in it's early stages of Puck development / integration. However I have every confidence that these glitches will be ironed out. Hopefully @Gordon will be able to contribute soon on this.

    As far as I see it, the main sticking point is in the lack of clear beginners guidelines on how we can communicate data from and to the Puck. Most of my time spent has been in trying to simply understand how to send a few bytes of data in the correct GAT standard. @dklinkman has some interesting posts on data formatting etc and worth checking out.

    I was wondering if it would be possible to start a Puck related "How To" guide for sending / broadcasting data using the correct data types and standards? Probably something that should be set up a bit like the Bluetooth SIG has for industry standard communications, but in more friendly to read way.

    All this being said, I think the Puck is a great project and will / is becoming an integral part of my currently growing IoT home.

    1. Battery - yes, it's a pain. As things improve I hope it'll become less and less likely that this is needed. Shorting out the battery for a fraction of a second isn't hugely bad - it'll just reduce your battery's life a bit - it's a dead easy way of resetting since power and GND pins are so close, but if you can do it by disconnecting the battery it's obviously much easier.

    2. Flashing - it'll be part of the Web IDE soon (at least via Web Bluetooth). Nordic have said they're coming up with a Web Bluetooth flashing tool, so I'm just waiting for them.

    3. Hub - I think a setup script could help a lot here... Just 'copy raspbian to the SD card, run this file'. I hope eventually I might be able to do pre-flashed SD cards or maybe even sell a kit with everything pre-installed, but the software isn't really mature enough right now.

    4. Serial/power: Adding USB-serial and USB power to the Puck only requires a $5 USB-TTL adaptor and soldering 4 wires though, so it's not too difficult a task.

    Puck Jacket - that could be interesting - perhaps it could use Pogo pins and a case could be 3D printed such that the existing silicone top still fitted? At some point it starts to be more sensible just to put the module itself onto a new PCB though - it's something I'm considering doing at some point (a small PCB in the Espruino WiFi form factor)

  • a simple way to "power-cycle" the puck is to get your fingernail under the battery and lift it slightly from the PCB towards the springs of the metal clip.

    Actually, that works very well. Thanks!

    1. I use a little cable between the GND and RST pads. Isn't that pretty much the same? Seems slightly more elegant?
    2. That will be so great!!!

    I would also certainly be very interested in the Puck Jacket. Especially if it has the pogo pins. If someone produces it, I will be in!

  • a simple way to "power-cycle" the puck is to get your fingernail under the battery and lift it slightly from the PCB towards the springs of the metal clip.

    Wow, just tried it - that does work well :)

  • Should there be Puck v2 would it be practical to add an accelerometer? I'm guessing in the absence of one, the magnetometer could be used as a proxy for this sort of thing by detecting speed of reading change?

    Also, my kids have acquired Microbits and I'm considering putting Espruino on them. Is any of the Puck BLE piece likely to make it's way into the Microbit espruino build? I had a brief look at the Microbit spec and it doesn't appear to be a Nordic chip so I'm guessing no or it would be a lot of work that's not worth doing given the user base?

  • I was wondering about an accelerometer - the issue really is price (and also board space). If I can source a cheap magnetometer + accelerometer chip that handles both then possibly, but I think it's unlikely I'd change it for a while.

    I guess potentially there might be an uber-puck that comes with loads of sensors (especially when the BME680 environment sensor finally gets released), but I wonder whether that'd be popular enough to justify it.

    http://www.espruino.com/MicroBit :)

    MicroBits do use a Nordic chip, but an older, cheaper one. There's very little RAM available, and it can't connect to other devices. However you can still connect to it (even from a Puck) and can do some fun stuff.

  • I missed the Nordic bit on Microbit as I was looked at the BBC website. Not being able to connect from it would limit it's utility though. Thanks for clarifying.

    1. I just use a good old match to pop out the battery. Maybe you could include a match with each Puck? :)
    2. A setup script for the hub would be nice. I'm hestitant to try my luck on my OpenWrt router after seeing how hard it can be to set it up on a Raspberry Pi. Hope it won't be too difficult to run the hub software on OpenWrt routers with Bluetooth 4.0 dongle.

    I'd definitely buy the uber-puck! Seems like we've been waiting for BME680 forever, maybe it's a vaporsensor after all :D

    Edit: Just found out that it's coming out in 2nd half of 2017. @Gordon, if you haven't already got samples of BME680 you can contact this guy who made a BBQ sensor with it (he works at Bosch, his email address is in the comments):
    https://www.youtube.com/watch?v=G0rcjygz­mwo

  • @Joakim thanks! I think right now I need to focus on the current Puck, but I'll see what happens in 6 months... The timing could end up being quite good.

  • Of course, I'm sure you have your hands full after the release of Puck.js. There's no hurry, we won't get bored with our Pucks any time soon :)

    I hope the BME680 lives up to the expectations. It's already used by a few projects, like uRADMonitor model D, and Radu seemed impressed.

  • @Joakim, Matchstick. Brilliant! I was skeptical at first of @ChristianW 's fingernail trick. But it works great and now that's what I usually do. Too much fun.

  • Post a reply
    • Bold
    • Italics
    • Link
    • Image
    • List
    • Quote
    • code
    • Preview
About

Some thoughts.

Posted by Avatar for Ollie @Ollie

Actions