Ideas for next evolution of pyboard

The official pyboard running MicroPython.
This is the reference design and main target board for MicroPython.
You can buy one at the store.
Target audience: Users with a pyboard.
Post Reply
Damien
Site Admin
Posts: 647
Joined: Mon Dec 09, 2013 5:02 pm

Ideas for next evolution of pyboard

Post by Damien » Fri Dec 04, 2015 11:31 pm

Hi MicroPython community!

It's been almost 2 years since the Kickstarter campaign was successful, and just this week we started receiving boards from the third manufacturing run. The pyboard has been upgraded to v1.1 (LDO improved, 8MHz crystal changed to 12MHz, addition of JST battery connector) and we now have a lite pyboard with an STM32F411 (512k flash, 128k RAM, 96MHz, less peripherals than the F405, much lower power consumption) that sells for under GBP20 / USD25.

The pyboard with its current form-factor and feature set will remain available and supported for some time to come, but I know that we also need to innovate and provide exciting new hardware that makes the most of the cool new MCUs out there. After all, hardware sales directly support the ongoing development of the software.

So, what would you like to see in the next evolution of the pyboard?

Think outside the box! Are you interested in low cost and basic functionality? Or a quad-sized pyboard with hundreds of IO pins? Or builtin Ethernet? Or a builtin display? Is connectivity really important? I always wanted to make a half-sized pyboard, but that would lack IO and the SD card would need to go (maybe that's not a bad thing?).

There is also the option to develop more "skins" (epaper perhaps?), or support more 3rd party add-on modules (eg LCDs).

Please voice your ideas and opinions!

JimTal001
Posts: 176
Joined: Thu Jul 30, 2015 4:59 pm

Re: Ideas for next evolution of pyboard

Post by JimTal001 » Sat Dec 05, 2015 1:06 am

Hello Damien,

1. A way to deploy protected pyboard code. Assuming someone develops a commercial product using the pyboard their code needs some level of protection. As I understand it, currently it is not possible or at least not easy.

2. Blue tooth on-board (with sleep mode)

3. Output 3.3 pin with On/Off ability

4. Ability to power on/off the SD card so less power is used when in suspend mode.
Last edited by JimTal001 on Mon Dec 14, 2015 6:07 pm, edited 2 times in total.

User avatar
dhylands
Posts: 3821
Joined: Mon Jan 06, 2014 6:08 pm
Location: Peachland, BC, Canada
Contact:

Re: Ideas for next evolution of pyboard

Post by dhylands » Sat Dec 05, 2015 1:34 am

As far as "big boards" go, Olimex has a couple of 407 boards,
https://www.olimex.com/Products/ARM/ST/ ... e-hardware
which has sdcard, USB Host, USB OTG, and sdcard, with a layout for an Arduino compatible interface.

and the https://www.olimex.com/Products/ARM/ST/ ... e-hardware which has ethernet and HS and FS OTG connectors.

Both have gobs of I/O available.

If I was going to create a new board, I think that having a board with a focus of low power would be good (perhaps using one of the new STM32L4 MCUs).

There already quite a few boards with SMT32F4xx MCUs available. So if you were going to create a new board, I think you'd want to avoid just duplicating functionality available on another board.

I think improving the peripheral support (ethernet, LCD, touch screen, camera, watchdog, USB Host) would be better than just creating more types of boards (IMHO)

mianos
Posts: 84
Joined: Sat Aug 22, 2015 6:42 am

Re: Ideas for next evolution of pyboard

Post by mianos » Sat Dec 05, 2015 1:43 am

Something ultra lower power would be fun, like the Kinetis L Series. Maybe with a button battery and a header for the serial programmer. Just looking at maybe the EXP432P401R from Ti. ARM and lower power. Plenty of ram.

SpotlightKid
Posts: 463
Joined: Wed Apr 08, 2015 5:19 am

Re: Ideas for next evolution of pyboard

Post by SpotlightKid » Sat Dec 05, 2015 1:48 pm

dhylands wrote:I think improving the peripheral support (ethernet, LCD, touch screen, camera, watchdog, USB Host) would be better than just creating more types of boards (IMHO)
Totally agree. Especially USB host support is still rare or difficult to implement on other boards.

Turbinenreiter
Posts: 288
Joined: Sun May 04, 2014 8:54 am

Re: Ideas for next evolution of pyboard

Post by Turbinenreiter » Sat Dec 05, 2015 2:35 pm

* a half-sized pyboard with nothing but the microcontroller, USB, LED, reset and voltage regulator
* a skin for that with micro-SD, JST connector and lipoly charger (like MCP73831), more LEDs and user-button

The half-sized pyboard could have 26 pins (2x8 on the side, 10 on the end, pin-compatible to the original board [at least the sides]).
It would be small (33x20mm).
It would be cheap (bare bones and 'smaller' microprocessor, i.e. STM32F401CE).
The skin would solve two main problems - storage and power.

I also think that a couple more skins would make good sellers:

* IMU-skin -> accel, gyro, mag + baro - similar to what we already did in the quadrocopter-thread.
* 3G+GPS-skin
* wifi/bluetooth-skin
* etc.

For the chip itself - there are a lot of crazy chips out there. Just look at this. This is a System in Package from Bosch, containing an Atmel SAMD20 (256k Flash, 32k RAM) and three axis accelerometer, gyroscope and magnetometer. I guess it's not enough RAM for MicroPython.
The question is whether or not chips like that are a good idea for a pyboard. If you don't need the integrated Mumbo Jumbo, be it sensors or wireless coms, you are paying for stuff you don't need. Well, except for the ESP8266, which is cheaper than anything else you can get. And while I think a low-price board would be good, I'd never ever compromise documentation and support for that. I worked with the MPU9x50 chips (combined accel, gyro and mag + 'sensor fusion core) and the main feature - the sensor fusion, was just not accessible, due to lack of documentation of a proprietary feature - which is why I switched to Bosch - they have much better docs and even Open Source drivers.
I also think that regular microcontrollers are more long-living than those crazy chips. I think it's more reasonable to assume that STM will update it's stm32f4x line in a compatible and sensible fashion than i.e. the ESP8266.

The other thing would be whether to stick to STM - less porting cost and more resources to improve support for that family - or to use another platform, like Atmel - more porting cost, but gaining wider audience.

blmorris
Posts: 348
Joined: Fri May 02, 2014 3:43 pm
Location: Massachusetts, USA

Re: Ideas for next evolution of pyboard

Post by blmorris » Sun Dec 06, 2015 4:17 am

One thing that I haven't seen on a commonly available STM32F4 dev board is a ULPI PHY (example: the USB3300 from Microchip) to provide a true HS USB interface. Many chips in the STM32F4xx family support it, and on the higher pin-count devices it wouldn't preclude access to most other peripherals.

I do realize that using the USB Mass Storage capabilities can be somewhat troublesome, especially when a MicroPython application wants to access a filesystem that the host thinks that it owns. However, using regular FS USB, it is cumbersome to get large data files on and off an SD card using just the pyboard. On the other hand, if HS USB were available on a pyboard, I probably would just use it as my primary card reader, in addition to all the other fun stuff it can do.

If I were going to use a pyboard that way, though, I would probably also want an actual mechanical switch to set whether the Host or the pyboard has exclusive access to the SD card or flash filesystem (after the next reset, of course.)

Regular readers here may also know that I am interested in I2S; if there is interest I could potentially distribute the I2S skin that I have developed.

-Bryan

User avatar
pythoncoder
Posts: 5956
Joined: Fri Jul 18, 2014 8:01 am
Location: UK
Contact:

What about the kids?

Post by pythoncoder » Sun Dec 06, 2015 8:05 am

Comparing the Pyboard with similar boards one thing stands out. It requires its users to be capable of soldering. Was this a deliberate decision to aim for the more experienced user?

Why not also offer one with sockets? The variant could be sold into markets like that for the Arduino, Pi and other plug and play solutions. Connectors might prompt the development of third party plug-in skins. The boards could even be marketed to schools...
Peter Hinch
Index to my micropython libraries.

stijn
Posts: 735
Joined: Thu Apr 24, 2014 9:13 am

Re: Ideas for next evolution of pyboard

Post by stijn » Sun Dec 06, 2015 8:37 am

^ good point. Also turbinereiter's idea is of course good (there's a reason it has been done so many times already) but from the company's point of view I think it's not ideal. There are alreay tons of stackable boards out there so it's at least doubtful whether spending time on such boards is worth it. Firstly that time could be well spent in another way, secondly: with so many other options out there it's not sure the time spent can also be monetized - as in: why would the people typically buying these boards buy yet another one with yet another layer stack when they already have X stacks.
So combining all those: I think there is definitely a market for a PyBoard with a socket layout compatible with standard Raspberry Pi or Arduino shields so people can just by the PyBoard and then start coupling it with existing stacks.

SpotlightKid
Posts: 463
Joined: Wed Apr 08, 2015 5:19 am

Re: Ideas for next evolution of pyboard

Post by SpotlightKid » Sun Dec 06, 2015 8:48 am

stijn wrote:So combining all those: I think there is definitely a market for a PyBoard with a socket layout compatible with standard Raspberry Pi or Arduino shields
That's basically the NUCLEO-F401, isn't it?

Chris

Post Reply