WIFIO is a dual Arduino WIFI module (ESP8266+ATMEGA328P) FCC approve-able with transferable licence. Can use the 328P for I/O expansion also...

Moderator: igrr

User avatar
By gregware
#31044 AFAIK, the very first ATMega+ESP initiative was an open-hardware board called the ESPToy.

It was a very slick design, used one of the AT Megas with 2 H/W serial, and the board design was publicly available. Second version of ESPToy dropped the ATMega, I think they realized it was too much hurdle getting the two mcus so tightly coupled for not much additional value.

Now, Amica, who sell Node-MCU style boards has tried to launch a discussion (both on this site and on the node-MCU/firmware GitHub) on how to evolve the Node-MCU board, which his also open hardware.

IMHO, it would be good to rather capitalize on Node-MCU, which is the de facto 'standard' board for ESP, and has managed to reduce overhead to $2 (Node-MCU sells for less than $5, just $2 above a bare ESP-12 module).
My feeling is that unless you are chinese, you will not be able to get a very low price, which is still the main selling point for ESP. A price point of $10 is too high, for that price you get a Linux SBC such as CHIP, with onboard Wifi, so why bother managing two lower-capability mcus?

I would very much like to see the Node-MCU ecosystem develop beyond the plain board itself, by adding 'shield' capability, and more modularity for the USB/TTL and power options.
Those shields could be ATMega328, or ATTiny85, or a simple I/O mux (74HC4067 provides 16 channel A/D mux for one dollar).
Doing this would cost much less as it is incremental over an already well established ecosystem with several actors, and would be more flexible (IMHO, still...)
User avatar
By mrburnette
#31074
Those shields could be ATMega328, or ATTiny85, or a simple I/O mux (74HC4067 provides 16 channel A/D mux for one dollar).


Well, that is not how I see it. Maybe an STM32F103'ish solution would be far more interesting - and more capable. Maple Mini clone boards with 128K flash and 20K SRAM are available for $4, delivered.
http://www.aliexpress.com/store/product/5PCS-LOT-leaflabs-Leaf-maple-mini-ARM-STM32-compatibility/213957_1400682373.html
The design is open hardware and easily morphed into a shield or incorporated on the main board.

The STM32F103 is very well supported using this core:
http://stm32duino.com
Coupling the Cortex-M3 with the ESP8266 is interesting, IMO. And, native USB too.

Richard's board, $15 - $20, may hit a sweet point provided it can deliver lots of functionality quickly. Yes, we can cobble together a bunch of components, but a go-to solution that covers 80% of most needs would be on my bench.


Ray
My Stuff: https://www.hackster.io/rayburne
User avatar
By gregware
#31107 STM32F could also be a daughter board, you pick the shield that suits your needs. So you end up using the ESP for Wifi, which is ok and was probably the initial intent.
But this will not be the only board in that game, starting with EMW3165, which is STM32F Cortex M4+ Wifi on a single small footprint board now down to $8 at Seedstudio (at the same time they sell NodeMCU for $12...) and Espressif is rumored to have higher-specs chips in the pipe, all of this could undercut an enhanced ESP8266 combo.

Above that there are already quite many linux capable WiFi SBCs for between $10 and $20 on the market (CHIP, Onion Omega, ...).

My take on all this is that ESP8266 has two correlated competitive advantages, first and foremost is the current $3 price point, and second is the very broad community supporting it.
At the same time as ESP8266 came up, there was MT7681 with similar spec and price, but somehow ESP got much more popular, probably in part because of the combined efforts of Espressif BBS and this web site (kudos to Richards for this!), so community support is a key success factor too. Not sure which will come ahead for the next generation, the bets are opened I'd say...

Greg
User avatar
By mrburnette
#31183
My take on all this is that ESP8266 has two correlated competitive advantages, first and foremost is the current $3 price point, and second is the very broad community supporting it.


Greg,
I'll give it two advantages ... cheap (your $3) and programmable via Arduino GUI (thanks igrr). Powerful, 32-bit with WiFi and enough flash/SRAM to actually do neat projects.

I've just completed the beta code and prototype for a GPS time/date "extender" using UDP ... GPS up in the attic with the ESP8266 broadcaster, ESP8266-enabled clocks around the house.
https://www.hackster.io/rayburne/tardis-time-esp8266-ap-webserver-gps

Point is, the ESP8266 is cheap 'nuff to be a utility device ... somewhat like I used to use atmega328P Minis.

Ray