Member #20924

Member Since: June 22, 2007

Country: United States

  • Took the words out of my mouth! My Imp projects I can and have handed off to friends, but my ESP8266 projects stay at home (or I have to configure before passing them out.)

    The ESP8266 is supposed to support, like the CC3000, some sort of UDP-based setup but I haven't heard of anyone getting that to work.

  • Oh, I know, and the thing is that I find the FTDI basic board doesn't supply enough 3.3V amperage to reliably reflash my ESP8266 - and I've yet to find an automatic solution (that doesn't require pressing buttons to reset and put it in programming mode.) I'd happily pay $50 for the ability to reflash without breaking my circuit and rebuilding it - right now I develop the network bits in isolation and then pair it with an Arduino (where I can rapidly iterate on the firmware while it's in circuit) out of necessity.

    I should say reliably reflash some of my ESP8266s - QC is...variable, so far.

    Also, I linked the wrong device, sorry! is the only thing that's reliably reflashed all of my random ESP8266-based hardware without a separate 3.3V supply.

  • Eh, it's more that the Imp is an iPhone and the ESP8266 is one of those non-branded grey-market Android phones.

    The Imp is a little more expensive - and a little less flexible - but just works, is secure, and has great support. The firmware's a binary blob you can't control at all. Key for me is that it's a secure solution - SSL to their cloud, and SSL to other points. I've used Imp-based production hardware and it's completely reasonable.

    The ESP8266 can be modified - flash whatever ROM you like, build your own, raw hardware access without a SWD programmer! - but there's a lot of grey area (like the licensing for the various components) and binary blobs around key features. SSL stopped working at 0.9.4 and is now back in 1.1.0...but it's limited to TLS v1.1/1.0 and I still don't see a way to either pin certificates or do certificate validation. The community's pretty good, and Espressif is helpful (and making smart hires like Sprite_TM) but nothing compared to the help I was able to get from the Imp founders. I'd feel really nervous shipping a solution based on the ESP8266 to customers - bootstrapping an ESP8266 on to WiFi (read as: without reflashing/UART) is really poorly documented, at least in English.

    The NodeMCU and NodeLUA projects are pretty interesting but, having used both them and the Imp, the Imp is vastly superior (and Squirrel vs Lua is largely a wash.) Really, one day I'd like to be able to buy Imp licenses for the ESP8266.

  • The problem with this breakout is that bootstrapping it to a network is moderately annoying; the SD card form factor + is much easier to work with. The Imp isn't just 3.3V GPIOs, it's relatively fragile 3.3V GPIOs IIRC, so I try to avoid using them directly. :)

    I really wanted to see the SparkFun take on - I'm still looking for a good ESP8266 development board with FTDI built in and reliable flashing characteristics - all of my boards require a separate 3.3V supply to flash, or a separate programmer specifically built for the esp8266. :(

  • server-side solution for the imp is something the ESP doesn't have at all

    Granted, knowing what I know about the ESP8266, it's probably not at all secure and unlike the Imp I'd never consider it for production...but the price is right.

    The Imp has made huge strides - and I'm a little out of touch - but it really wasn't great at offline for a long time. There's also no way that I'm aware of to push native code to the Imp, so certain low-level operations need to be creative (like using SPI commands to drive IR LEDs / 433mhz radios) or can't be done. Of course, setting up a working ESP8266 SDK (not NodeMCU, but the actual C SDK) is all kinds of fun.

    To be completely fair, I tend to use both chips in combination with some kind of Arduino anyways to get around the various limits involved (and the fact that it's just easier to rapidly iterate on the Arduino, and if I screw up, it's cheaper to replace.)

  • Hmm. Tried to answer this one myself and it seems like you're familiar with the products in question...

    Any advantage of this board (or the eval board you mentioned) over the Freakduino?

    It's based on the AT86RF231 radio with a ATMEGA328 - I assume this design (being single-chip) might have better power characteristics? Arduino compatibility (especially with all the shields I have kicking around) is a big plus for me.

  • Based on the datasheet (but it's not clear) and based on user reports - the WiFly is not 5V tolerant. Using 5V will, long or near term, lead to your WiFly burning out.

    3.3V for the data pins, 1.5V for the sensor pins.

No public wish lists :(