ESP8266–An Interesting Development

Updated Sunday night:

I got up this morning to see if my ESP-01 board now set to run at 57600 was being any more reliable, constantly polling the state of an output and sending it back over WIFI… erm, no, the board had reset for some reason. Still struggling with the official software. Anyway, I thought I’d scan the web to see if anything was new. I’ve been using Google since it was invented and I’ve only just discovered the option to see only what’s new (I know, sad).

Anyway, I stumbled across this link – something called esp8266-frankenstein – “alternative firmware for ESP8266 modules”. In there is a BIN directory and in there are two .BIN files. I took my new FLASH DOWNLOAD_TOOLS_V0.9.2 tool (for Windows) using PYTHON (essentially loaded Python onto the PC (couple of button presses) and run the code… that lets me download multiple .BIN files onto the ESP8266 boards… I’d never had any luck up to now but then, these two .BIN files load at 0x00000 and 0x40000 respectively. I punched in the information and… GOOD GRIEF – it worked.

tmp6CBCThis is completely new software for the ESP8266, it says “release candidate”, Andrew has even made a very nice prototype board up to add reset and other bits to the board… but up to now I can find zero documentation because there isn’t any – this is in fact just very early alpha stage software – but apparently an update will appear late tonight – worth keeping an eye out? The new software does not use AT commands, but has a pretty verbose setup… here it is in action when I type help. As you can see, nothing like before. When I asked it to show a list of access points, it showed a disappointing one access point (normally shows 3 or 4) but I’ve no way to understand right now why the discrepancy.  If you’ve figured out how to blow .BIN files into your #esp8266 board you’ve nothing to lose but time by having a look at this to see if it has wings.. and it is Sunday after all. As things stand, the commands are indecipherable to me but then it may be this is some standard I’m not aware of – or maybe I’m just missing the point.

Has anyone else spotted this software (well, you have now)? At least out of this I got the confidence that the Python-based chip blowing software actually works so if nothing else that’s a plus.

As an update, a new set of binaries were uploaded tonight and some instructions – but this loops permanently for me – I’ve been in touch with the author and as soon as I get it working (next iteration hopefully I’ll try some simple TCP/IP socket commands – if this works – could this be an alternative to the buggy AT command set??? Time will tell.


16 thoughts on “ESP8266–An Interesting Development

  1. Yes, I saw this just a minute ago on their BBS, and immediately thought I might know someone who saw it before me… yup.
    Looks like another clever guy to me. Be interesting to see where he goes with this.
    I joined the BBS yesterday along with details of how I would like to use the device etc etc but explained that I could not entertain it in product whilst the on going bug was present. Maybe I’ll get the same answer you were given …. we’ll see.

  2. I’ve looked through the source code and it reminds of me of the Dr Who Tardis -there doesn’t appear to be enough in there… I can’t see a list of commands and what they do… might be missing something. I’m thinking.. “breakfast” and that might be damaging my concentration. A short break then I’ll take another look. Do let me know if you get anywhere with the bug. I am dying to use this device in something useful instead of devoting my time to trying to make it work.

  3. I’ll let you know as soon as I hear anything positive… this thing is bugging me in a similar way to you …. constant nag, constant thoughts of a possible fix. I’ve sent off for more brain cells, but can’t remember where I was supposed to pick them up.

  4. I did kindly get a response from ‘Necromant’, the author of the Frankenstein work, in relation to him possibly seeing where a problem might be, associated with the old ‘busy s’. He stated, probably making a load of sense, that writing clean routines is a lot easier than figuring out an issue with other writers work. He seems to be progressing at a some pace, so hopefully we may have something considerably more stable quite soon.

    I’ve ordered a few more units, so I will be crashing in there myself soon. I hate bricking things without immediate replacements. Sometimes I run out of expletives to call myself! At least not too often.
    Still, the reply from espressif bbs is still in the wings.

  5. Well, I got no useful info from espressif. Didn’t really expect any. But, I’ve been playing with the unit up at 921K baud. Interesting. Can’t get it to do a busy lock up. I’m using a web page that has 12 gauges, 4 sliders and 10 toggling switches, all of which are requiring feedback for verification. Cipsend is sending back 850 characters of response and xml format. Sure I can lock the page up by going silly with it, but it hasn’t busy locked, which was so easy to do at 115K. There is no reset failsafe on this one.
    Up at 921K the response time is below 100mS.
    I’m using a cheap STM8 which could easily be stuck on the back of it, which I might well do, if it pans out.
    More work to do, of course! My opinion is that sure, a lot of ‘great academic’ work is being done out there, but how about a solid and stable unit first?


  6. That’s really interesting Dave… I wonder why it would appear to be more reliable at higher speed… I agree with you about the stable unit and while it’s great that there are two or 3 groups or -people out there developing alternatives, one has to wonder what’s wrong with the manufacturers that they are not knocking this on the head by simply producing a fix. They must’ve had enough feedback by now, surely.

  7. To me, getting reminiscent of the dotcom boom. Hype, living the dream, easy money, but no real substance. ( bit heavy there). The trick was to actually know what was real and of actual value.
    Ok, so it doesn’t have a hyperwarp 7th dimensional phase limited thermodriven facet wangle, but it works, is solid and won’t let you down. That might not impress those that want to create an environment where all you do is find problems to fix, and spend a life time doing it, without benefit to anyone.
    For me it would be finding the real applications for such a device, know your ‘real’ market for the level of product you have, and getting it out there. Make a mark11 after that. …. enough of that!

    I have asked on their bbs if they could just have a source for the latest ‘fixed’ or updated single bin files, and why 921k operation is more stable. The cloud. I don’t think so.

    So, having found that the high baud rate ‘appears’ to be better, I’ll work backwards, then experiment with cipsend lengths etc. I would have thought they would have made sure that the chip worked at the fastest rate and geared around that. Once that was established I suppose it would be presumed that anything slower could easily be coped with. Maybe their uart patch fixes it.
    ( ok. just let me have the solid bin file, where’s the problem with that).

    Anyway, I’ll dabble a bit more and see what happens with the four devices I now have.
    ( apologies for my rambling on).


  8. O.k. Just went back to Netio app on phone, same software as before on STM8 but changed to 961K. Interval at 330mS, 2 sliders, toggle switch, feed back of slider vals, 2 adc’s sent back. Whatever I do, fast buttons, swap netio configs, switch off, change ap’s then back etc etc. No sign of one busy.
    Hmmm. might start to use this now. I’ll thrash it tomorrow some more…. nice meal coming up, want to keep happy, not the other way round:)
    Using a cheap STM8 discovery board by the way.

  9. Oh dear. I have no idea how to convince an Arduino-type board to work above 115,200 baud.. otherwise I’d follow you and test NETIO (which I use all the time)… hmm, I’ll try Googling that one..

  10. It’s getting worse – The software has now been updated to 9.25 – but I can’t seem to change the baud rate from 9600 now ???

    The Electrodragon site says this..

    Download the bin file
    Set the module to update mode, connect the module : choose “tools” – “configure device”
    Upload bin file: API Test – Flash image download
    upload the bin file at 0x0000
    upload the second bin file at 0x4000

    Well I have not idea what the first bin file reference is about – API TEST??
    I loaded the at 0 and the at 0x4000 as instructions..

    AT+GMR reported exactly the same software as before


    But then the baud rate command would not do anything … and checking for routers returned the dreaded BUSY message…

    Is anyone else getting this?

  11. Hmmm … not so clever then….
    I decided to neaten it all up, ready for the next stage. 4 switched leds, 4 pwm leds, 4 pots for a/d. socketed esp on the STM8 card. What? busy again? Slow? Yup. ok, what’s different. The messing about stuff was bird nest wiring. So, ok try 10cm thin ribbon cable to the esp. Yup, back to faster again. Put the board back in the same position with the cable attached, no problem. Hmmm, Mysteries of the deep eh? Bit of inductance and no caps right on the board.
    Phone twice as fast as the pad for response to sliders and switches.

    Think I might just put my drink down, get up and turn the volume up instead!
    Alright, I’ll just have another drink and crack on. Got some other units arriving…

Leave a Reply

Fill in your details below or click an icon to log in: Logo

You are commenting using your account. Log Out /  Change )

Google+ photo

You are commenting using your Google+ account. Log Out /  Change )

Twitter picture

You are commenting using your Twitter account. Log Out /  Change )

Facebook photo

You are commenting using your Facebook account. Log Out /  Change )


Connecting to %s