ESP8266 Community Forum

- Sat Oct 01, 2016 1:54 pm #55958 Interestingly, while searching for clues, I found a YouTube video that addresses this very issue. I am in the process of the work around now. I'll report my success/failure later!

- Sat Oct 01, 2016 5:34 pm #55960 OK all. So as mentioned above, I came across this You tube video https://www.youtube.com/watch?v=uXqTw5eO0Mw that sounded like my problem. I can say it worked at least to test a flash. I say that because I followed the procedure making path changes and it appeared to work. But when I closed and reopened to command window, it seemed to drop it. So I redid it all to continue the test. I was indeed able to flash new firmware! (woot woot). Be advised that the video was written for Windows 8.1; the process should be the same (I used it on win10) for Windows 10 however the GUI's are a bit different. Once I was able to get the system to recognize python and esptool I was able to do the flash successfully .....but....OMG all that typing I thought we'd come so far. I felt like I stepped back 40 years to DOS...or worse yet RealTime Exec or CP/M. I hope someone gets the NodeMCU Flasher up to speed. On the other hand, I did need to practice my typing . Many thanks to Marcel and Martin. I owe you both a beer or if you prefer, the wife conjures up a pretty good Piesporter

- Sat Oct 01, 2016 9:39 pm #55966 At this point I am compelled to mention the fact that what started this was an issue concerning connection in AP mode. (See the initial post of the thread). Sadly after 3 days of searching trying and finally getting the latest version of NodeMCU firmware uploaded, the module is worse than before. Not only will it not connect in AP mode but station mode only sporadically connects. I admit there may be some differences in the versions as pointed out in the docs :

It goes without saying that you shouldn't expect your NodeMCU 0.9.x Lua scripts to work error-free on a more recent firmware. Most notably Espressif changed the socket:send operation to be asynchronous i.e. non-blocking.

but if this means many things need to be rewritten to work........
So we discovered the NodeMCU flasher didn't perform; and after fooling around for several hours I find that the same can be said for ESPlorer. I actually liked ESPlorer on 0.9.x even in it's uncompleted state. Lualoader seems to work a bit better but it lacks some of conveniences of ESPlorer IE: Button pushes for many functions like chip info, flash id, heap, file listing, snippet storage and so on. But the primary function of uploading code doesn't seem to be compatible with the new version of NodeMCU. I know I'll keep several . modules on 0.9.6 for a while, in hopes this all gets fixed. Again a hearty thanks to Martin and Marcel for their much appreciated guidance.

- Mon Oct 10, 2016 12:56 pm #56326 I am posting this as a continuation of the events of this thread in hopes of some insite, and to present some help to others who have experiences like mine. For that purpose I'll state that I've been working with 2 different ESP-12-E's. One somehow just seemed to brick, although that isn't the case. The other, wanting to learn the proper commands of the newer nodeMCU firmware, had a sketch uploaded that caused a panic loop that I couldn't get out of. So both were rendered useless. Referring to that one;

I re-flashed it using esptool. A couple innteresting side notes on this. I am using a windows 10 machine, so it was a bit of a challenge and learning curve gettng that going. I erased the flash memory, and then reflashed the module as described in the instructions on the nodeMCU github page. I found issues when trying to flash with my FTDI converter, with it stalling at the beginning waiting for a header. I changed over to a ch340 based converter and it flashed fine. (Forgive me for being a skeptic in the past on that issue; even though that converter works on some things it didn't for this flash.)

Now after the flash, the module appeared to put out constant garble but when changing the baud rate on the serial monitor to the native 74880, I found it putting out a description of room & tail information. ( assume this has to do with the file system) After a few resets, it put out a message "please wait, formatting file system". This took quite a bit of time so watch for that. After a period of time it reset and came up.

So, I thought I was out of the woods and tried the same procedure on the other one; erase, flash, go. NOPE.
On this one after the flash the only thing I could get to was the continuous output at the native 74880 baud of:
ets Jan 8 2013,rst cause:2, boot mode:(3,7)

load 0x40100000, len 25644, room 16
tail 12
chksum 0x6d
ho 0 tail 12 room 4
load 0x3ffe8000, len 2256, room 12
tail 4
chksum 0x06
load 0x3ffe88d0, len 8, room 4
tail 4
chksum 0x52
csum 0x52
rf_cal[0] !=0x05,is 0xFF
It is obviously resetting and telling me why, but I don't know what it means I've been here in the past and somehow gotten out if it. But I honestly don't know how?!?