ESP8266 WiFi Rant.

[Yansi]

Hi
a few weeks ago, me and a few of my colleagues have decided to buy ESP8266 based WiFi modules and test them, just being curious what they are capable of. Well, as we've learned today, they are capable of pissing not only one engineer really off.

We bought two models of the modules: that cheapest rubbish with the nonshielded board and onboard etched antenna, this: http://www.aliexpress.com/item/1pcs-ESP8266-remote-serial-Port-WIFI-wireless-module-through-walls-best/32280714838.html and one looking much better: http://www.aliexpress.com/item/ESP8266-serial-WIFI-model-ESP-07-Authenticity-Guaranteed/32226744057.html

I'd like to share an experience with these modules and some subjective review based on short toying with these modules.

As an engineer, the firs think you will likely do is to search for documentation.  We have found only some marketing rubbish pdf, which mostly doesn't help with anything. Search for a datasheet specific for the two modules gave us no result. Nothing's been found. Only some "OHSW arduino type" schematic plans using silly images of the components not standardized schematic symbols.

Connecting the module. Looked to be easy. 3V3 supply, ground, RX, TX and that CHPD (chipselect/powerdown) signal tied to VCC. Cause the "OSHW arduino schematics" even were much consistet with how to connect Rx and Tx pins (some "instructables" were to connect Rx to Rx and Tx to Tx), a simple resistor loading test were conducted.  Of course, Rx is the input on the module, Tx the output respectively.

By the way, the chip on the unshielded module gets hot rather quickly - don't understand that at all. And also two decoupling caps were installed on the module - seemed none is there. (wtf?!)

Trying to send an AT command. Seemed easy, until... the fuck what speed does the bastard use? "OSHW inctructables" also weren't much clear about that, nor you will get that from the pdf immitation of datasheet. I mean this: https://nurdspace.nl/images/e/e0/ESP8266_Specifications_English.pdf

After some time labouring with the speed setings in the terminal, the colleague's unshielded module caugth somewhat working at 9600 8N1. But the web instructed mostly using speeds like 115200 or 57600. But the fun haven't ended there. My shielded module has been just sitting there doing nothing. Only lighting the red LED.

Troubleshooting the metal can module. I have noticed, that sometimes on power-up, my metal-can-shielded module produced some glitch in the terminal, puked there some random characters, regardless of the speed was set. I couldn't make tha sucker print anything normal  at any speed from 600bps to 115200. Then I grabbed a DSO a look at the mess on the TX pin.  Looked definitely like UART signal. I have decoded first two characters and got 0x0D 0x0A. (CR LF)...  Then I calculated the speed. 13.2us per bit.  Roughly 75700. Wtf shit is that speed? Some chinese standard?

Trying to communicate with the metal-can module. After setting the terminal to 75700 it finally produced some readable text. Some module ID, two-number bootcode (dunno what that means but gave me I think 7,7). Nice. Then I tried typing AT. No response. Then tried AT+RST. Nothing again. Then I started messing with CR/LF line endings. Nothing doing anything. Only gettin that short infotext message after reset or powerup.

Another web search followed. I learned, that the modules comes with differnt firmwares. And you simply don't know, which one you have in, when you buy the module. So you must first flash it using some (as I read on the web) very buggy flasher tool, which even doesn't support COMs with higher number that 6. I haven't tried to flash the gizmo with anything so far, so stay tuned for update. Quick sniff into the web raised some alarms, that the flashing procedure isn't just that simple.

To sum up the today's few hour mess-around with these modules, I'd say these are pretty crap and an example of BAD product design. Yes, they are cheap, OSHW, Arduino, anything, but more trouble than helpful. Pretty disappointed   

[sleemanj]

You're rant says more about your own inability to perform research, put in time and effort, and recognise than a new extremely cheap device under heavy active development, from China, is not the same as buying a mature stable well documented expensive device from a respected manufacturer.

There is a large quantity of information out there, both manufacturer and community derived, you just have to Google for it.

Communicating with the ESP over it's AT interface is not that hard, and there are alternative firmwares, and the ability to code for it directly if you want to avoid AT command communication.

Yes the documentation is a bit rough and ready, it's an experimental device from China, if you really wanted to use this device you would have no trouble to work it out.

[mtdoc]

Maybe you've seen this, in case not:

esp8266 forum

[stevech]

er, please edit language to yield a "G" rating.

[Yansi]

You're rant says more about your own inability to perform research, put in time and effort, and recognise than a new extremely cheap device under heavy active development, from China, is not the same as buying a mature stable well documented expensive device from a respected manufacturer.

There is a large quantity of information out there, both manufacturer and community derived, you just have to Google for it.

Communicating with the ESP over it's AT interface is not that hard, and there are alternative firmwares, and the ability to code for it directly if you want to avoid AT command communication.

Yes the documentation is a bit rough and ready, it's an experimental device from China, if you really wanted to use this device you would have no trouble to work it out.

True. I am used to use devices and read datasheets from respected manufacturers. So quite shocking, how the "OSHW arduino stuff" is inefficient at documentation level.

I don't see anything hard at working with AT commands, but the device should listen them first.  Just a few minutes ago, I noticed on some random ESP-arduino-something-oriented webpage that GPIO15 should be tied to GND on the "SMD module". I don't know, if the module with metal can is called "SMD module", equally I don't know why I should ground the GPIO15, but I am going to try, what happens.

[PeterFW]

a few weeks ago, me and a few of my colleagues have decided to buy ESP8266 based WiFi modules and test them, just being curious what they are capable of.

The same thing i dit about 2 weeks ago, the documentation all over the web.
The firmware running on the chip is available to study.
Even the whole datasheet for the chip is on the web.
If the datasheet, wich is fairly comprehensive, does not tell you all you need to know, what else do you want?

"The chip runs hot" is no failue description, you power supply or your meter should tell you what power the consumed.
Idling the chip draws between 40 and 80mA at about 3.3V, that is ~1/4W max.

The fault is all on your side, lets call it overconfidence and a pretty huge gap between actual and percived knowledge.

[dannyf]

a few weeks ago,

Wow. Multiple ST embedded engineers couldn't get a simple esp8266 to work.

Wow!

Hopefully that's not indicative of ST's engineering prowess.

[Yansi]

PeterFW: Simple. If one chip remains only warm, why should the other get too hot to touch? Same supply, same setup.

Dannyf: Dont understand your point. I really would like to see you, what you would be capable of doing with that just in two hours, with only such pispoor documentation available and no previous experience with these suckers. But sorry to bother you.

[zapta]

To sum up the today's few hour mess-around with these modules, I'd say these are pretty crap and an example of BAD product design. Yes, they are cheap, OSHW, Arduino, anything, but more trouble than helpful. Pretty disappointed   

The 8266 is a game changer. Dirt cheap chip, has open SDK and includes an MCU that can run application code. Compare with similar modules from other vendors.

The 75k boot message is bizarre but just ignore it. The stock AT commands look good initially but when you get to the details they are hard to work with. There is a community activity to have better AT commands but am not sure how about the progress. I ended implementing a small Lua program that implements a a better interface with the MCU (an ARM M0) and it works very well. Actually I am surprised how well it works considering that small printed antenna and non shield board. The next step once my C++ application code is done is to port it to run directly on the ESP8266 and make its SSL to work with 2014 bit keys so it can talk with real servers (stock SSL limited to 1024 bits).

The 8266 is new but is surprisingly open and there is a strong community around it so it should do well. Will be interesting to see how the competition react.

[PeterFW]

PeterFW: Simple. If one chip remains only warm, why should the other get too hot to touch? Same supply, same setup.

Because you do not run if of a proper power supply, dit not bother measuring the load and you have pins floating.
If you leave the wrong pins floating the board will pick up noise, transmit full tilt garbage and draw above >200mA.

You are in pretty big denial

[Yansi]

The power supply is not at any fault, to be sure it was also checked with a scope so was added decouping directly on the module. Still you think the PSU was the issue? Howewer the whole OSHW stuff probably is. I havent seen any arduino instructable suggesting to connect any of the remaining gpios, have you?

[PeterFW]

Still you think the PSU was the issue?

I think i have fed you enough allready, i will put my bag of chow back in the cellar, it is not nearly moldy enough annyway...

[Yansi]

Yeah, fed some inconcrete stuff, ignoring what I have said you. So please have a rest. I am going to do the same now. Tschus und gute Nacht.

[IconicPCB]

ESP8266 can now be programed via Arduino IDE.

Arduino1.61 will recognise the hardware and compile code so that the application runs on the 8266 chip.


Documentation is still a bit wanting but systematic search will uncover a lot of information. You need to search yourself into the hardware.

[Tuppe]

Dude, those modules are DIRT cheap for a reason. At this stage they're more for hobbyists, I wouldn't get professionals working on them.

There's a lot of tutorials online:
http://hackaday.com/2015/03/18/how-to-directly-program-an-inexpensive-esp8266-wifi-module/

There's also the latest(at the time writing this) ESP-11 model available. Problably ESP-25 is already out before I finish this message. I don't know does the chipset change.

If you want proper well documented gear, just use TI's CC3200 for example. You will save money in a long by run using more expensive and reliable tools.
Because of the enormous success of ESP8266(yes, I also have 5 of those. My cat has couple more), I bet markets will be flooded with fake ones that have no guarantee of working to specifications.
It's cheap, but I wouldn't design any serious product around it.

[Yansi]

Did you know ESP8266 requires some pins be pulled down and some pins be pulled up at power up in order to operate properly? Did you went through all available documents before bitching it?

Yes, it sucks, and nobody wants to use it for high bandwidth apps, but it is designed for WiFi toys, where cost it the most important factor. There is s reason why its documents are not complete, because the maker doesn't want you to know the internal approaching.

With published documents, you can surely set up an environment for it, you just have to dig deeper.

How should I know it, if none of the instructables cared about the GPIO pins? Is it my fault, that there no serious documentation exists? Would you imagine, how the engineering world woud work if anything had that much poor documentation, where each information source almost contradicts the other about what should be connected and what should not?

If some pins are needed to be pulled up or down, they should've use pull-up/down-s on the board or on the chip. This is just improper design to save a few tenths of a cent. The same for the missing proper decoupling.

If you'd say that it is better to start with the module by flashing it to a known firmware, rather than relying on what can be found in five minutes, then OK, I get it.

See:
http://www.instructables.com/id/Using-the-ESP8266-module/step1/Obtaining-and-preparing-your-8266-module/
see again? Where are the GPIO connected? And again claiming 115200-8N1. But not true, worked somehow at 9600-8N1. (Is it my fault? No, it is NOT).
http://www.espruino.com/ESP8266
Again, do you see any GPIO connected? I do not. Some mentioning about the 9600-8N1 (yeah, we later found that too the hard way).
http://rancidbacon.com/files/kiwicon8/ESP8266_WiFi_Module_Quick_Start_Guide_v_1.0.4.pdf
What GPIOs are connected there?

Then there are pages, that tell you some more, like:
https://github.com/esp8266/esp8266-wiki/wiki/Hardware_versions
or
http://www.electrodragon.com/w/ESP8266
and this one was the last useful, about the GPIO15, which I've talked about somwhere in the beginning of this thread.

So whose fault it is, that no serious documentation exists for this thing?  I am not at fault, I am only the pissed usser of this thing. The deeper you go into this thing the more you understanded, how bad the documentation is.

There is s reason why its documents are not complete, because the maker doesn't want you to know the internal approaching.

What is the reason? ESP manufacturer doesn't want us to know how to use their product? Have I understood that well? Jeeez...

I think the rant is pretty much warrantable.


https://github.com/esp8266/esp8266-wiki/wiki/Hardware_versions
On this page, there some models listed. Mine is not there, which is I think ESP 7. Found almost nothing for it in a quick search. Yea, someone made a breakout board. Schematic is where?  Again, lack of documentation. Shit. So I must obviously analyze the photos, to get a clue what one is buying. And by the way, ceramic cap directly parallel with the button is a no no.
https://www.tindie.com/products/Ba0sh1/esp8266-esp-0712-full-io-breadboard-adapter/


IconicPCB: I do not care much about arduino, but I appreciate the added functionality. The documentation definitely should be improved.


Tuppe: I know the CC3000, which is NRND now. The CC3200 looks much more as a standalone MCU, pretty.
Fake ones... huh, one more reason for the rant (against the fakers). :-)

It's cheap, but I wouldn't design any serious product around it.

Propably might be done, after rewriting the whole firmware in that thingy. But that was never my plan to dig deeply into that sucker.


Maybe I will toy with that nonresponding ESP07 metal-can gizmo after having a lunch.

[dannyf]

How should I know it, if none of the instructables cared about the GPIO pins?

Time to expand your universe beyond "instructables".

I really would like to see you, what you would be capable of doing with that just in two hours

give me two years I may not be able to figure it out.

But then I don't work for ST so the expectation for me is much much lower.

[amyk]

The official documentation is in Chinese, so be prepared to translate and look into the Chinese part of the Internet.

[Yansi]

dannyf, please STFU or GTFO or both. Never seen you making any useful comment. Noone is interested in your offtopic hate.

After some afternoon internet labour, I finally got that metal can working. None complete english documentation exists, you really must to dig in the web. Kind of compensation of the low price.

Update to the 75kilobaud initial message: After working which pin must connect which level at startup, the module still sends the 75.7kbaud info header after pwrup and then switches to 9.6kbaud.  Interesting to note, that after using the CIOBAUD command,  both the header baud and communication baud seems to get the same speed.

I have toyed with the module for some hours, seems it works just fine. But to get it to work, is just another lesson. Have no problems with the module opration yet. Only ranting about the documentation. If it had some good quality complete manual, it'd be unbeatable product for its price.

Chinese manual... unfortunately I can't speak or read chinese. 

One other sub-rant: Two LEDs on the module. Only wasting power. I appreciate someone likes blinking christmas trees, but for any battery operated project, one must kick em out to preserve some power. 

So the overall impresion of the module is:

+ Really cheap. I think one cannot complain about the price.
+ It packs useful set of features to make things really simple
+ I think the radio part of the chip works quite good
+ Direct programmability of the chip
+ Claims FCC compliance on some modules (quite impressive)
+ High output power: claims +25dBm (300mW) - if true... (note: beware of legislation about Wifi maximum EIRP in EU region)
+ Can do almost megabit over the uart (supports 921.6kbaud)
- Lack of serious EN documentation (Haven't found any yet)
- Information scattered all over the web and you must dig for it
- Incomplete listings for the AT commands (a lot of digging and poking for some parameters and functionalities)
- Sometimes the AT command outputs are formatted ugly - the text parsing software should be more robust to handle the outputs right
- two unnecessary onboard LEDs wasting some power

Even though I could find some more negatives, I still think it is a good product. At least I don't know any better for the price. I only imagined better and complete EN documentation existed because of the ESP being so popular. It would have made life easier.

Now I consider this little experiment closed and I hope, you enjoyed my little rant, only subjective and non-scripted.

[dannyf]

After some afternoon internet labour, I finally got that metal can working.

congratulations in order.

Question: how many ST engineers does it take to get a simple esp8266 to work?

[Yansi]

The question is inapropriate. It should be: How many OSHW arduino fanatics does it take to produce usable complete documentation for their toy?

Last warning. If you have nothing to the topic, STFU & GTFO of my thread.

[dannyf]

How many OSHW arduino fanatics does it take to produce usable complete documentation for their toy?

Infinitely less than the number of ST engineers it took to get it to work?

The Arduino folks, for their lack of technical skills that you so insist on, seem to have no problem getting it to work. Then a group of technically superior ST engineers like you seem to struggle a lot to get it going.

Why is that?

Last warning. If you have nothing to the topic, STFU & GTFO of my thread.

I think the best way to avoid people adding to your thread is not to post it to a public forum.

[Yansi]

What module have you used?  My colleague's  ESP-01 worked almost "out-of-the-box", only the right baudrate needed to be found. On the other hand my ESP-07 module was more troublesome, but finally working.

Recently I came across an issue, when the AT command response from the chip got delayed until next character had been sent in. But not sure, if it is not an issue in the terminal software I am using now. Will need doublechecking.

And also one very useful thing: Just to turn off the annoying localecho done by the ESP chip itself, use the ATE0 command (ATE1 respectively). Why the hell haven't I seen the command listed anywhere? One must dig really for everything... 


UPD: Wow, it really has FCC. But the chip itself, not as a module :-)
https://apps.fcc.gov/tcb/GetTcb731Report.do?applicationId=794971&fcc_id=2AC7Z-ESP8266EX

[zapta]

Try to flash the Lua image. It's makes it very easy to program and you can have your own replacement for AT commands. I programmed mine to connect automatically to the AP and remote server such that the host MCU doesn't need to desk with connections and just sees a bidirectional stream of data bytes.

[PeterFW]

NOTE: This message has been deleted by the forum moderator Simon for being against the forum rules and/or at the discretion of the moderator as being in the best interests of the forum community and the nature of the thread.
If you believe this to be in error, please contact the moderator involved.
An optional additional explanation is:

[Stonent]

My esp8266 showed some signs of life while I was trying to get a serial connection to it. I could see it on a WiFi scan. Then suddenly it got really hot and stopped working despite  nothing being shorted or wired up wrong. This was after having it for several days   It just suddenly said I'm tired of this crap, I'm outta here.

I checked all connections. No shorts.

[PeterFW]

I checked all connections. No shorts.

That is the price you pay for incredibly cheap china hardware, that is something wich has to be expected.
If you pay 3 bucks for something that usually costs 30, there has to be a catch

[Stonent]

I checked all connections. No shorts.

That is the price you pay for incredibly cheap china hardware, that is something wich has to be expected.
If you pay 3 bucks for something that usually costs 30, there has to be a catch

I bought the one with all the little castellations on the sides because I wanted the I/O.  I took some perfboard and put some .1" headers on it and used resistor leads coming up angled inward slightly and soldered them to the cutouts.  The 8266 hovered over the perfboard after doing that but I could plug it into a breadboard at least.

[Stonent]

I checked all connections. No shorts.

That is the price you pay for incredibly cheap china hardware, that is something wich has to be expected.
If you pay 3 bucks for something that usually costs 30, there has to be a catch

Where is it sold for $30? Aside from it being nowhere, I don't accept the logic of your initial sentence.

I understood the statement as a similar device that has an MCU and wifi on one board.

[Frost]

Where is it sold for $30?

An Atmel SAM W25 should be somewhere in this region.

[PeterFW]

Where is it sold for $30? Aside from it being nowhere, I don't accept the logic of your initial sentence.

Something similar, comparable, same functionality.
Board, antenna, antenna connector, RF, stack, logic and so forth.

Last time i checked every other WiFi product was around that $30 price range in low/single quantity.

[Tuppe]

Where is it sold for $30? Aside from it being nowhere, I don't accept the logic of your initial sentence.

Something similar, comparable, same functionality.
Board, antenna, antenna connector, RF, stack, logic and so forth.

Last time i checked every other WiFi product was around that $30 price range in low/single quantity.

Close, but not quite. Last time I quickly checked (5 minutes ago)

WiFi SoCs 1 piece price:
MT7681 6€
AR9331 16€
CC3200 16€

Most of those don't come in header form, but whacking it on a standard breakout isn't hard to do.
Obviously ESP8266 has driven those prices down single-handledly.

[dannyf]

Seems to be most people have no problem getting those things to work, with some exceptions,

Obviously ESP8266 has driven those prices down single-handledly.

Yeah. If you go back just a little bit, those modules were considerably more expensive. $30 was a good number.

[Yansi]

Some people have problem staying off this forum for a single hour. 

[SL4P]

...? Howewer the whole OSHW stuff probably is. I havent seen any arduino instructable suggesting to connect any of the remaining gpios, have you?

You call yourself an engineer, but the extent of your understanding this chip is based on 'instructables'...LOL.

[Yansi]

And on what should I base my knowledge of this chip, if there is none documentation? Sorry, I cant do chinese nor have I willingness to learn it.

If you know, where full english documentation is, provide us with it. I think a whole bunch of people will become satisfied more than enough.

[dannyf]

Some people have problem staying off this forum for a single hour.

No need to despair. Your problem is fairly easy to solve: ask your parents to install parental control software on  your computer and wola, your problem gone and life is good again.

[PeterFW]

And on what should I base my knowledge of this chip, if there is none documentation? Sorry, I cant do chinese nor have I willingness to learn it.

I think you may be a good example of the modern school system wich values the skill of memorizing more then thinking, understanding and associating.

You bought something without anny reasarch, expected someone else to allready have done all work and now you are throwing a tantrum...

The second result for "ESP8266 pinout" gives you every information you have to know on how to get the board to respond to a ping.
Allthough the second hit may be google search bubble enhanced... personal experiance may vary.
http://www.electrodragon.com/w/ESP8266

Oh... my biggest problem was that just nothing worked, no blinky, no data, no nothing, only a constand 80mA current draw.
Took me a while to figure out that it was my power supply, i had set it to 150mA.
During bootup the board draws way more for a very, very short time.
So short that it dit not trip the over current protection but kept to board from booting properly.

[Yansi]

I think you may be a good example of the modern school system wich values the skill of memorizing more then thinking, understanding and associating.

Yes, example of what people like you have made the today's school systems look like. Not fault of my generation. I haven't built the school system. And please stop mentioning school here, because school gave me nothing, compared to what I've learnt myself. And I also do NOT agree with the state of our local school system. (Not only school, the whole system here is a pile of turd, recently I have read, that a crippled shit missunderstood by our dumb politicians in our law possibly makes a total ban of coin cell batteries in our country in October this year). I am looking forward for their next action.

You bought something without anny reasarch, expected someone else to allready have done all work and now you are throwing a tantrum...

If you buy for example TL074, would you expect that you will have to dig for for every parameter of it throughout the web? That someone measured its bias currents, someone else figured out the pinout, etc... or would you just expect downloading a full featuring datasheet, then read it and make the thing working using its specsheet?

Do you think, that expectation of proper documentation is fault of school education?    That was a good one...

No, this page wasn't so close to the top here. This one I have dug out later. Although it is a good one to start with, I'd not consider it as documentation. It is quite incomplete. or example, where can I list the WHOLE set of AT commands for specific FW version? The one listed there is quite short.

Still no explanation why my module boots at 75.7kbaud. But I hope someone will come and say I haven't done research or haven't read the documentation. (which one?)  :-D


Tommorow I am going to put the module on power analyser to see the real consuption (also in time). And will compare consuption on different module types.

For those wandering what I consider good documentation, see for example this nice GPRS module: http://source.sierrawireless.com/resources/#tags=AirPrime%7Cq2686 (Hope the link will bring you where I expect)

[PlainName]

If you buy for example TL074, would you expect that you will have to dig for for every parameter of it

Don't know about  them, but I would have checked what documentation is available and then figured whether it was worth my time to hack it or best look at something more professional.

Though having said 'professional', I can assure you (and I would have thought you would know if you are in the design business) that really quite serious kit can suffer from lack of documentation, or even misleading documentation.

Mind, I can remember when acquiring documentation meant phoning the manufacturer and then waiting for some books (if you're lucky) or leaflet to drop through the letterbox. The likes of Google have spoiled today's generation, where you can go from zero to apparent expert in a few mins of browsing. I keep wondering what's going to happen when the original sources of info - guys like the ones you don't want to listen to in this thread - are gone and replaced by only consumers of info - guys that rely on others having done the work for them, via instructables or whatever. Seems to me that Googling is going to start coming up short, and development skills will again become pretty useful.

Fortunately, I came late to this thread

[zapta]

The rants so far are about documentation (datasheet), firmware (e.g. AT commands and 1024 only SSL) but not about the ESP8266 itself, so that's a good sign because these are things that can be improved by the vendor and the community.

I would be more concerned if the ESP8266 itself would be faulty and unreliable.

[Yansi]

Mmmm no, I am not from design. Design is my everyday hobby, so far. Not enough knowledge yet, one must successively make his way to the design world.

Don't know about  them, but I would have checked what documentation is available and then figured whether it was worth my time to hack it or best look at something more professional

Thats true, I can't complain about that. It successfuly  completed our objective, which was mostly just to "test it" and see, where it gets you. I kinda haven't expected so much scattered documentation. I am used to use standard more-or-less of the shelf parts, which mostly have accessible complete documentation, which can be simply applied and the product used.

Maybe I overreacted a bit (it happens, I am bit of choleric person; luckily no hammer is never near me bcs I would widlarize stuff rather quickly),  after going through some tens of pages, one can learn anything he needs about the ESP. Then it is really no hard process to deal with the traps the module throws at you.

Ultimately, if it works, it works fine or fine enough that it is a good value for the money. Hopefuly I'll work out some handy usage for the module.


zapta: By mentioning the reliability.. tommorow I'll conduct a consuption analysis. I suspect the other module is somewhat damaged, it got hot rather quickly. It worked, but the operating temperature was interesting. My module I have here on the table doesn't get even near that hot when operating. It is only warm. I will share the results.

[Simon]

Some people have problem staying off this forum for a single hour.

No need to despair. Your problem is fairly easy to solve: ask your parents to install parental control software on  your computer and wola, your problem gone and life is good again.

If you don't have anything to contribute to the actual thread shut up and go away!

[zapta]

zapta: By mentioning the reliability.. tommorow I'll conduct a consuption analysis. I suspect the other module is somewhat damaged, it got hot rather quickly. It worked, but the operating temperature was interesting. My module I have here on the table doesn't get even near that hot when operating. It is only warm. I will share the results.

Do a visual inspection of the modules. There were some reports on poor workmanship such as crystals assembled in a wrong orientation.

My guts feeling is that the ESP8266 itself is a solid chip but we actually buy modules from small vendors that add the typical quality you expect from ebay/China. For example, the standard flash IC (that's second IC on the module) seems to have low endurance rating (number of good writes) and I had to replace a few but that not the ESP8266 itself.

Olimex now have their own OSH modules with Eagle files and, I think, better flash and workmanship so basically all we need is a good supply of the core ESP8266.

https://www.olimex.com/Products/IoT/MOD-WIFI-ESP8266/open-source-hardware

[dannyf]

I would have checked what documentation is available

Documentation doesn't seem to be a problem for those amaturer Arduino guys.

Only the "professionals" seem to struggle with the "lack of documentation" here.

[Simon]

I would have checked what documentation is available

Documentation doesn't seem to be a problem for those amaturer Arduino guys.

Only the "professionals" seem to struggle with the "lack of documentation" here.

And I now look forward to 2 weeks of peace from these idiotic posts!

[Yansi]

zapta: Will do a visual check. 

From what I can tell after one afternoon of experiments - I observed, that the chip probably has good performance radio.  The sensitivity with an external antena chopped of an old Asus router is surprising. Finds about 7 to 10 networks, compared to anything I have at home which can find 4 of them at best. For those far networks the ESP gives reading of under -80dBm, record is network at a level of -95dBm.

Has anyone measured the real output power? The claimed 300mW (+25dBm) is quite respectable power, considering that maximum 100mW EIRP is allowed here.

I also searched for bare chips, some time ago. I thing the can be found, but the prices were higher than the modules, if I remember that right. I should look again now. I don't like sticking any unnecessary modules to my boards. And custom implementation seems possible here, only the output LNA trace should have desired impedance of 50 ohm.

//seems OK now: http://www.aliexpress.com/item/Free-shipping-10pcs-lot-IC-CHIP-ESP8266-Serial-wireless-WIFI-chip-100-New-and-original-esp8266/32256889807.html

[zapta]

//seems OK now: http://www.aliexpress.com/item/Free-shipping-10pcs-lot-IC-CHIP-ESP8266-Serial-wireless-WIFI-chip-100-New-and-original-esp8266/32256889807.html

Good to know, I will order me a few. Since Olimex released the eagle layout I will just copy the RF section and will flatten my design.  Also ordered these ones, I think they are replacements for the flash ICs.

http://www.ebay.com/itm/251711518945?_trksid=p2057872.m2749.l2649&ssPageName=STRK%3AMEBIDX%3AIT

[zapta]

And I now look forward to 2 weeks of peace from these idiotic posts!

Not fair, Danny will have a life for two weeks while we are squabbling here.

[Yansi]

Bringing the measurements.  Unfortunately I forgot I've wanted  to measure also startup currents. The results look pretty interesting, I must say. Sometimes, the peak goes a bit over 300mA.

Also looks like the MCU sleeps and is periodically woken up.

[zapta]

The stock AT commands look good initially but when you get to the details they are hard to work with.

Correction, the ESP8266 AT commands work quit good. The trick is to disable echo ("ATE0", not documented) such that it doesn't interfere with "+IPD" RX data.