These are chat archives for esp8266/Arduino

14th
Jan 2016
0xPIT
@0xPIT
Jan 14 2016 00:00
this is my exception.
yes, ConfigFile.ino works.
Martin Ayotte
@martinayotte
Jan 14 2016 00:00
Oh ! ... maybe it is an issue with flush(), I don't use it. Try commenting it ...
0xPIT
@0xPIT
Jan 14 2016 00:01
nope.
if I comment the flush(), it will crash at close().
if I also commend close(); it will print the last message of the function and then crash.
very strange.
Martin Ayotte
@martinayotte
Jan 14 2016 00:04
So, it must be somewhere else in you code.
0xPIT
@0xPIT
Jan 14 2016 00:04
no. my code is fine.
Martin Ayotte
@martinayotte
Jan 14 2016 00:04
Strip ever other code to figure out
0xPIT
@0xPIT
Jan 14 2016 00:04
just found out: it works when i use it in setup()
i did that already, stripping it.
Martin Ayotte
@martinayotte
Jan 14 2016 00:05
and where it is when crashing ? for me, I'm doing those write() in loop() without problem
0xPIT
@0xPIT
Jan 14 2016 00:07
ok, so it works when I call my function saveParams() from setup() but when I call it from a serial port command handler, it does crash.
Martin Ayotte
@martinayotte
Jan 14 2016 00:09
which kind of serial handler ? did you try also in loop() ?
0xPIT
@0xPIT
Jan 14 2016 00:11
i'm trying to set a bool in the serial handler and call save() from loop()...
no. same problem when I call from loop().
Martin Ayotte
@martinayotte
Jan 14 2016 00:13
In my case, I'm doing SPIFFS.begin also in loop(), maybe this can help ?
0xPIT
@0xPIT
Jan 14 2016 00:16
ok. looking at the code I think thats not necessary. Plus, SPIFFS.begin checks and returns immediately if it has already been called.
I've tried it nevertheless, no effect.
Martin Ayotte
@martinayotte
Jan 14 2016 00:18
then, you will need to find where the crash happen using above exception address against objdump output
Stavros Korokithakis
@skorokithakis
Jan 14 2016 00:25
@tzapu man, i'm looking for the datasheet for HT7333 again and it's looking pretty shitty, only 250mA output... what was i looking at before?
@martinayotte which looks better to you? HT7333 or SPX3819M5-L-3-3?
0xPIT
@0xPIT
Jan 14 2016 00:31
i'll try, thanks.
Martin Ayotte
@martinayotte
Jan 14 2016 00:34
Welcome
@skorokithakis , The SPX look fine, question of 500mA and small SOT-23 package, but the HT7333 has lower Quiescent Current
Stavros Korokithakis
@skorokithakis
Jan 14 2016 00:37
@martinayotte that's the current the regulator itself consumes, right?
Martin Ayotte
@martinayotte
Jan 14 2016 00:37
Yes
Stavros Korokithakis
@skorokithakis
Jan 14 2016 00:37
hmm, most of the time i need a regulator it's for powering from a power plug
so it looks like the SPX is a better regulator
@martinayotte which is your favorite 3v3?
Martin Ayotte
@martinayotte
Jan 14 2016 00:38
But in fact, even if both have 10x order magnitude, it is FAR from the big Quiescent Current
of the AMS1117
Souper time ... I will be back in 40 mins
0xPIT
@0xPIT
Jan 14 2016 00:42
@martinayotte so you say the the epc1 address should match the exception when I objdump -D the ELF file??
ok, the exception is in my timer interrrupt. seems that disabling the interrupts does either not work or does not affect timer0.
Harrison Mclean
@h4rm0n1c
Jan 14 2016 00:46
Bugger, I just ordered a reel of 50 AMS1117, oh well, I did design a heatsink into the board for it...
Stavros Korokithakis
@skorokithakis
Jan 14 2016 00:46
@h4rm0n1c why did you order those?
they don't look that good
Chris Elsworth
@celsworth
Jan 14 2016 00:46
AMS1117 are cheap, tried, tested
Harrison Mclean
@h4rm0n1c
Jan 14 2016 00:46
Yeah, that
Chris Elsworth
@celsworth
Jan 14 2016 00:46
good enough for arduino to put on millions of boards :)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 00:49
what, the uno? :P
all the breakouts i've seen use SPX
for the ESP, i mean
Chris Elsworth
@celsworth
Jan 14 2016 00:52
nodemcu is 1117 right?
yep, looking at one now :)
Harrison Mclean
@h4rm0n1c
Jan 14 2016 00:53

Most of my stuff uses 5v, I like my through hole TTL parts still, the Woz in me won't let go.

So yeah, I figured dropping less than 2v wouldn't be an issue for the 1117

Stavros Korokithakis
@skorokithakis
Jan 14 2016 00:54
@celsworth which one?
actually i think i looked at the wemos d1 mini and the huzzah
both spx
Martin Ayotte
@martinayotte
Jan 14 2016 01:06
(Souper faster than expected)
@0xPIT , good to know that you find clues ! please provide details to help others if you find the guilty line
About AMS1117, if we don't care about Quiescent Current, it is good because of 1A. But I think I will order some SPX since they are so small ... :-)
Maybe my 50x AMS1117 will start crying in their drawer :-)
0xPIT
@0xPIT
Jan 14 2016 01:10
Hm, I've tried to manually disable interrupts, this does not work for Timer0.
if I use ETS_CCOMPARE0_ENABLE() / ETS_CCOMPARE0_DISABLE() the code stops. All other ways I could find to disable all interrupts left timer0 unaffected and active. I've tried ÌnterruptLock, plus I tried to modify the interrupt levels, andets_intr_lock()/ets_intr_unlock()`. non works.
thanks for the help @martinayotte :beer:
my workaround for now is to detach the timer ISR and reattach after the SPIFFS write is complete.
if anyone knows how to really disable intterrupts including timer0, I'd be happy to receive a tip. (twitter: @0xPIT) thanks.
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:20
@martinayotte i agree about the quiescent current
looks like the AMS1117 has 5V max input voltage, though?
Chris Elsworth
@celsworth
Jan 14 2016 01:22
no?
15
but it looks like they'd rather you didn't go above 12
Martin Ayotte
@martinayotte
Jan 14 2016 01:23
Yes, 15V
Chris Elsworth
@celsworth
Jan 14 2016 01:23
above 12 looks to be out of guaranteed regulation, but will still work
Martin Ayotte
@martinayotte
Jan 14 2016 01:23
But it would be absurd to put those voltage for efficiency ...
Chris Elsworth
@celsworth
Jan 14 2016 01:24
yeah I imagine it'll get pretty hot
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:24
aw
Chris Elsworth
@celsworth
Jan 14 2016 01:24
need a buck for that really?
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:24
really? i'm designing a board that will have 12V input
Chris Elsworth
@celsworth
Jan 14 2016 01:24
LDOs are not the way to do it
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:24
what should i use to regulate to 3.3v?
:(
what should i use then?
Chris Elsworth
@celsworth
Jan 14 2016 01:25
buck regulator, imo
but I will defer to @martinayotte :)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:25
aw, that's huge :/
i don't need 3A either, i only need to power the ESP from it
Harrison Mclean
@h4rm0n1c
Jan 14 2016 01:26
What did the ESPlux use, that's 12v
Martin Ayotte
@martinayotte
Jan 14 2016 01:26
no, they are about 1cm x 1.5 cm
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:27
@martinayotte that's almost as big as my entire pcb :/
i was making a 2.5x2.5cm pcb
what will the problem be with the LDO and 12V?
it's rated for up to 16
Chris Elsworth
@celsworth
Jan 14 2016 01:27
it will burn up, probably :)
yes, at very low currents
(like.. uA)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:28
oh :(
Martin Ayotte
@martinayotte
Jan 14 2016 01:28
I would like to find complete DC-DC inside an SOT-23, but it doesn't exist yet
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:28
yeah :/
Chris Elsworth
@celsworth
Jan 14 2016 01:29
@skorokithakis for the AMS1117, look at "Thermal Considerations" in the datasheet
Pd = (Vin - Vout) * Iout
so you can see immediately, the larger the voltage drop, the hotter its going to get for a given output current
Martin Ayotte
@martinayotte
Jan 14 2016 01:30
Yes, this is true, but ESP don't take the full 1A
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:30
yep :/
Chris Elsworth
@celsworth
Jan 14 2016 01:30
plug some numbers in, see how hot it'll get
all the info is there to work it out
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:32
how can i calculate temperature from power dissipation?
Chris Elsworth
@celsworth
Jan 14 2016 01:32
TJ = TA(MAX) + PD(Thermal Resistance (junction-to-ambient))
Martin Ayotte
@martinayotte
Jan 14 2016 01:32
@skorokithakis , but at first, why you input should be 12V ?
Chris Elsworth
@celsworth
Jan 14 2016 01:32
same section
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:32
@martinayotte that's what the LED strip operates at
@celsworth oh, right
Chris Elsworth
@celsworth
Jan 14 2016 01:32
lets assume 0.2A? so (12-3.3)*0.2 = 1.74W
Martin Ayotte
@martinayotte
Jan 14 2016 01:33
But you can add the small DC-DC in the same casing of your PSU then ...
Chris Elsworth
@celsworth
Jan 14 2016 01:33
ambient, lets say 20C? so TJ is 20 + 1.74*55 = 115C
thats using the best case thermal resistance from that table
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:33
@martinayotte the PSU is sealed, i can't open it :/
Martin Ayotte
@martinayotte
Jan 14 2016 01:33
Cooking eggs ...
Chris Elsworth
@celsworth
Jan 14 2016 01:33
and: Maximum junction temperature must not exceed 125°C.
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:33
ouch
Chris Elsworth
@celsworth
Jan 14 2016 01:34
115C is very close the borderline, assuming perfect thermal dissipation
I conclude 0.2A is too much to risk :)
for that drop, anyway
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:34
yeah :/
and i wouldn't want that to get to 115 degrees anyway
i can boil water with that
Chris Elsworth
@celsworth
Jan 14 2016 01:34
no, definitely not
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:34
bleh
that's too bad
Martin Ayotte
@martinayotte
Jan 14 2016 01:35
I would go for DC-DC
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:35
yeah, i may have to :/
Martin Ayotte
@martinayotte
Jan 14 2016 01:35
a small as postal stamp
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:36
that's as big as the entire rest of the circuit though :/
no choice, i guess
by the way, i'm guessing the same thing happens with the big TO-220 regulators?
because i used one of those to regulate 9v down to 3v3
Chris Elsworth
@celsworth
Jan 14 2016 01:37
would need to check their datasheets, but the fundamentals are the same
Harrison Mclean
@h4rm0n1c
Jan 14 2016 01:37
Yeah, linears.
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:37
hm right
Martin Ayotte
@martinayotte
Jan 14 2016 01:37
with TO220, it will be bigger than DC-DC
Heat-Sink, etc
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:38
@martinayotte yeah, but i already used one for something else and i'm wondering whether it'll catch on fire
Chris Elsworth
@celsworth
Jan 14 2016 01:38
well, feel it :)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:38
i'll do that when i go back home
Chris Elsworth
@celsworth
Jan 14 2016 01:38
if you can't put your finger on it, maybe time to check :)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:39
haha, yes
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:39
i'll buy those converters
@martinayotte how do you configure the output voltage?
Harrison Mclean
@h4rm0n1c
Jan 14 2016 01:40
Variable resistor and a voltmeter I'd guess.
Chris Elsworth
@celsworth
Jan 14 2016 01:41
yeah that could be a pot
Harrison Mclean
@h4rm0n1c
Jan 14 2016 01:41
Looks like a smt pot to me
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:42
great, tahnks
Harrison Mclean
@h4rm0n1c
Jan 14 2016 01:43

One of those had an SS34 diode on it, that's rated for 3A.

I'd feel uncomfortable pushing 3A through something that small...

*a board
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:43
yeah, doesn't look safe
Martin Ayotte
@martinayotte
Jan 14 2016 01:44
@skorokithakis , yes this is a small trim pot to adjust output voltage
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:44
@martinayotte great, bought! i can solder it on the bottom of my pcb
is there a similar problem passing 12V through a MOSFET with a 3V3 gate?
i guess no
by the way, does anyone know what the footprint difference between 0805 and 0805_handsoldering is (in kicad)?
Martin Ayotte
@martinayotte
Jan 14 2016 01:46
I don't understand the question : you MOSFET is still on the 12V side, and the gate is Vgs on 4V side ? so, nothing to change here
I use plain 0805, but maybe the other has bigger pads
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:47
yeah, i ordered one with handsoldering, and the pads are huge and farther out
i wonder how hard it would be to solder the non-handsoldering ones
Martin Ayotte
@martinayotte
Jan 14 2016 01:48
As I said, I use plain ones, and it quite easy to solder, they still been 0805, not like those 0402 :-)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:50
@martinayotte so you select the 0805 footprint? not 0805_handsolder?
Martin Ayotte
@martinayotte
Jan 14 2016 01:50
@celsworth , Atmel-vs-Microchip ? ... ouch !
I don't remember where I took those footprint, but they are really name 0805, not 0805_handsolder
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:52
i see, thanks
Martin Ayotte
@martinayotte
Jan 14 2016 01:52
Remember that I'm still use older Kicad, even if I've installed newest on another machine
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:52
must be the default kicad ones
ah yes
i just checked and the handsolder pads are just much, much longer
Martin Ayotte
@martinayotte
Jan 14 2016 01:53
Maybe it is for those who have parkingson ;-)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 01:53
hahah
so mean
Martin Ayotte
@martinayotte
Jan 14 2016 01:55
That make me remember my first iron solder when I had 7 year old : a Weller that really look like an gun45 (I should try to find picture)
BTW, it is now 45 years later ... ;-)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 02:00
haha, my dad still has one of those
his has a light bulb under the metal bits
Harrison Mclean
@h4rm0n1c
Jan 14 2016 02:01
The wiring inside is... clever, and kind of frightening...
Martin Ayotte
@martinayotte
Jan 14 2016 02:02
Yes, problably the guys who took that picture broke the buld and didn't replace it.
Stavros Korokithakis
@skorokithakis
Jan 14 2016 02:03
right
Martin Ayotte
@martinayotte
Jan 14 2016 02:03
This is a big transformer producing heat on the tip, the housing was becoming a huge magnet.
Harrison Mclean
@h4rm0n1c
Jan 14 2016 02:04

http://kingstonemodelworks.com/MRH/RSUb.jpg

No wonder china loves turning out copies of this... so simple, but I don't want to imagine the current on that secondary...

Martin Ayotte
@martinayotte
Jan 14 2016 02:04
I think I still have it in one of my junk boxes
Harrison Mclean
@h4rm0n1c
Jan 14 2016 02:04
Def a liability around magnetic storage, lol
Martin Ayotte
@martinayotte
Jan 14 2016 02:05
I don't care much for that, floppies are well obsolete too ;-)
The good thing with those guns, they were able to solder 10AWG wires
Not good for LQFP64 thought :-)
the routing is ugly as hell
but it's a full LED strip controller
in 25mm^2
Martin Ayotte
@martinayotte
Jan 14 2016 02:47
This is the addition of Power Barrel Connector ?
I would suggest to still kept an LDO on the board for 3.3V and the Barrel can be supply with 5V taken from the DC-DC
The 5V can be useful for something else.
Martin Ayotte
@martinayotte
Jan 14 2016 03:31
BTW, guys, I've received my first WeMos D1 Mini today. Looks very good for the price, $4 on AliExpress, USB-Serial with CH340. VReg from 5V USB, although I don't see big capacitors on PCB, ESP seems to be stable ! It provide RES switch, but there is no BOOT switch, so I presume I will have to place a wire on GPIO0 for my next upload ...
Chris Elsworth
@celsworth
Jan 14 2016 03:31
might be like nodemcu and controlled from DTR?
nodemcu has boot/reset switches but I found I never need to use them once esptool is configured correctly, all the necessary wires are hooked up to make it handsfree, quite nice
Martin Ayotte
@martinayotte
Jan 14 2016 03:36
Maybe ... will check tomorrow !
beikeland
@beikeland
Jan 14 2016 04:06
@martinayotte it flashes fine from arduino ide with nodemcu reset method. So reset and flash signals are hooked up to handshake signals on the ch340
My only issue with the D1 mini is that is seems to sometimes boot into flash mode instead of my application, needing a reset. Havent looked into why in any detail yet though
Martin Ayotte
@martinayotte
Jan 14 2016 04:12
@beikeland , yes ! I've just done my second upload, it worked. the plateforms.txt commands handled it properly ! I think I will re-iterate with new order of 2x boards. :-)
Martin Ayotte
@martinayotte
Jan 14 2016 04:23
Order re-iterate done ! :-)
beikeland
@beikeland
Jan 14 2016 04:52
can i test incomming http requests for If-Modified-Since headers? if serving static files from spiffs; it'd be okay to respond 304 to any such requests wouldn't it? Asuming i reboot the esp if i ever update the spiffs
bidyutper
@bidyutper
Jan 14 2016 04:59
@skorokithakis mqtt helpin me alot....nw I can control multiple mqtt clients over mqtt broker...& lie above tcp/ip protocol
beikeland
@beikeland
Jan 14 2016 05:04
server.hasHeader("If-Modified-Since") should get me some of the way; is file modification time available on the spiffs?
tzapu
@tzapu
Jan 14 2016 06:51
@skorokithakis yes, 250ma on ht7333. I plan some battery sensors as well, so it makes sense to have the lowest quiescent current one for me at least. Maybe you should have both plans on your github ? download this for ht7333, download this for the spx...
btw, i do have the ht7333 in "production". been on a battery operated sensor that waks every 10 minutes, reads ds18b20, posts to emon cms over http and goes to sleep and it s been running for around 6 months now, so the 250ma + cap seem fine for the esp
hmm, seems it s connecting to mqtt and sending a messages as well, so doing a bit of TX
i plan on using both your designs anyway :D
Mario Mikočević
@mozgy
Jan 14 2016 08:32
@skorokithakis I'm using those Mini-360 DC-DC and they're really good
so good that I don't plan to add power related circuit on my boards
beikeland
@beikeland
Jan 14 2016 08:33
I have a ton of them as well, my goto thing to get 5v down to 3.3v. sometimes just add an extra cap.
Mario Mikočević
@mozgy
Jan 14 2016 08:34
my only beef on them is that they're not good for deepsleep :(
waiting this https://www.pololu.com/product/2122 to became cheaper
tzapu
@tzapu
Jan 14 2016 09:24
that pololu is really impressive 'power-saving feature maintains high efficiency at low currents (quiescent current is less than 0.1 mA)'
Mario Mikočević
@mozgy
Jan 14 2016 09:26
on/off feature is what I want
atm I can't use LiPo as batt source 'cos it's too high voltage
fresh LiPo is 4.2V
tzapu
@tzapu
Jan 14 2016 09:28
well, i m using lipo with a ht7333, but obviously that won t work lower than 3v most likely
on/off/standby i think i saw that on the reg @skorokithakis was using. may be wrong
i think the reg still works when below 3.3 - drop voltage
and the esp also works below 3.3
lipo will get damaged at 2.7 maybe
don t recall all this...
also for lipo i got one of these http://www.ebay.com/itm/281652406065 but can t seem to find a way to keep the esp running when charging, because it doesn t cleanly terminate the charging
tzapu
@tzapu
Jan 14 2016 09:36
for what it s worth one of my esps, with ht7333 on one 18650 (4.2v when charged) seems to have worked down to 2.9V read by readVCC
Mario Mikočević
@mozgy
Jan 14 2016 09:37
I'm using LiFePO4 and that works up to 2.7V :)
tzapu
@tzapu
Jan 14 2016 09:38
so 2.9 after regulator, not much use for a boost covertor really...
Mario Mikočević
@mozgy
Jan 14 2016 09:38
or should I say 'down to' :)
tzapu
@tzapu
Jan 14 2016 09:39
this one probably does as well, it s an ecig battery with protection
again, if memory serves, they cut off at 2.7-2.8 i think
Helio Machado
@0x2b3bfa0
Jan 14 2016 10:17
Good morning!
tzapu
@tzapu
Jan 14 2016 10:18
good morning
Helio Machado
@0x2b3bfa0
Jan 14 2016 12:00
@/all: What would be the right sorting for these includes?
#include <ESP8266WiFi.h>
#include <ESP8266mDNS.h>
#include <WiFiUdp.h>
#include <ArduinoOTA.h>
#include <ESP8266WebServer.h>
#include <ArduinoJson.h>
#include <DNSServer.h>
#include <FS.h>
Markus
@Links2004
Jan 14 2016 12:10
@crushedice2000 right is every order that not give you an error.
the rest depends on personal code style.
Helio Machado
@0x2b3bfa0
Jan 14 2016 12:11
@Links2004: Mainly ascii art :smiling_imp: However I think that would be a method to determine the better order for they.
Harrison Mclean
@h4rm0n1c
Jan 14 2016 12:11
In the order of correct code dependency resolution, which will depend on what is defined where.
Chris Elsworth
@celsworth
Jan 14 2016 12:35
in an ideal world, each .h would include all its own dependencies and would have an #ifndef guard, so any order would work ;)
Markus
@Links2004
Jan 14 2016 12:37
the biggest problem is when you get a include dead lock.
Stavros Korokithakis
@skorokithakis
Jan 14 2016 12:44
so i j ust soldered the entire 0603 in my newly-arrived SMD practice board
and then realized that I was soldering 0805 components on the 0603 row
and wondering why the pads were small
@martinayotte that's a good idea, maybe i'll switch and use the 5V in the barrel jack, i'm going for small though, so i'll have to see
@martinayotte i also got two D1 minis, they look great so far, good form factor
@tzapu that's a good idea, but i don't want to maintain two versions every time i make a change :/
Mario Mikočević
@mozgy
Jan 14 2016 13:00
@skorokithakis go go go ! nice practice :D
Stavros Korokithakis
@skorokithakis
Jan 14 2016 13:02
not great, you should see the components :P
Chris Elsworth
@celsworth
Jan 14 2016 13:03
that doesn't look too bad
got tweezers to straighten them out?
Stavros Korokithakis
@skorokithakis
Jan 14 2016 13:04
those with the 102 text are 1206 type
Chris Elsworth
@celsworth
Jan 14 2016 13:04
if you get one side done while its held with tweezers and kept straight, the other side is then easy :)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 13:04
those with 821 are 0805
so i soldered each thing on its smaller component row
Chris Elsworth
@celsworth
Jan 14 2016 13:05
oops :)
Stavros Korokithakis
@skorokithakis
Jan 14 2016 13:05
soldered the 0805 on the 0603 and the 1206 on the 0805
Helio Machado
@0x2b3bfa0
Jan 14 2016 13:07
Only a small question: How can I limit my ESP access point to one client?
Chris Elsworth
@celsworth
Jan 14 2016 13:26
meh. wish the esp-01 and all its deriatives would die tbh
its only useful for AT, the 8266 can do so much more :D
Stavros Korokithakis
@skorokithakis
Jan 14 2016 13:40
you mean the 12E?
oh in general
tzapu
@tzapu
Jan 14 2016 13:48
@mozgy i keep looking at that version and at the breadboard adapter they do, and always come back thinking it s noce, too bad it s so expensive :P
Martin Ayotte
@martinayotte
Jan 14 2016 13:56
Yes, too expensive !
Chris Elsworth
@celsworth
Jan 14 2016 13:56
bare esp12f modules, sorted :)
doesn't get much cheaper, I think mine were like $2.50 each, got a stack of em
Stavros Korokithakis
@skorokithakis
Jan 14 2016 13:57
what's good about the f?
i just bought five 12Es last night :(
Chris Elsworth
@celsworth
Jan 14 2016 13:58

prepare for chinglish:

ESP8266-12 is an enhanced version of the ESP8266-12F, improve the peripheral circuit, the four laminates plate process, enhanced impedance matching, signal output is better, whether it is stable or anti jamming ability, PCB antenna after professional laboratory testing, perfect matching, after ROHS certification, have been greatly improved!

Stavros Korokithakis
@skorokithakis
Jan 14 2016 13:58
aw
Chris Elsworth
@celsworth
Jan 14 2016 13:58
12F isn't new ,its been out a while
Stavros Korokithakis
@skorokithakis
Jan 14 2016 13:58
i knew the one on the WeMos looked different from the E
the antenna is different
Chris Elsworth
@celsworth
Jan 14 2016 13:58
realistically speaking, you probably wouldn't even notice the differences, so don't worry too much about the E
yep
easiest way to tell, shape of the antenna
Stavros Korokithakis
@skorokithakis
Jan 14 2016 13:58
yeah, but if i'm buying, should have gotten the Fs
oh well
this button pcb is pretty handy
it's basically an amazon dash button
i made it play a random show on tv when i press it
Chris Elsworth
@celsworth
Jan 14 2016 13:59
nice :D
tzapu
@tzapu
Jan 14 2016 14:00
i still think the bring bacon use is the winner..
Stavros Korokithakis
@skorokithakis
Jan 14 2016 14:04
haha
i'll use it for that today
ystrem
@ystrem
Jan 14 2016 14:15
@mozgy very nice but pricy :D
Helio Machado
@0x2b3bfa0
Jan 14 2016 14:16
@mozgy @ystrem: Very pricey but nice :smiling_imp:. Is there a cheaper variant?
tzapu
@tzapu
Jan 14 2016 14:25
not that i found, would love an esp01 form factor like that with adc and/or sleep :D
like this one
only at a third up to a half of the price tops
beikeland
@beikeland
Jan 14 2016 14:28
but as long as it doesn't fit on a bradboard why is it better than any of the others?
Chris Elsworth
@celsworth
Jan 14 2016 14:30
@skorokithakis speaking of smd soldering, my smd rework hot air station just arrived :D I will have to try my smd practise board with it too
Stavros Korokithakis
@skorokithakis
Jan 14 2016 14:31
@celsworth ooh, photo?
tzapu
@tzapu
Jan 14 2016 14:33
@beikeland for a few projects i don t need all the pins and hassle of setting up an esp12, esp-01 at least is more ... plug and play.. :) for me anyway
yeah @celsworth photos :D
Chris Elsworth
@celsworth
Jan 14 2016 14:39
when I've reworked something with it :)
it came with a neat smd sucker part picker
tzapu
@tzapu
Jan 14 2016 14:40
wouldn t mind seeing the station either :P
Helio Machado
@0x2b3bfa0
Jan 14 2016 14:41
@celsworth: Congratulations!
beikeland
@beikeland
Jan 14 2016 14:46
Speaking of rework; you guy got tips for removing the rf shields? I tried removing on one of my nodemcu's. The esp module came off before the rf shield came off :P
Mario Mikočević
@mozgy
Jan 14 2016 14:59
@beikeland bigger hammer ?
beikeland
@beikeland
Jan 14 2016 14:59
@mozgy i'm not an mechanical engineer, i only have small hammers ;)
Mario Mikočević
@mozgy
Jan 14 2016 15:00
;)
beikeland
@beikeland
Jan 14 2016 15:01
we actually have a saying what separates the "fine" from the "coarse" mechanincs .. size of sledgehammer :D
Me No Dev
@me-no-dev
Jan 14 2016 15:08
@beikeland I used the same "effect" to replace the ESP module on my NodeMCU with QIO one :)
patience gets the top off after
Anyone here have a clue how to get a crosstool-ng build env on Windows? (cygwin)
Helio Machado
@0x2b3bfa0
Jan 14 2016 15:12
@me-no-dev: Sorry, all but W$. However there are many users here. :worried:
@beikeland: I removed the RF shielding of a ESP-12E at ~420 celsius degrees with the soldering station.
Me No Dev
@me-no-dev
Jan 14 2016 15:15
@crushedice2000 yeah I'm all but MS as well. Looked all over the net and there is nothing but prebuild toolchains and instructions for linux and osx
beikeland
@beikeland
Jan 14 2016 15:16
Do they use the same solder for the shielding, or is it just that effective at dissapating the heat? I did get it off, but suprised the whole module came off first
Helio Machado
@0x2b3bfa0
Jan 14 2016 15:18
@beikeland: My module remained on its position when removing the shield. Later I removed the module.
beikeland
@beikeland
Jan 14 2016 15:20
I used hot air, and module came off first, despite focusing air on the shield. later turned the temp up from 320 to 380 and shield came off
@beikeland: I started at 420 and with a 1cm nozzle. The shielding came off after ~12s
@both: Now I'll travel. I won't be "turned on" again up to tomorrow. Sorry :(
Markus
@Links2004
Jan 14 2016 15:41
@beikeland i use a gas burner short hold it on the metal and then use centrifugal force (Slapping on the table) and the metal is removed.
beikeland
@beikeland
Jan 14 2016 15:53
so the lesson i'm learning is crank up the temperature and get it off before anything else comes off. thanks!
FWeinb
@FWeinb
Jan 14 2016 18:39
Is there a lightweight templating language for embedded webservers, would be great to have such a lib on the ESP?
xbary
@xbary
Jan 14 2016 19:08
hi
I have a problem with loop to "LmacRxBlk: 1" for too many requests from the browser.
xbary
@xbary
Jan 14 2016 19:15
is there any way to get out of this loop is not resetting ESP?
Michael Miller
@Makuna
Jan 14 2016 22:32

I am finally back on my big project since August, and it doesn't compile anymore.

C:\Users\Michael\AppData\Local\Arduino15\packages\esp8266\hardware\esp8266\2.0.0\cores\esp8266/pgmspace.h:18:51: error: __c causes a section type conflict with __c

 #define PSTR(s) (__extension__({static const char __c[] PROGMEM = (s); &__c[0];}))

                                                   ^

C:\Users\Michael\AppData\Local\Arduino15\packages\esp8266\hardware\esp8266\2.0.0\cores\esp8266/WString.h:38:76: note: in definition of macro 'FPSTR'

 #define FPSTR(pstr_pointer) (reinterpret_cast<const __FlashStringHelper *>(pstr_pointer))

                                                                            ^

C:\Users\Michael\AppData\Local\Arduino15\packages\esp8266\hardware\esp8266\2.0.0\cores\esp8266/WString.h:39:34: note: in expansion of macro 'PSTR'

 #define F(string_literal) (FPSTR(PSTR(string_literal)))

                                  ^

C:\Users\Michael\Documents\Arduino\Esp8266CoapServer\Esp8266CoapServer.ino:328:20: note: in expansion of macro 'F'

     Serial.println(F("we be running"));

                    ^

now the strange thing is, if you build a simple app with just the serial.println it compiles and runs fine.

Has the size of the .irom.text changed dramatically? I have a ton of constant strings, but pretty sure its less than 4K.
Markus
@Links2004
Jan 14 2016 22:35
this not look like a .irom.text problem, the error is during compiling not lilking
pgmspace.h:18:51: error: __c causes a section type conflict with __c
Michael Miller
@Makuna
Jan 14 2016 22:36
ok, but how could other code effect this one line other than consuming the memory in that section?
Markus
@Links2004
Jan 14 2016 22:37
have you a variable named __c ?
Michael Miller
@Makuna
Jan 14 2016 22:38
nope, I never use (double underscore) as generally its reserved for system level features.
Markus
@Links2004
Jan 14 2016 22:39
ok have you put some code in the .irom.text section?
define ICACHE_RODATA_ATTR
Michael Miller
@Makuna
Jan 14 2016 22:43
I haven't, the only other use I can see is in osap.h
#ifdef USE_OPTIMIZE_PRINTF
#define os_printf(fmt, ...) do {    \
    static const char flash_str[] ICACHE_RODATA_ATTR __attribute__((aligned(4))) = fmt;    \
    os_printf_plus(flash_str, ##__VA_ARGS__);    \
    } while(0)
#else
extern int os_printf_plus(const char * format, ...) __attribute__ ((format (printf, 1, 2)));
#define os_printf    os_printf_plus
#endif
Markus
@Links2004
Jan 14 2016 22:44
its a variable too, so no conflict there.
Michael Miller
@Makuna
Jan 14 2016 22:45
p.s. I even changed the name in the macro to __cUniqueAsShit and it still had a problem.
Markus
@Links2004
Jan 14 2016 22:45
section type conflict sounds like that there is something in the .irom.text what not shut be there like a function.
Michael Miller
@Makuna
Jan 14 2016 22:57
I think its related to the preprocessor. I noticed since then some other things that don't work like they used to, like in the INO, you could have a function at the bottom but it would still find it without having a prototype above the call site. This is no longer true.
Markus
@Links2004
Jan 14 2016 22:58
yes may some arduino ide garbage, some where in the temp dir is the combined file that the compiler get to see.
Michael Miller
@Makuna
Jan 14 2016 22:59
I cleared the temp dir thinking some partial compile could be it, that didn't help.
Markus
@Links2004
Jan 14 2016 23:01
may its linked too #1194 and #1066
in eclipse the F macro usage works fine.
in arduino only init serial and use F macro in print working too.
so there is some interaction with your code.
Michael Miller
@Makuna
Jan 14 2016 23:10
Yeah, a simple app with the same lines works fine, this was what was leading me to think I may have ran out of flash memory and get this strange error rather than an error that you ran past the memory limit.
Most of my code is in a Coap library, with barest in the ino (and h) that define the constants.
Markus
@Links2004
Jan 14 2016 23:11
the memory is checked during linking not during compile, I am 99.99999% sure its not this.
Michael Miller
@Makuna
Jan 14 2016 23:13
Any thoughts on what could then effect this? I spotted one mention of this error (and its not related to esp) that compiler changed and required an extra ",r" after the .irom.text to allow it to be used with consts.
Markus
@Links2004
Jan 14 2016 23:14
since it working without in a "normal" sketch i think that is not our problem,
but you can try it.
Michael Miller
@Makuna
Jan 14 2016 23:21
I tried the ,r, didn't help. this project compiled back in August, so something new since then is causing an issue, which may still be mine, but knowing of anything changes may help lead me to the problem.
My divide and conquer method has so far led me to that a library that uses irom data is the issue.