WARNING: Potential ESP brick issue with deepsleep!!

Moderators: grovkillen, Stuntteam, TD-er

Post Reply
Message
Author
Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#11 Post by Drum » 01 Jun 2016, 13:15

I have been using Deep Sleep on a 10 minute cycle on 2 devices since February, no problems. Yet...

tozett
Normal user
Posts: 734
Joined: 22 Dec 2015, 15:46
Location: Germany

Re: WARNING: Potential ESP brick issue with deepsleep!!

#12 Post by tozett » 01 Jun 2016, 13:29

+1 thanks for updating and ''DE-warning''

shuttle2k
New user
Posts: 1
Joined: 20 Jun 2016, 11:10

Re: WARNING: Potential ESP brick issue with deepsleep!!

#13 Post by shuttle2k » 20 Jun 2016, 11:15

Hi,

has the issue already solved or still in there?

Thanks

sebnania
New user
Posts: 1
Joined: 17 Oct 2016, 09:59

Re: WARNING: Potential ESP brick issue with deepsleep!!

#14 Post by sebnania » 17 Oct 2016, 10:11

hi all,
have you got some new informations about this issue?

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#15 Post by Drum » 17 Oct 2016, 14:54

Still have 2 ESP8266 running on Deep Sleep waking up every 10 minutes, since February.
This problem has not come up for me yet.

DeNB3rt
Normal user
Posts: 120
Joined: 15 Dec 2015, 14:07

Re: WARNING: Potential ESP brick issue with deepsleep!!

#16 Post by DeNB3rt » 17 Nov 2016, 15:29

so no deepsleep issue for >1min boot? Want to try it for my weather station :)
Located in Belgium, Bruges. Working on a full DIY domoticz setup with ESPEasy.

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#17 Post by Drum » 20 Nov 2016, 06:44

Still have the same 2 units running with no problems. They are set for 10 minutes deep sleep.

lunix80
New user
Posts: 1
Joined: 06 Dec 2016, 17:40

Re: WARNING: Potential ESP brick issue with deepsleep!!

#18 Post by lunix80 » 06 Dec 2016, 17:43

Hi,

I have two esp-01 bricked by using deep-sleep mode.
They sleep for 5min, send temperature to mqtt broker then go to sleep again. They runned for a couple of days, than nothing.

I don't know if the battery voltage going low was the problem or the sleep-reset mode.

Anyone?

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#19 Post by Drum » 07 Dec 2016, 06:33

I do not use any esp-01 modules, not enough exposed pins and you have to solder a wire to the chip just for deep sleep to work. Closer examination of ,micro soldering job would be my first step.....

When you say "bricked" what have you tried to recover? What release are you using? Schemstic? Picture? More information is needed to even think about it. Nothing you mention should "brick" the device. What voltage regulation are you using?

By the way, is there a reason you are using esp-01s? I have noticed several people whoo seem to be locked in to the 01s when more flexible versions are available. Just curious....

NightBird
Normal user
Posts: 24
Joined: 02 Dec 2016, 17:54

Re: WARNING: Potential ESP brick issue with deepsleep!!

#20 Post by NightBird » 12 Dec 2016, 01:08

Does it mean the ESP write something in flash memory at every reboot ??? :o

As I understand deep sleep, the chip is reset by internal timer, just like a "classical" reboot ?
ESPEasy running on ESP-12F self-made modules (with various sensors) and talking to Domoticz (hosted on RPi2)

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#21 Post by Drum » 19 Dec 2016, 12:31

That would be my understanding.
I went back through the thread and realized my test-bed for this is sleeping too long to be a useful test.
Once I get things unpacked after my move, I will re-configure for a shorter sleep time to better see how it works.

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#22 Post by Drum » 17 Feb 2017, 11:03

I reconfigured my 2 test setups a couple of weeks ago to sleep for 1 minute. I checked this morning and 1 is running and posting like clockwork the other is not, but I expect it is a power issue as it is running on a LiPo / solar cell rig and not getting much sun these days. I'll check it see what is going on.

In any case I should be well over the 100,000 limit by June.

Just to update, both are running and posting to a sqlite database every minute now. It is only getting a couple of hours of sun a day (and most days are cloudy) which is not enough to keep it running at 1 minute deep sleep.. 10 minutes sleep it runs for a good 2 weeks with no sun. The second one has a larger solar panel so it seems to do better with the limited sun during the Italian winter.

I need to add the rule to post the IP address on connection to make this easier to track.

vojtishek
Normal user
Posts: 70
Joined: 12 Jan 2017, 08:00

Re: WARNING: Potential ESP brick issue with deepsleep!!

#23 Post by vojtishek » 10 May 2017, 09:13

Hi, I just running wemos d1mini with dht shield and as it's heating and increasing temperature of sensor then I decide to use deep sleep to decrease chip heat.
I've used 60 seconds interval without any trouble in deep sleep for appr. 24 hours, then It's stucked. At the moment every time I enable deep sleep, then stucked after appr.5-10min.
I'm using power adapter and wemos is just next to wifi router.

I've tried to change interval on 5 and 10 min but still same issue. Sensor delay is same like in device sensor setting delay.

Is there something else what I can do with that?

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#24 Post by Drum » 10 May 2017, 12:27

I would set the weblog level to 4 and check the logs to see what is going on. What version are you using and how is it configured? For the most part I am using 144 to 148, I have not converted to the new scheme yet so I really can't offer any help with those

vojtishek
Normal user
Posts: 70
Joined: 12 Jan 2017, 08:00

Re: WARNING: Potential ESP brick issue with deepsleep!!

#25 Post by vojtishek » 10 May 2017, 15:10

I´m using firmware image R147_RC8.
I will try to get some logs and come back with results.

vojtishek
Normal user
Posts: 70
Joined: 12 Jan 2017, 08:00

Re: WARNING: Potential ESP brick issue with deepsleep!!

#26 Post by vojtishek » 11 May 2017, 08:00

So It looks there was a problem with power adaptor, which I´ve used, cheap one from aliexpress.
Most probably some drop of voltage during regular wake ups.
Now I' m using another one from old cellphone and 12 hours without any problem, so let's see.

Stupid question, but where I will find the weblogs, in domoticz or somewhere in espeasy?

User avatar
grovkillen
Core team member
Posts: 3621
Joined: 19 Jan 2017, 12:56
Location: Hudiksvall, Sweden
Contact:

Re: WARNING: Potential ESP brick issue with deepsleep!!

#27 Post by grovkillen » 11 May 2017, 08:17

vojtishek wrote: 11 May 2017, 08:00 So It looks there was a problem with power adaptor, which I´ve used, cheap one from aliexpress.
If you use USB power adaptor:

I recommend anyone buy this type of voltmeter to test power adaptors (or power banks). It is really great to find out how much juice the unit is given and also at what ampere the power adaptor starts to choke. To test at what amp/current the power adaptor starts to give in I use this unit with variable resistor level. To test a powerbanks mAh I use a simpler one that have 3 levels of resistors and is not as likely to overheat. You will be struck by the fact that the power adaptors are not a constant 5.0V level. I got adaptors that give anything from 4.7V to 5.4V.

Another kit that I use is this one to test my DIY usb cables. So I really got two voltmeters to test the cables upstream and downstream. This is another type of kit that I use a lot.
ESP Easy Flasher [flash tool and wifi setup at flash time]
ESP Easy Webdumper [easy screendumping of your units]
ESP Easy Netscan [find units]
Official shop: https://firstbyte.shop/
Sponsor ESP Easy, we need you :idea: :idea: :idea:

vojtishek
Normal user
Posts: 70
Joined: 12 Jan 2017, 08:00

Re: WARNING: Potential ESP brick issue with deepsleep!!

#28 Post by vojtishek » 11 May 2017, 09:36

grovkillen wrote: 11 May 2017, 08:17
vojtishek wrote: 11 May 2017, 08:00 So It looks there was a problem with power adaptor, which I´ve used, cheap one from aliexpress.
If you use USB power adaptor:

I recommend anyone buy
Wow, thanks for that. Just added into my cart, definitely something what missed in my accessories.
Just wondering, what exactly happened in cheap power adaptor, because wemos was still visible in my wifi network, but not sending updates.
If there are any drops in voltage which causing fail of sending updates, then how to monitor them. Do you have any idea how to get voltage logs from some kind of measure device?

Shardan
Normal user
Posts: 1156
Joined: 03 Sep 2016, 23:27
Location: Bielefeld / Germany

Re: WARNING: Potential ESP brick issue with deepsleep!!

#29 Post by Shardan » 11 May 2017, 10:38

grovkillen wrote: 11 May 2017, 08:17 I got adaptors that give anything from 4.7V to 5.4V.
For ESP devices that are driven by 5 V a range from 4,7...5,4 V is completely OK - if there are real 4,7..5,4V!!
They usually have a LM1117-3.3 or similiar voltage regulator that gets voltage down to 3,3V. They will work
with 5,4 V as good as with 4,7V. For those devices directly running on 5 V.....
well, for a Display it does not matter, for a PIR sensor neither.

The problem is on another note, it is not the voltage itself.

If you use a digital voltmeter for checking power consumption of an ESP, you will get an average
current of 50..70 mA depending ont he type of ESP. If you use a good Voltmeter, you may notice
that the value "jumps" now and then to a higher value. But most digivoltmeters are too laggy, it
won't show the true peak value.
If you use an oscilloscope or a digivoltmeter capable of measuring peak values for measuring the
current you will get other values. A usual ESP-12E in my test circuit takes up an average of ~65 mA.
Peaks can go up to 350...400mA for a short moment when WiFi activities fire up.

The "chinaware" power supplies often can't handle these peaks. It needs some big capacitors
to handle and those are expensive. This is contradictionary to the cheap price those items are sold for.
If those power supplies get such a current peak from an ESP, the voltage breaks down for a short moment.
It might even run under the voltage level the regulator on the WeMos or nodeMCU board can handle, this
bringing down the ESP to crash.

A standard voltmeter or USB-Checker will not show these breakdowns in voltage, they are too laggy.

Another issue with cheap power supplies: They tend to oscillate.
You put a voltmeter on the output and see some voltage. Let's say 5V for an example.
You put an oscilloscope on and you see: The voltage oscillates fast between 4V and 6V.
I've seen oscillation frequencies between some KHz up to some MHz!
Such oscillations are too fast for an usual voltmeter, it just shows the average value.
In most cases this points to the manufacturer left off some blocking capacitors to save money.
Anyways it points to a strange behaviour of your ESP. It does not boot up and you have to
reset several times? Crashes completely unpredictable? This might be the reason.

So my advice would be:
Don't buy cheap power supplies and an USB voltage checker on top to find out your new power supplies are classy scrap.
Save your money for all this and get real power supplies. It saves a lot of headaches and on long turn it saves money too.

A really good 5V / 1A supply comes around 10..15€.
It's worth it! It has those capacitors and other parts that make a reliable power supply.

Even more: This is the part that isolates you from mains voltage when handling your ESP.
Saving Money by buying cheap power supplies can get expensive.......

Regards
Shardan
Regards
Shardan

User avatar
grovkillen
Core team member
Posts: 3621
Joined: 19 Jan 2017, 12:56
Location: Hudiksvall, Sweden
Contact:

Re: WARNING: Potential ESP brick issue with deepsleep!!

#30 Post by grovkillen » 11 May 2017, 11:27

Good points Shardan.
ESP Easy Flasher [flash tool and wifi setup at flash time]
ESP Easy Webdumper [easy screendumping of your units]
ESP Easy Netscan [find units]
Official shop: https://firstbyte.shop/
Sponsor ESP Easy, we need you :idea: :idea: :idea:

Shardan
Normal user
Posts: 1156
Joined: 03 Sep 2016, 23:27
Location: Bielefeld / Germany

Re: WARNING: Potential ESP brick issue with deepsleep!!

#31 Post by Shardan » 11 May 2017, 11:36

Thanks :)

If i get such a power supply on my desk again i'll do some screenshots from the oscilloscope
and make an extra posting in the Wiki... might be helpfull.

Regards
Shardan
Last edited by Shardan on 11 May 2017, 11:37, edited 1 time in total.
Regards
Shardan

vojtishek
Normal user
Posts: 70
Joined: 12 Jan 2017, 08:00

Re: WARNING: Potential ESP brick issue with deepsleep!!

#32 Post by vojtishek » 11 May 2017, 11:37

Thanks Shardan, completely make a sense.
Still thinking about distribution of power for sensors in ceiling through whole house with one reliable source of voltage.

Shardan
Normal user
Posts: 1156
Joined: 03 Sep 2016, 23:27
Location: Bielefeld / Germany

Re: WARNING: Potential ESP brick issue with deepsleep!!

#33 Post by Shardan » 11 May 2017, 11:39

vojtishek wrote: 11 May 2017, 11:37 Thanks Shardan, completely make a sense.
Still thinking about distribution of power for sensors in ceiling through whole house with one reliable source of voltage.
A central source has it's advantages and disatvantages.
But we should discuss that in a new thread, it's getting somewhat off-topic here :)
Regards
Shardan

vojtishek
Normal user
Posts: 70
Joined: 12 Jan 2017, 08:00

Re: WARNING: Potential ESP brick issue with deepsleep!!

#34 Post by vojtishek » 11 May 2017, 11:41

Yes, definitely. Electricity in my new house is in plan during October so I will make new topic until that period :)

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#35 Post by Drum » 11 May 2017, 16:05

Just to get back on topic, the 2 test units are still running strong, although the small solar cells are not keeping the batteries charged with 1 minute deep sleep so there have been some gaps. I will have to see what I can do about that next week. I need to do some work in nodered to set up some kind of notification on the voltage levels and if they have not connected in more than a few minutes.

I have also noticed they can get stuck when the voltage is low and they stay until the power is cycled. Just charging the battery is not enough, and any current from the solar cell seems to keep it in this state, although I have not tried resetting it yet. So when working with solar power, a voltage regulator which can shutoff power completely when the power is too low is probably a good idea for better reliability.

Shardan, any recommendations on an inexpensive oscilloscope which will work for these and not take up much space? Mostly I expect I would use it for looking at power on the ESPs and maybe some sensor communications. I try with ina219s but they are too slow for measuring mA the esp uses.

Shardan
Normal user
Posts: 1156
Joined: 03 Sep 2016, 23:27
Location: Bielefeld / Germany

Re: WARNING: Potential ESP brick issue with deepsleep!!

#36 Post by Shardan » 11 May 2017, 20:00

Drum wrote: 11 May 2017, 16:05 Shardan, any recommendations on an inexpensive oscilloscope which will work for these and not take up much space? Mostly I expect I would use it for looking at power on the ESPs and maybe some sensor communications. I try with ina219s but they are too slow for measuring mA the esp uses.
I'm using a Peaktech 1255 which is not really a cheap one (around 750€)
If you don't need a storage oscilloscope with a lot of gadgets i'd try to get a used Hameg 203-7 or similiar.
There are several offers on Ebay for those.

Digital scopes are somewhat expensive. Watch out, they should have 0,5..1 GS/sec. (Giga Samples)
Examples may be the GwInstek GDS-1072A-U https://www.reichelt.de/Oscilloscopes-S ... ARCH=%252A
or a UNI-T UTD 2102 CEX https://www.reichelt.de/Oscilloscopes-S ... ARCH=%252A.
Both around 450€.

Beware: Good instruments are never inexpensive..... but bad instruments are even more expensive on the long term.
2nd beware: Calculate about 10..20 hours to learn the "HowTo's".......

Cheaper ones with lower bandwith and (more important!) lower sampling rate are not a really good choice as they
might fail with quick changing values or high frequencies.

Regards
Shardan
Regards
Shardan

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#37 Post by Drum » 14 Jul 2017, 14:58

I have one unit which is not well over 100,000 deep-sleep restarts. I have 100,000 restarts documented today. I need to look to see if I can get through the rest of the data before I added a rule to post the IP on re-boot to get a better count. No problems, but I will keep it going until I see problems.

I have found a potential issue when using solar panels, but I will put that in a new post.

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#38 Post by Drum » 12 Oct 2017, 00:55

Just an update,
Unit 1 over 173,000 restarts
Unit 2 Over 219,000 restarts

Unit 2 is still plugging away, but Unit 1 is giving me errors, but I think something may be corrupted and I need to figure out how I fixed this the last time it happened.....

User avatar
ThomasB
Normal user
Posts: 1064
Joined: 17 Jun 2018, 20:41
Location: USA

Re: WARNING: Potential ESP brick issue with deepsleep!!

#39 Post by ThomasB » 20 Dec 2019, 19:39

Two years later: It appears this is an old problem that has been solved. I suggest it be removed from the pinned topics.

User avatar
grovkillen
Core team member
Posts: 3621
Joined: 19 Jan 2017, 12:56
Location: Hudiksvall, Sweden
Contact:

Re: WARNING: Potential ESP brick issue with deepsleep!!

#40 Post by grovkillen » 20 Dec 2019, 19:50

ThomasB wrote: 20 Dec 2019, 19:39 Two years later: It appears this is an old problem that has been solved. I suggest it be removed from the pinned topics.
Done
ESP Easy Flasher [flash tool and wifi setup at flash time]
ESP Easy Webdumper [easy screendumping of your units]
ESP Easy Netscan [find units]
Official shop: https://firstbyte.shop/
Sponsor ESP Easy, we need you :idea: :idea: :idea:

Drum
Normal user
Posts: 300
Joined: 07 Feb 2016, 11:56

Re: WARNING: Potential ESP brick issue with deepsleep!!

#41 Post by Drum » 20 Dec 2019, 19:56

By the way 1 is still working, although with the short days and lack of sun is not running today (would be better if it changed it back to 10 minute deep sleep).
The other one, I did something (can not remember exactly what I did) but I think it just needs a reflash, but just have not tried yet.

Now my problem is deciding if I want to try to save just under 9 MILLION lines of data in the sqlite database tracking everything from my ESPs.
I did upgrade it to a SSD recently but it is still a lot for a Raspberry Pi3 to sort through.
The thing is running Mosquitto, Node-red, Sqlite3 and Pi-hole with no problems ;)

User avatar
grovkillen
Core team member
Posts: 3621
Joined: 19 Jan 2017, 12:56
Location: Hudiksvall, Sweden
Contact:

Re: WARNING: Potential ESP brick issue with deepsleep!!

#42 Post by grovkillen » 20 Dec 2019, 20:02

:o :) thanks for the update.
ESP Easy Flasher [flash tool and wifi setup at flash time]
ESP Easy Webdumper [easy screendumping of your units]
ESP Easy Netscan [find units]
Official shop: https://firstbyte.shop/
Sponsor ESP Easy, we need you :idea: :idea: :idea:

Post Reply

Who is online

Users browsing this forum: No registered users and 28 guests