Let's Control It

I highly suspect the WEMOS D1 (clone) voltage regulator to be part of the reboot problem.
This low-dropout or LDO regulator has SMD marking code 4A2D which leads to the Torex XC6204. From this LDO there are 150mA and 300 mA versions.
The 300mA is the XC6204 E to H type. The SMD marking code 4A2D gives no clue on a 150 or 300mA version.

The original WEMOS D1 documentation shows a ME6211 LDO with 500mA current.
A search on internet shows ESP8266 current peaks of 200 up-to 430mA.

So the regulator of the WEMOS D1 clone seems to be under-engineered.

I have made 6 previous projects with NODEMCU boards running v2.0.0-dev7 up-to v2.0.0-dev11 with 8-12 DS18B20, which are extreme stable (1 automatic reboot in 49 days), but never tested these version on WEMOS D1.
The NODEMCU boards do have a LM1117 800mA regulator.

So I flashed v2.0.0-dev7 and v2.0.0-dev11 on WEMOS D1 clones.
Configurations with 2# DS18B20 (from the 3.3V from WEMOS D1) and a optional a Pulse Counter task connected to a frequency generator (see note) with 1 - 80 Hz (on 5V via level shifter)
Reporting temperature and Counter frequency every 30 sec. to Domoticz.
Reboots are send to a Domoticz switch for logging, via a rule: The result:
Even the rock-solid v2.0.0-dev7 and v2.0.0-dev11 reboots on WEMOS D1 with only 2# DS18B20!!
With pulse counter .. more reboots. Higher frequency ... more reboots.

The reboot soup:
- WEMOS clone with under-engineered 3.3V regulator for ESP8266 chip
- DS18B20 connected to under-engineered 3.3V regulator
- DS18B20 address and resolution problems
- A lot of interrupts
- A lot of WiFi transmissions
- 802.11 B/G/N mode (B more power-hungry than N ??? ...see https://www.itead.cc/wiki/ESP8266_Serial_WIFI_Module)

Breadboard assumption validation:
I disconnected the 5V connection and used a 3.3V regulator LM1117T-3.3 with 800mA specification. Thus by-passing the on-board regulator.

This unit with v2.0.0-dev7 is now running for 17 hour with a 60Hz Counter task. ... without reboots !!

Next step will be a 3.3V modification of some of my WIP units to test v2.0.0-dev7 and v2.0.0-dev11
See: viewtopic.php?f=2&t=5955&p=32473#p32473
Further tests will be with recent releases.

Will keep you informed,
Domosapiens


Note: frequency generator: €1.73 ,highly recommended:
https://nl.aliexpress.com/item/Signal-G ... Title=true

Yikes! FWIW, another issue with the XC6204 is that Vin MUST be at least 2V greater than Vout (2V dropout). So it's definitely a bad choice for the Wemos. BTW, the LM117 dropout is 1.2V, so besides its increased current capabilities it can tolerate USB powered operation (~5V Vin).

I'm running mostly NodeMCU's (LM117) with extra caps (470uF bulk + 0.1uF decoupling). But some of my ESPeasy devices are reflashed Sonoff THT10 / Basic. Reboots on all of them have been a persistent issue, with random occurrences every day. But in your situation with the inadequately designed Wemos clones, I think your VReg fix will reduce their reboot problems a lot.

- Thomas

Thanks for your response ThomasB,

I'm running mostly NodeMCU's

Are NodeMCU's free of the reboot problem ?

another issue with the XC6204 is that Vin MUST be at least 2V greater than Vout (2V dropout)

Where/how did you find that ?
This document https://www.torexsemi.com/file/xc6205/XC6204-XC6205.pdf
states in it's intro:
Dropout voltage 200mV@100mA
further down the line (page 12), I see a max. figure at 3.3V:
330mv@100mA
The worst diagram on page 21 suggests:
650mV@200mA

Anyhow, the voltage regulator seems to be underrated during Wifi operations.
Too much underrated to compensate it with a capacitor, during WiFi burst operations.
Probably it will effect the WiFi calibration during start-up.
As measured here here .. 430mA spikes ....

Are NodeMCU's free of the reboot problem ?

No, they randomly reboot. So do my Sonoffs. For reference, the Sonoffs don't reboot on ESPEasy R120.

Where/how did you find that ?
This document https://www.torexsemi.com/file/xc6205/XC6204-XC6205.pdf
states in it's intro:
Dropout voltage 200mV@100mA

Sorry, my mistake. Dropout won't be a problem here, but as you have identified the currents are a bit too high for the XC6204. So your proposed VReg upgrade is a good idea.

- Thomas

the issue is not dropout voltage - its most likely how steep the slopes are. how fast can the capacitors and regulator catch up?

try adding a 220µF low esr cap to the input of the regulator and 1µF to the output. also put a 100nf to each 3.3v pin on the nodemcu. for the 0.1µF i use 0805 smd types, they fit inbetween the pins

Thanks a lot! I have one project on wemos D1 clone and was wondering why it reboots frequently. Now I know why

That is good thinking of you Domosapiens.

I have a ESP that is also rebooting a lot, on average 1 hour. It is good that after a reboot the ESP comes back but it is still annoying.
I ordered a couple of: https://www.aliexpress.com/item/Free-sh ... 96723.html
and will hot-air and replace the 4A2D/XC6204 regulator with this S2LD/ME6211 to see if that solves the problem. Will let you know in a few weeks.

This post on reddit talks about the same problem: https://www.reddit.com/r/esp8266/commen ... 150ma_33v/

Sorry to say ...but ...

I disconnected the 5V connection and used a 3.3V regulator LM1117T-3.3 with 800mA specification. Thus by-passing the on-board regulator.

is not THE solution.
Further testing goes on ..

Due to 'strange' behaviourof a setup with multiple sensors interfacing to á WEMOS D1_Pro (and then getting 3.3V from the WEMOS), I changed the setup in such way that for power the sensor-PCBs are now bypassing the WEMOS and running from 5V Power Supply as provided by an addon Power-Shield.
The subject sensor configuration (for multiple measurement of light) is 1*BH1750+2*TSL2561+1*GUVA_S12SD.
These 4 sensor-PCBs should be compatible with 5V_Power, but the GUVA now 'strikes' and has a frozen output of 7-out-of-1024
Reason to plan exchange of the GUVA for an ML8511_PCB running at 5V_Power.
Alternative (not yet tested) would be to run the GUVA at 3.3V from a stepdown-converter hooked to the 5V-Power from the Shield.

Any comparable experiences?

I posted last week:

I have a ESP that is also rebooting a lot, on average 1 hour. It is good that after a reboot the ESP comes back but it is still annoying.
I ordered a couple of: https://www.aliexpress.com/item/Free-sh ... 96723.html
and will hot-air and replace the 4A2D/XC6204 regulator with this S2LD/ME6211 to see if that solves the problem. Will let you know in a few weeks.

A few days ago I received some original RT9013 voltage regulators. These are specified for 500mA max, definitly much better that XC6204 and still better than ME6211.
So last wednesday I replaced the 4A2D smd regulator with the RT9013 smd.
I cannot say that the stability is good now. It still resets once in a while. It is not easy to compare because last sunday there was a network update on the local wan where this wemos is situated. The network update did already improve the stability. It went from on avarage less than an hour to on average more than 10 hours between the reboots.
Since the replacement of the regulator it looks the same, it still reboots dayly.

So I cannot believe that the frequent reboots is a hardware issue, I start to believe it is either a software issue or maybe the wan network has still issues that causes the reboots.
The ESP Mega version is the experimental one from March 15 which I compiled myself. The system values for this wemos are about 14500 free ram and a load between 30-34%.
ASAP i will update the software with the newest precompiled version and see if that helps.

ESPEASY randomly reboots.
It is the fact and must be sw problem.
I have got about six wemos D1 with ESPEASY ( Rxxx, v2.0 and mega releases tested ).
Two DS18B20 and relay ( Wemos module relay board ).
All reboots randomly.
The same Wemos works like ethernet gateway for MySensors network - RFM69 radio module connected to it. So more power consumption then alone Wemos.
Runs weeks without problems.
Power from USB adapters for all of them.
Your sw and hw design must be prepared for this situation.
(I am sending values to my Wemos periodically every minute end after reboot I am setting safe initial values. )