Let's Control It

Thankyou so much for this page - I've had 2 boards that I've been battling with trying to work out why they are intermittently working from 5v, even tried 3.3v - and the fix.. just cut one of my many spare USB cables and use that instead - From what I can tell it looks to be working smoothly now!

What baffled me was USB worked absolutely fine.. but everything else was intermittent!

Might be a hardware aspect.

Sounds strange, but USB-cables are not all identical, and that is not visible from outside:
- some are sold for power supply only (and only have 2 cores for 0/GND and 5VDC)
- others are fully equipped for transfer of power&data.
- quality of manufacturing might be an issue, why some transfer is not OK [ price & quality have relation]

Therefore your approach is practical:
if errors, try a few more, and keep the best suitable for your application => mark the 'weak ones' (to prevent future frustration)!
If failing for power supply, then probably also little value for other applications, but as your experience shows still data transfer might be OK.

Why not doing something like that ?
I have cut the 5V input track and solder the 5V pin to the 5V USB arrival with the red wire, so far it works.

Be careful with such a mod, not to power the board and connect it to a PC to program it.
Normally there is a diode to protect the PC.

Yes you are right,
when connecting to a PC, I switch off the external 5V power supply.

But, except the first time, I am using OTA firmware updates.

The problem is solved a little easier than mentioned above. It is necessary to cut the trace from pin 5v to the cathode of the diode through which power is supplied from the USB connector. You need to cut it next to pin 5v. Then solder a 1n4148 diode between pin 5v and the anode of the aforementioned diode. The 1n4148 diode is soldered anode to pin 5v.

Hi guys,
I had these problems for a while since I had dozens of these Wemos D1 Pro mini Clones, at least 70% in a batch of 100 were faulty. I've been trying to find the problem, and I found that the UART IC is faulty out of the factory. Replacing it with the one from the functional board solves the problem, but I don't have those, so I had to find another solution. Testing the difference in the behavior of the UART IC powered from the USB port and from the external 5V pin, I've noticed a difference in voltage on the RST pin. Since the reset pin is powered directly from ESP IC by 3.3V, the reset function is performed by the pulldown RST pin to the GND. Powering board from 5V externally, measured voltage on RST pin was 2.1V which tells me that something draws current from ESP8266 IC and it keeps the ESP IC in reset state preventing it to boot up. I figured out that UART IC keeps DTR pin in high state (which normally sends reset signal to ESP) and it draws current from RST pin through the Dual Digital Transistor UMH3N when it's powered from 5V externally. Then I came up with this workaround, which is the best one so far, and it works fine. I've shorted two pins together on UART IC. To be precise, I've shorted pins 5 and 6 (VIO and VDD), and in that way, the voltage supplied to the RST pin is 3.3V no matter how you power the WEMOS board, from an external 5V or USB, and the board works just fine. It can be programmed via USB, it boots up just fine, and it can be powered from 5V and from USB (Do not connect both simultaneously). Wifi works as it should. Take a look at the image I attached. It is the best and easiest solution that works for me. I hope this helps you too.

Cheers!

Looks like a valid fix to me.
It is also used like this on the Adafruit Feather board: https://cdn-learn.adafruit.com/assets/a ... 1604095577

Also this note in the datasheet for the CP2104:

Note 3 : VIO can be connected directly to VDD or to a supply as low as 1.8 V to set the I/O interface voltage.

Great !

thanks much.

cvijanovic84 wrote: ↑26 Dec 2022, 14:51 ... shorted pins 5 and 6 (VIO and VDD), and in that way, the voltage supplied to the RST pin is 3.3V no matter how you power the WEMOS board, from an external 5V or USB, and the board works just fine. It can be programmed via USB, ...

Question from a non-expert: „can be programmed via USB“ means a USB software mod is required to make the device work or that the device can still be flashed as usually? (Stupid question; I assume it’s the latter).

Also: Can this solder bridge 5-6 be safely applied to all units even if it unknown whether they have the flaw or not?

This afternoon I was working with one of the D1 Minis then found it could not power on via the 5V pin, but could power via USB. I changed two other boards which seemed okay initially but showed the same situation when connected. I eventually found a soldered short circuit between the ground pin and the adjacent D4 pin. This has damaged the 3 boards such that they cannot work from the 5V pin. Maybe this could help with the diagnosis being discussed here. Also found the in-built LED no longer works. Any ideas?

EDIT. Soldered short circuit was on my pcb panel holding the D1 Mini.

I would like to thank everyone who participated in this thread and helped me resolve the issue I was facing with my Wemos D1 Mini Pro v1.0 !

I'm powering my Wemos D1 Mini Pro v1.0 externally using a 3.3V supply. However, since this voltage mirrors onto the regulator input, the exact same issue arises: the USB chip occasionally enters an undefined state, which in turn messes up the boot procedure.

Unfortunately, LOLIN, the original manufacturer, no longer produces this module (they now sell version 2.0 which includes battery charging features, etc.). On the bright side, some third-party manufacturers have updated the module to version 1.1, as seen here:

https://www.ebay.com/itm/255283221967?m ... media=COPY

I've come across two updated designs. The first one is a simple design that replaces the protection diode with a 0R jumper. The second, a more sophisticated design, powers the USB chip exclusively with 3.3V, keeps the protection diode and also adds a fuse for external 5V.



You can see the schematics of the original (v1.0) and the elaborate solution (v1.1) at the end of this page :

https://nettigo.eu/products/wemos-d1-mi ... ifi-module

In my case, I've replaced the diode on my existing modules with a 0R jumper, and I'm happy to report that they're all working perfectly now. Once again, I want to thank everyone for their valuable input!"

Just make sure to never connect it to a PC when you have the external power applied.

TD-er wrote: ↑12 May 2023, 09:47 Just make sure to never connect it to a PC when you have the external power applied.

I won't !

However, I think that this danger applies to external 5V only.
I am using 3.3V, which comes after the regulator.

That depends on whatever is output to the "5V" pin of the USB connector, or the D+/D- pins.

I would expect USB ports to have some diode themselves, so it will probably not be too catastrophic, but still it may be tricky as not all logic chips may be happy with voltages applied exceeding their Vcc/Vdd voltages when the device (e.g. smart charger or PC) is powered off.

@vpapanik, thanks for the information about the revisions.

- Thomas

This guy sells new modules with the CH9102 chip. Do you think that this could resolve the problem without having to manually bridge the protection diode ?
https://www.aliexpress.com/item/1005005933528007.html

vpapanik wrote: ↑23 Jan 2024, 19:00 This guy sells new modules with the CH9102 chip. Do you think that this could resolve the problem without having to manually bridge the protection diode ?
https://www.aliexpress.com/item/1005005933528007.html

The only way to know for sure is to purchase such unit and test for yourself...

BUT:

When buying new devices, especially if you go for 16MB Flash, PLEASE buy ESP32, ESP32-S3 or ESP32-C3 units, with 16MB Flash if available (probably not for ESP32-C3), so you can have all plugins and features in 1 package, and also a larger file-system up to 8MB. These ESP32 units are much more capable, with more RAM and CPU power, and fully supported by ESPEasy. The ESP8266 line isn't developed/improved by Espressif any more, just produced for those 'fixed' on those chips.

Yep please do not use ESP8266 anymore for new projects/designs.

I really like ESP32 and I know that it's better than ESP8266.
However, for my project, ESP8266 does the job fine.
80/160 MHz speed and 16M (even 4M) is enough for me.
I will definitely move to ESP32 when I am in need for more GPIOs.

However, I am really curious to know specific ESP32 advantages in more detail.

I'm spending serious amounts of time trying to make it all fit in the limited resources for the ESP8266.
That's for one a very strong reason to go for ESP32-variants.

TD-er wrote: ↑23 Jan 2024, 20:57 I'm spending serious amounts of time trying to make it all fit in the limited resources for the ESP8266.

I am personally very grateful to you for that. Thank you!