New version (v2.2.0) of the D1 Mini
It looks like the D1 Mini has been upgraded (link to Aliexpress)
- Seems like @wemos has added a 500mA resettable fuse like the D1 Mini Lite.
- There's a new 6-pin IC that is connected to the DTR and RTS lines off the CH340G, so I suspect it's doing the work of booting into firmware mode. Maybe the UMH3N that is on the D1 Mini Pro?
- Instead of an ESP-12F, it's been changed to an ESP-12S module, without the extra (unconnected) pins
Anyone know what the square piece is above the CH340G? And what is the voltage regulator? It doesn't have the same markings as the RT9013 on the old D1 Mini/Mini Pro
That square component is the quartz oscillator for the CH340G. It's also the same as on the D1 Mini Lite.
@deshipu Yep, you're right. You can see the traces from the crystal nicely on the Lite.
I got tripped up by the lite having 2 crystals, but the other one is clearly feeding the ESP8266.
The ESP-12S advertises better antenna performance. Anyone care to compare one of these new ones with an older D1 Mini?
Was hoping they'd do this and also hoping they do one with the ESP-07S still too ;)
I have a question for @wemos too, the ESP-12S has the built in pull-up/down resistors required to run the module yet the v2.2.0 of the D1 Mini still appears to have these in addition - any reason why?
What about power CH340G from USB and not from 3.3V line? (I remember several post in old forum speaking about low power consumption, and one of the most thing to improve it was to power CH340G only when wemos is connected to USB, as you can power it, for example, from a LiFePO4 battery directly in 3.3v pin.
Have they changed it ?
@juan3211 hard to say without a schematic.
I think this is not easily possible.
According to my understanding of the CH340G datasheet, the device supports 5V and 3.3V operation, but the output voltage is VCC-0.5V (min). In case you would apply 5V as supply voltage to VCC, the outputs will have 4.5V at least or in other words, it will operate as 5V device. What we want is a 3.3V I/O voltage. FTDI supports this, maybe some others, too. In case of the CH340G two regulators would be needed, one for the ESP12S and one for the CH340G. Don't know what's happening if the CH340G is not powered, but 3.3V signals are applied ( parasitic power supply ?)
From the layout I believe it's easy visible that they use a similar approach as with the previous version. The CH340G is connected to the regulator output, which should be 3.3V. The USB-5V is connected to the VIN and EN of the regulator.
I heard back from Wemos on the new regulator. They've replaced the RT9013 with a ME6211. There's a slightly higher input voltage, up to 6-6.5 volts. But it looks like quiescent current is slightly higher (not 55A as the datasheet says, but 55 uA), and the dropout voltage looks to be slightly higher as well.
It'll be interesting to test the current in deep sleep. I think it will be 20-30 uA higher than the old D1 Mini
I've measured power draw on these new v2.2.0 D1 minis. The new regulator is pretty good for quiescent current, better than the RT9013 that was there before.
- 70-75 mA average with wifi on
- 15.4 mA with wifi off
- 115 uA in deep sleep (comparable to the D1 mini pro, and ~30% lower than the RT9013)
Nice job @wemos
Has anyone looked at how clean the voltage is on the 3.3V side with the new voltage regulator?
I found that the WEMOS mini D1 V1 had quite a noisy 3.3V that sensors like the BMP280 didn't like.
Do you think it would be possible for D1 mini that was more battery friendly could be made with a switch or jumper that isolated the voltage regulator and UART chip?
Obviously it would only be a 3.3V system, but I think for most people this would be suitable. Alternatively WEMOS could release a module that was designed specifically for running on battery.
The new voltage regulator on the v2.2 board is an improvement, but not really enough to be fully useful.
I haven't hooked it up to a scope, but I've had no problems with v2.0.0, 2.1.0, and 2.2.0 and the BME280 or any sensors due to noisy voltage. Have worked with a lot of D1 minis, and a lot of BME280s (>50).
The ESP8266 is really sensitive to noisy voltage, and I'd think you'd be more likely to have problems with them than with the sensors.