Lolin32 power draw
I measure the following on the current dev version of nodemcu firmware:
Awake, wifi enabled: 114 mA
Awake, wifi disabled: 114 mA
Deep sleep: 1.05 mA
Obviously, the firmware may have a ways to go, but that's a bit disappointing.
It's got a ME6211 and CP2104, but neither of those should account for the high draw in sleep. Perhaps its the circuitry to charge the battery?
Anyone know what the ESP-WROOM-32 should draw in sleep by itself?
@wemos: A schematic would be very helpful. Also, one of the very useful things that makes the D1 mini an improvement over many other ESP8266 boards is this power usage. A sleep current of 1 mA really limits the usefulness of the lolin32.
Where did you make your current measurements ?
If they were made at the USB cable connector they likely include the battery charger current (a little even if no battery) and the CP2104 may itself be active even if the ESP is in deep sleep because of the Vusb voltage.
Try measuring the current at the battery connector (using only a 3.7v lipo battery and no usb) as this does not power the battery charger and I think the CP2104 will go to sleep. Your results may be better.
For a developing schematic see this
Please post your findings - we are all learning at the moment !
Using an external 12V 2A power brick with LM2596S buck boost converter, set to output 5V.
Buck boost output connects to GND and 5V pins, with a multimeter (in DC 10A ammeter mode) connected in series between the + wires.
[wall] -- 240V --> [AU power adapter] -- 12V --> [buck boost] -- 5V --> [DMM] --> [Lolin32] -- ~4-5V --> [lipo]
0.120A idle, awake, with default firmware, scanning for Wifi APs. Red CHG off.
0.630A when charging empty 150mAh lipo. Red CHG on.
Slowly falls down to around 0.170A then the red CHG led switches off when the battery is full (~4.15V).
Drops around 0.01-0.02A every few minutes when I checked it.
Drops quicker towards the end.
I can graph it properly if anyone wants.
Similar numbers when charging empty 1200mAh lipo and 4400mAh twin ICR18650 pack.
Same when I switched to a XL4015E1 buck boost.
@dtricker: My measurements were with the 5V pin, not USB power. No battery hooked up.
Yes, it's very possible the battery charger current is wasteful. I'm less interested in running it via lipo though, so that would be unfortunate.
@wemos Thanks for the schematic, its really useful
@mickles From the schematic R5 would appear to add a fixed 500uA to 5v current draw. It is used to turn the MOSFET Q2 on when only being powered by a lipo cell. So in theory you could change it to a higher value or remove it altogether if never using a cell.
The TP4056 charger is also specified to take around 100uA in standby, again if never using a cell you could disconnect it.
The CP2104 is likely to amount to about 200uA in standby. I assume the function of the Q1 and the current mirror is to disconnect the Vbus rail from the 5v supply when powering from the 5V pin. I think this will ensure the CP2104 realises it is not connected to the USB and go into standby, also it will disconnect the usb pullup resister saving more current.
Anyone else have any ideas about the current mirror, I'm not sure how much current it actually uses ?
@wemos Please create new version LOLIN32 without CP2104 (as WEMOS XI) and add switch for TP4056. I need deep sleep with 5 uA. Thank you.
@josef I also hope those things get updated in new version, but it will surely take some time to release a new version. For my project I need Li-Po pin and 3V3 LDO enable pin, so I can turn off things while charging the Li-Po. @wemos thanks for the super great work you are doing at an unbeatable price, keep it up!
@josef I think that in such a situation you should just use a bare EPS32 module, and add only the parts that you need. It's very unlikely that someone will make a custom version of a devboard with exactly the things you need, and without any extras.