Power consumption of TTGO T8 Board

[ThomasChr]

Hello Forum,

I‘ve got a TTGO T8 v1.7 Board, details can be found on Github https://github.com/LilyGO/TTGO-T8-ESP32 and I‘ve attached the schematics. The board can be powered by a Battery and this battery does only provide the 3.3V rail. Seems pretty right to me!

But when I use machine.deepsleep() and provide power via the battery the consumption only goes down to 3.5 mA. I‘ve desoldered the SIL2104 Usb/serial chip which is on the 3.3V rail and got it down to 3.1 mA. But thats still way to high. Has anyone any clues what is wrong here?
Is it the crystal? Or a cheap 3.3V regulator?
The spiram and flash seems to be powered by the ESP and I assume they got powered down in deepsleep. I‘m just curios what is eating that much power away and any help on this will be greatly appreciated!

Thomas

[jimmo]

What's your setup for measuring the current?

Are you able to verify with a multimeter that VDD_SDIO is in fact being turned off in deep sleep?

The regulator is a buck converter... the datasheet doesn't give a good indication of what its low-load current is like but of course the efficiency does drop at lower output currents.

[ThomasChr]

Hi jimmo,

thanks for your help!

This is the code I use, maybe I can switch something of in the Code?

Code: Select all

import time
import machine

led = machine.Pin(21, machine.Pin.OUT)

while True:
   # 30 Secs blink
   for i in range (0, 15):
       time.sleep(1)
       led.value(1)
       time.sleep(1)
       led.value(0)
   
   # 30 secs deepsleep
   machine.deepsleep(30000)

I‘m measuring the current flowing from the battery (my lab bench power supply) at 4V with Dave Jones uCurrent. I noticed that decreasing the Voltage reduces the power consumption clearly and thought that this must be a not very efficent power regulator. (out of my head the numbers are something like 0.5 less mA if you reduce the power by 0.5V - the drop seems to be fairly linear so I don‘t think it‘s some kind of brownout)

I will measure VDD_SDIO tomorrow when I‘m back at home, a good idea. I have no problem with damaging the board if that is needed to satisfy my curiosity.

I also will post the exact consumption at 4V, 3.8V and 3.5V - without the SIL2104.

Thank you very much so far!

Thomas

[ThomasChr]

VDD_SDIO is not pulled low when entering deepsleep. So flash and psram aber active in deepsleep. Any ideas how I could change that?

[ThomasChr]

Some news here:
1. VDD_SDIO does not go low when entering deepsleep. It seems that the Resistor R20 is placed on the board and pulling VDD_SDIO to 3.3V. Removing it helped and I lost one more mA (1.5 mA total in deepsleep)
2. I powered the board directly on the 3.3V rail, so the SYS8008B Voltage regulator was not involved and I'm down to 295 µA.

So still there are some 280 µA to catch. Anyone got an idea what I can try now?
Of course, everything is allowed, even experiments which will destroy the board, all to satisfy my curiosity!

Thomas

[ThomasChr]

Maybe it‘s a Pullup Resistor burning power? But I‘m not sure which one... I‘m assuming the GPIOS of the ESP32 are floating in deepsleep and so not sourcing any current to a pullup resistor like R14 which is pulling GPIO12 high.

[Roberthh]

Even if you source by 3.3V, the power regulator is still connected at it's output.

[ThomasChr]

So blowing the power IC off the board could help?

[Roberthh]

Hard to say. I do not know the internals of that chip, and the data sheet does not tell much about reverse sourcing.

[ThomasChr]

I‘ll try it tomorrow and will report the outcome here!