Skip to main content

ESPCLOCK4 - Reduce operating current by removing UART module

The ESP32 Mini uses a CP2104 UART module to drive the USB port.  The CP2104 is configured in self-powered mode, which means that even if nothing is plugged into the USB port and the board is powered by a 3.3V source, it is still drawing ~100uA of current. I wanted to see the effect of removing this module on the power consumption of my test circuit.

Removing the UART module means losing access to the USB port. However, I should still be able to program it using a standard FTDI programmer, which I wanted to verify. So I took out the same FTDI programmer that was used for programming the ESP8266, and built a custom holder for the ESP32 Mini.




I started with the same connections as used with the ESP8266:
  • ESP32 VIN <-> FTDI 3.3V
  • ESP32 GND <-> FTDI GND
  • ESP32 TX <-> FTDI RX
  • ESP32 RX <-> FTDI TX
However, I could not get it to work. After some fiddling around, I discovered I could only get it to work by connecting RX - RX and TX - TX. Not sure why.

But now I am confident I can still program the board without its USB, so I proceed to remove the CP2104 module (marked in red) with a hot air gun.


After removing the UART module,  I verified that the board was still working by uploading the code here, and plugging it into the test circuit. Sure enough, the clock started ticking. It worked!

I then tried measuring the current consumption using the LTC4150. Unfortunately, the current seems too low to be measured. After doing some research, I decided to swap out the 0.05ohm sense resistor with a 1ohm resistor. Because I didn't have a 1% part, I simply used a normal 5% metal film resistor. However, I did calibrate it against a multimeter with a 10kohm resistor over a 3.3V power source. I found the measurement to be within ~20uA tolerance.

Note: with such small current, the power used by the LTC4150 itself becomes very important. From the datasheet, at 2.7V, its typical power usage is 80uA, with a max of 100uA. With a 10Kohm passive load, I got a reading of 430uA. Substracting 90uA from the reading gives me 340uA, which compares well with the multimeter reading of 326uA.

Using a 1ohm sense resistor means I can no longer use WiFi in my test code. The 500mA current drawn will cause the board to brown out and reset.

So finally, with the UART module removed, the current draw is 650uA - 90uA = 560uA, or ~0.6mA. This is a hugely improvement over the 1.13mA that I measured previously with the UART module intact.

So it looks like I am going to go ahead and create the full ESPCLOCK code with this setup.

Labels:

Comments

Post a Comment

Popular posts from this blog

Update: Line adapter for Ozito Blade Trimmer

Update (Dec 2021): If you access to a 3D printer, I would now recommend this solution , which makes it super easy to replace the trimmer line. I have been using it for a few months now with zero issue.
5 comments
Read more

Attiny85 timer programming using Timer1

This Arduino sketch uses Timer1 to drive the LED blinker: 1 2 3 4 5 6 7 8 9 10 11 12 13 14 15 16 17 18 19 20 21 22 23 24 25 26 27 28 29 30 31 32 33 34 35 36 37 38 39 40 41 42 43 44 45 46 47 /* * Program ATTiny85 to blink LED connected to PB1 at 1s interval. * Assumes ATTiny85 is running at 1MHz internal clock speed. */ #include <avr/io.h> #include <avr/wdt.h> #include <avr/sleep.h> #include <avr/interrupt.h> bool timer1 = false , led = true ; // Interrupt service routine for timer1 ISR(TIMER1_COMPA_vect) { timer1 = true ; } void setup() { // Setup output pins pinMode( 1 , OUTPUT); digitalWrite( 1 , led); set_sleep_mode(SLEEP_MODE_IDLE); // Setup timer1 to interrupt every second TCCR1 = 0 ; // Stop timer TCNT1 = 0 ; // Zero timer GTCCR = _BV(PSR1); // Reset prescaler OCR1A = 243 ; // T = prescaler / 1MHz = 0.004096s; OCR1A = (1s/T) - 1 = 243 OCR1C = 243 ; // Set to same value to reset timer1 to
2 comments
Read more

3D Printer Filament Joiner

I have been looking at various ways of joining 3D printing filaments. One method involves running one end of a filament through a short PTFE tubing, melting it with a lighter or candle, retracting it back into the tubing and immediately plunging the filament to be fused into the tubing: One problem with this method is that you can't really control the temperature at which you melt the filament, so you frequently end up with a brittle joint that breaks upon the slightest bend. Aliexpress even sells a contraption that works along the same line. As it uses a lighter or candle as well, it suffers from the same weakness. I am not even sure why you need a special contraption when a short PTFE tubing will work just as well. Another method involves using shrink tubing/aluminium foil, and a heat gun: But a heat gun is rather expensive, so I wanted to explore other alternatives. The candle + PTFE tubing method actually works quite well when you happen to melt it at the rig
1 comment
Read more