Skip to main content

ESPCLOCK4 - Design

The idea for ESPCLOCK4 came after encountering this article. It made me realize the successor to the ESP8266 - the ESP32 - has an ultra low power (ULP) coprocessor that can perform simple processing eg. toggle pins, read ADC etc. while the power hungry main CPU is in deep sleep. 

After reading through the specs, it seems the ULP can substitute for the ATtiny85 in the ESPCLOCK3 design. The advantage is that the I2C bus is not longer required, and the circuitry can be vastly simplified. 

Just like the watchdog timer on the ATtiny85, it appears the real-time clock (RTC) that drives the ULP is not very accurate (5% jitter). However, I think I can use the same calibration technique employed to adjust OCR1A on the ATtiny85 in the ESPCLOCK2 design to dynamically tune the RTC. Basically, by using the difference between the computed network time and actual network time, the RTC counter can be increased or decreased accordingly.

So here's the preliminary design. The ESP32 will again serve only as the conduit to the Internet to retrieve the network time. It will be in deep sleep otherwise to conserve power.

The ULP will perform 3 tasks:

  • It will advance the physical clock by generating the tick pulses.
  • It will check to see if battery voltage is low. If so, it will wake ESP32 up, which will write the physical clock time to EEPROM and shutdown.
  • It will monitor a reset button and perform a software reset when the button is pressed.
As before, a supercapacitor will be used to supply power to the circuit until the ESP32 has time to properly shutdown when power is removed.

I am interested to see what's the power profile of this approach compared to ESP8266 + ATtiny85.

Comments

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.

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

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