Skip to main content

ESPCLOCK2, Part 1 - Getting timezone and local time

I am almost done with the implementation of ESPCLOCK2, so I will start a series of posts detailing the obstacles encountered and lessons learnt from the project.

The original ESPCLOCK obtains the geolocation of the clock during setup from the user's browser, and use the geolocation to convert to local time using Google's Timezone API. As noted here, this approach hits a major snag as recent browsers started to restrict the use of navigator.geolocation.getCurrentPosition() to Javascript running from SSL-secured websites. Since it would be extremely unlikely for WifiManager to support SSL hosting, another approach is needed.

After considering the alternatives, I think the simplest approach will be to host a simple PHP script on a server. This has several advantages: 1) There are no API keys to muck with. All timezone APIs I looked at require you to register for some kind of API key. 2) You won't be affected by API changes, or the service shutting down. You can run the script from any machine that is properly NTP-synchronized (most OS do it by default these days, even Windows). I am hosting the script here, which is open to public access. It is basically a simple 3-liner:

1
2
3
4
5
6
7
8
9
<?php

if (isset($_REQUEST['tz'])) {
  $time = new DateTime();
  $time->setTimezone(new DateTimeZone($_REQUEST['tz']));
  print $time->format('h:i:s');
}

?>

It takes a timezone string found from the list here, and returns the local time for that timezone. The script has very low memory, network and disk access overhead.

There is still the problem of trying to detect the timezone of the user from the browser during setup. This is done by using Intl.DateTimeFormat().resolvedOptions().timeZone, which is supported by more recent browsers. If for some reason the browser does not support this, the user can always populate the field manually.

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