Skip to content

A portable, efficient, passive, consensus based (relative) time synchronization mechanism over radio waves. Cost-effective radio receivers too!

Notifications You must be signed in to change notification settings

kholia/ConsensusBasedTimeSync

Folders and files

NameName
Last commit message
Last commit date

Latest commit

 

History

4 Commits
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 
 

Repository files navigation

What?

We present a portable, efficient, passive, consensus based (relative) time synchronization mechanism over radio waves.

We demonstrate the portability of this mechanism on a bunch of common computing platforms:

  • AMD64 PCs

  • Raspberry Pi(s)

  • Android phones

  • Raspberry Pi Pico (Yep!)

Why?

FT8, WSPR, and other digital modes need the receivers and the transmitters to be synchronized in time.

What if there is no external and reliable reference available for time synchronization?

  • RTC clocks can drift over time.

  • GPS system can be jammed.

    In comparison, a passive receiver system (i.e. our system) operating over a wide range of amateur radio frequencies may be a bit harder to jam.

  • Internet may NOT be available in remote areas.

  • Sundials can be cumbersome to deploy, calibrate, and perhaps not accurate enough ;)

Problem Statement: How do we ensure that we are still able to communicate over amateur bands using time-synchronized digital modes in absence of external and reliable references of time?

How?

Initially, we don't know the (relative) time. We try to decode FT8 traffic with a timing fudge factor of 0. If we don't succeed to decode any traffic, this fudge factor is incremented. This process is repeated until we are able to decode FT8 traffic successfully. The final value of the fudge factor gives us the (relative) time offset we need.

Hardware

We use Si4732-A10 IC for HF SSB reception. The Si5351 module provides a clock output to this Si4732-A10 IC. The other outputs of the Si5351 module can be used to implement a transmitter.

Thus the whole system is actually a low cost, portable, and efficient digital transceiver!

The whole system is beginner friendly, and can be built easily on a breadboard in under an hour or so.

The Si4732-A10 IC is fairly easy to solder on a SMD-To-DIP adapter PCB - no need to worry about it ;)

For a low-cost alternate option to Si4732-A10, checkout the CD2003 option included in this repository.

Tip: Use KiCad 7.x for handling these designs.

Update (September 2023): After on-air testing, we recommend building the CD2003-BoB-v2 design included in this repository over other designs for HF! The CD2003-S-BoB design works too but BFO frequency tuning seems like a hard, finicky process.

Update 2 (November 2023): Please use the CD2003-BoB-Final version. It has some minor correctness fixes in it.

Update 3 (March 2024): Please use the CD2003-BoB-Final-v2 version. It can be automatically assembled by JLCPCB!

Pinout

ATTENTION: The Si4732 IC uses 3.3v signaling! Don't attach it to a 5v MCU without a bi-directional level shifter.

For sample programs, see https://github.com/pu2clr/SI4735/tree/master/examples.

We include a fully-functional FT8 receiver firmware (for Raspberry Pi Pico) in this repository. It works on 20m (14.074 MHz) by default. To use this firmware, write firmware/run_ft8.uf2 to a Raspberry Pi Pico board.

#define CAL            15 // Automatic calibration entry
#define SDA_PIN        12 // GP12 on Pico
#define SCL_PIN        13 // GP13 on Pico
#define RESET_PIN      10 // GP10, Si4732 stuff

Si5351 module's CLK 0 goes to RX_OSC of Si4732. This has the 32.768 kHz clock on it.

Si5351 module's CLK 2 goes to GP15 (CAL) of Pico. This is used for automatic calibration of Si5351 module.

Audio output comes from L and R pins of the Si4732 BoB.

Pico's Pinout

Action!

Action 1

Credits

https://en.wikipedia.org/wiki/Standing_on_the_shoulders_of_giants

Prior Work

JTSync and JS8Call are capable of doing time synchronization but they are NOT portable enough and can't run on microcontrollers!

Future

We are also interested in automatic, microcontroller based (full) time synchronization systems powered by WWV and other time sources.

And of course, we will be using DT values from the decoded FT8 traffic to derive a more accurate fudge factor soon!

References

About

A portable, efficient, passive, consensus based (relative) time synchronization mechanism over radio waves. Cost-effective radio receivers too!

Resources

Stars

Watchers

Forks

Releases

No releases published

Packages

No packages published