Injection locking: Where Christiaan Huygens meets the modern world...

[LittleWatchShop]

I designed the synthesizer and VCO for the DSP on an integrated circuits used in disk drives (circa 1995).  These designs went into WD, Seagate, and Quantum disk drives.  The VCO was slaved to the synthesizer (which was slaved to a crystal oscillator) and was designed to be adjusted according to where the read head was on the rotating disk (inside radius slower than when at the outside edge of the disk).

What we discovered was that as you dial the VCO near the frequency of the synthesizer, it just locked in and would not budge until you pushed it well past the reference frequency given by the synthesizer.  Wow...we were scratching our heads.  Somebody on the team discovered that we were not the first do notice this issue.

Christiaan Huygens also discovered that when clocks mounted on the same beam while closely regulated, will synchronize themselves.  In modern times, we call this "injection locking."

This bit of insight was exciting for me...being the son of a watchmaker.

[jdm]

16 minutes ago, LittleWatchShop said:

In modern times, we call this "injection locking."

Or more popularly,  spontaneous synchronization

 

[LittleWatchShop]

5 hours ago, jdm said:

Or more popularly,  spontaneous synchronization

 

very nice. so after my experience, i bought two pendulum movements with the goal of doing this very thing. never made it work. the reason was that they have to be the same frequency, but out of phase 

[LittleWatchShop]

Just now, LittleWatchShop said:

very nice. so after my experience, i bought two pendulum movements with the goal of doing this very thing. never made it work. the reason was that they have to be the same frequency, but out of phase 

i did not put the effort in to getting them calibrated

[LittleWatchShop]

5 hours ago, jdm said:

Or more popularly,  spontaneous synchronization

 

This stuff gives me a hard on...just awesome...and I have been part of it!

[HectorLooi]

Does this work even if the fundamental frequency is off, say by 5%?

[jdm]

5 hours ago, LittleWatchShop said:

never made it work.

First of all you need a medium that transmits resonance. The UCLA video cheats hugely with the board sliding in the same vector plane as the metronome pendulums. 

[LittleWatchShop]

4 hours ago, jdm said:

First of all you need a medium that transmits resonance. The UCLA video cheats hugely with the board sliding in the same vector plane as the metronome pendulums. 

i did. Also, the two clocks need to be regulated to near identical to each other.

[LittleWatchShop]

8 hours ago, HectorLooi said:

Does this work even if the fundamental frequency is off, say by 5%?

at some % difference, they will not lock. that is what we observed with the vco 

[KarlvonKoln]

Following this thread, for the cool stuff!