Email: Password: Remember Me | Create Account (Free)

Back to Subject List

Old thread has been locked -- no new posts accepted in this thread
???
05/07/11 11:12
Read: times


 
#182189 - Sensitivity since lack of "mass" to uphold frequency
Responding to: ???'s previous message
The voltages where a gate input switches from low-to-high or high-to-low is affected by any noise on GND or VCC.

And the chips consumes extra power during switching.

Two oscillators running at different frequencies will have a resulting interference frequency, i.e. they will regularly get turn on/off times close to each other. And at these times, they are very sensitive to influences on GND and VCC. If the "bumps" on VCC/GND are long enough to span the difference in half-period time then the two oscillators can quickly force each other into getting in lock-step.

The closer the two frequencies are, the harder it is to manage to filter the supply enough that they will not interfere.

The capacitor in a RC circuit doesn't have any real "mass" forcing a constant frequency. If the gate switches 10mV earlier, the capacitor will happily accept that the circuit switched between charge/discharge that bit earlier so the frequency change can happen very quickly.

In the case of the pendulums, there is a mass in the system, so they required some time to transfer energy between each other before they managed to tune in to the same common resonance frequency of a single "super-pendulum".

An electronic design that have some form of "mass" can better manage the bumps in VCC/GND at the interference frequency. They still get their own frequency affected - there is more jitter in the output frequency - but the two oscillators must be much closer together in frequency for them to jump into lock-step.

List of 18 messages in thread
TopicAuthorDate
An excellent demonstration of coupling between oscillators            01/01/70 00:00      
   Many many years ago...            01/01/70 00:00      
      You don't suppose that was for a reason, do you?            01/01/70 00:00      
         Sensitivity since lack of "mass" to uphold frequency            01/01/70 00:00      
            Even in well-filtered applications ...            01/01/70 00:00      
               Ground bounce...            01/01/70 00:00      
                  Length of gnd/vcc "bumps" in relation to period difference            01/01/70 00:00      
                     That was an interesting problem            01/01/70 00:00      
                  Signetics (later as Philips) made 74HC7046            01/01/70 00:00      
   That's a really nice demo            01/01/70 00:00      
      One of many            01/01/70 00:00      
   On a similar note            01/01/70 00:00      
      No coupling            01/01/70 00:00      
      Very nice...            01/01/70 00:00      
         Easy if they have a screw adjustment + web cam            01/01/70 00:00      
            It remembers me...            01/01/70 00:00      
         huh?            01/01/70 00:00      
            Just different location of screw adjustmenton clock pendulum            01/01/70 00:00      

Back to Subject List