Return to Answer

Copy edited (e.g. ref. <https://en.wikipedia.org/wiki/Sine_wave>, <https://en.wikipedia.org/wiki/Phase_noise>, <https://en.wikipedia.org/wiki/Ohm#Definition>, <https://en.wikipedia.org/wiki/Volt>, <http://english.stackexchange.com/questions/15953>, and <https://en.wikipedia.org/wiki/Slew_rate>).

Source Link

edit approved Apr 28, 2018 at 11:14

Peter Mortensen

1.7k
3
17
23

That sinwavesine wave, having a finite and easily computed slope, along with some noise floor in the internal circuitry that DOES square up the resonator signal, cause a predictable phasenoisephase noise or TimeJittertime jitter.

Use the formula

TjitterT_jitter = VnoiseV_noise / SlewRate

to predict the timing wander of this clock source.

Beware that any other circuits will only further add jitter. Use the same formula.

Assume your sinsine-to-square circuit has 10Kohm10 kohm Rnoise. This is 12 nanoVoltsnanovolts/rtHz thermal random/Johnson/Boltsmann noise density. If bandwidth is 100MHz100 MHz, the total input noise voltage is 12nV12 nV * sqrt(100MHz100 MHz) = 12nV12 nV * 10^4 = 120 microVoltsmicrovolts RMS.

Assume the crystal frequency is 10MHz10 MHz, with +-1volt1 volt peak sinsine amplitude. The slewrateslew rate is 1v1 V * 6.28 * 10MHz10 MHz = 63 volts/uSµs.

What is the edge jitter? TjT_j = VnoiseV_noise / SlewRate

TjT_j = 120 microVoltsmicrovolts / (63 volts/uSµs) = 2 picoseconds.

That sinwave, having finite and easily computed slope, along with some noise floor in the internal circuitry that DOES square up the resonator signal, cause a predictable phasenoise or TimeJitter.

Use the formula

Tjitter = Vnoise / SlewRate

to predict the timing wander of this clock source.

Beware that any other circuits will only further add jitter. Use the same formula.

Assume your sin-to-square circuit has 10Kohm Rnoise. This is 12 nanoVolts/rtHz thermal random/Johnson/Boltsmann noise density. If bandwidth is 100MHz, the total input noise voltage is 12nV * sqrt(100MHz) = 12nV * 10^4 = 120 microVolts RMS.

Assume the crystal frequency is 10MHz, with +-1volt peak sin amplitude. The slewrate is 1v * 6.28 * 10MHz = 63 volts/uS.

What is the edge jitter? Tj = Vnoise / SlewRate

Tj = 120 microVolts / (63 volts/uS) = 2 picoseconds.

That sine wave, having a finite and easily computed slope, along with some noise floor in the internal circuitry that DOES square up the resonator signal, cause a predictable phase noise or time jitter.

Use the formula

T_jitter = V_noise / SlewRate

to predict the timing wander of this clock source.

Beware that any other circuits will only further add jitter. Use the same formula.

Assume your sine-to-square circuit has 10 kohm Rnoise. This is 12 nanovolts/rtHz thermal random/Johnson/Boltsmann noise density. If bandwidth is 100 MHz, the total input noise voltage is 12 nV * sqrt(100 MHz) = 12 nV * 10^4 = 120 microvolts RMS.

Assume the crystal frequency is 10 MHz, with +-1 volt peak sine amplitude. The slew rate is 1 V * 6.28 * 10 MHz = 63 volts/µs.

What is the edge jitter? T_j = V_noise / SlewRate

T_j = 120 microvolts / (63 volts/µs) = 2 picoseconds.

Source Link

created Apr 28, 2018 at 3:40

analogsystemsrf

34.6k
2
19
48

Use the formula

Tjitter = Vnoise / SlewRate

to predict the timing wander of this clock source.

Beware that any other circuits will only further add jitter. Use the same formula.

Assume the crystal frequency is 10MHz, with +-1volt peak sin amplitude. The slewrate is 1v * 6.28 * 10MHz = 63 volts/uS.

What is the edge jitter? Tj = Vnoise / SlewRate

Tj = 120 microVolts / (63 volts/uS) = 2 picoseconds.