This means that the value of the negative input resistance must be greater than the series resistance (ESR) of the crystal.
Only then will there be a net amplification that leads to oscillation.
F.2: Influence on the transient response safety
If the value of the negative input resistance is too small (i.e. -Rneg is too weak), the following happens:
- The quartz receives too little energy → slow oscillation or no oscillation at all
- Oscillation only starts at a higher supply voltage or temperature
- Start-up in low-power operating modes becomes unreliable
Typical cause:
Some modern MCUs have weak oscillator amplifiers for efficiency reasons, which means that -Rneg is smaller than in older IC generations. At the same time, many designs work with small load capacitances or long traces, which increases parasitic losses.
Q.3: Why are crystals with low ESR particularly important?
The ESR of the crystal defines its internal losses. A low ESR means
- lower losses
- lower necessary counter-amplification
- high transient stability even with weak oscillator stages
- shorter start-up time
- More stable oscillation over temperature
Practical recommendation:
- Use crystals whose ESR is significantly below the maximum value specified by the IC.
- If a microcontroller specifies a maximum ESR of 70 Ω, for example, a crystal with an ESR of 30-50 Ω is ideal. This leaves a sufficient safety margin against a possibly low negative input resistance of the oscillator.
F.4: Conclusion on the interaction of -Rneg and quartz ESR
The transient response reliability essentially depends on the following:
- the internal oscillator provides sufficient negative input resistance,
- the crystal has a sufficiently low ESR,
- the load capacitances are correctly dimensioned.
Only if the ratio of -Rneg to ESR is correct will the crystal start quickly, reliably and over the entire temperature and voltage range.