IEC 60444 6 Edition 2 0 2013 06 INTERNATIONAL STANDARD NORME INTERNATIONALE Measurement of quartz crystal unit parameters – Part 6 Measurement of drive level dependence (DLD) Mesure des paramètres des[.]
Reversible changes in frequency and resistance
Reversible changes refer to variations in frequency and resistance that occur under consistent drive levels, observed after repeated measurements taken alternately at low and high levels, or following continuous or quasi-continuous measurements from the lowest to the highest level and back These changes are considered reversible if they remain within the limits of measurement accuracy, indicating a stable and reliable measurement process.
Irreversible changes in frequency and resistance
Irreversible changes are significant changes in frequency and/or resistance occurring at low level after an intermediate measurement at high level e.g when a previously high resistance at low level has changed in the repeated measurement to a low resistance Especially, when the crystal unit has not been operated for several days, its resistance may have changed back to a high value when operated again at a lower level Greater attention should be paid to the irreversible effect since it can significantly impair the performance of devices, which are operated only sporadically.
Causes of DLD effects
Whereas the mostly reversible effects are due to excessive crystal drive level, the irreversible effects are due to production, especially to imperfect production techniques Examples of causes are:
– Particles on the resonator surface (partly bound by oils, cleaning agents solvents or bound electro-statically);
– Mechanical damage of the resonator (e.g fissures due to excessively coarse lapping abrasive which may increase in size);
– Gas and oil inclusions in the electrodes (e.g due to a poor vacuum or an inadequate coating rate during evaporation);
Inadequate electrode mounting can significantly impact performance, often due to poor contacting between the electrodes and the mounting surface This can be attributed to various factors, including the use of conductive adhesives with insufficient metal components, inadequate baking or overheating during the manufacturing process Furthermore, excessive contact resistance between the conductive adhesive and the electrodes or mounting can also hinder optimal performance, emphasizing the need for precise mounting techniques to ensure reliable connections.
– Mechanical stresses between mounting, electrodes and quartz element
4 Drive levels for DLD measurement
For the DLD measurement, multiple drive levels are applied, including a low level and a high level, with the latter being the nominal drive level that matches the steady-state application level.
The drive level for crystal units should be set below the maximum applicable level outlined in Annex A, with a standard crystal current of 1 mA assumed if not specified, corresponding to a velocity of 0.2 m/s for AT-cut crystals To calculate the drive level in watts, the mean value of the specified maximum and minimum resistances is used.
Determining the minimum drive level at the start-up of an oscillator is a challenging task, as it can only be measured in a few cases using active or passive methods due to the noise limits of measuring instruments According to IEC 60444-1, the minimum measurable level is approximately 1 nanowatt (nW) or 10 picoamperes (pA), highlighting the limitations of current measurement techniques.
(depending on the equipment, the lowest power value can be reduced to 0,1 nW or 1 àA)
A velocity v max = 0,01 m/s, corresponding to 50 àA for AT-cut crystals, has proved to be practical value for π-network measurements (see “Method A”)
In the following, two methods and one referential method of DLD measurement are described
“Method A” is based on the π-network method according to IEC 60444-1, which can be used in the complete frequency range covered by this standard It allows the fast selection of drive level sensitive quartz crystal units by a sequence of three measurements The allowed variation of the resonance resistances given in Figure 1 is based on long-term examinations of crystal units of different manufacturers and proved to be a reliable indicator for crystal units showing start-up problems If necessary, this method should also is extended by measuring a large number of different drive levels However, in practice, this is not necessary in most cases (see 5.1)
“Method B” is used for devices where strict oscillation start-up requirements have to be fulfilled and for high reliability devices
“Method C” in Annex B is an oscillator method, which is especially suitable for measuring fundamental mode crystal units in larger quantities with fixed measurement conditions
(maximum drive level, R r max) in an economical way
If the proposed measurement techniques are not sufficient in special cases, the user should have an original oscillator with slightly reduced feedback or an original filter
“Method B” is stricter than “Method A”
“Method B” is based on the π-network method or reflection method according to IEC 60444-1,
IEC 60444-5 or IEC 60444-8, which can be used in the complete frequency range covered by this standard
Recommendation: These methods can be used for all types of crystals, however:
– “Method A” is recommended for filter and oscillator crystals
– “Method B” is recommended for applications with strict start-up conditions, for high reliability and for high stability applications It is the reference method for failure analysis etc
– “Method C” in Annex B is a go/no-go measurement technique for oscillator crystals
Method A (Fast standard measurement method)
Testing at two drive levels
Here is the rewritten paragraph:Testing is conducted at both low and high drive levels, as outlined in Clause 3, with measurements of resonance frequency and resistance taken in accordance with IEC 60444-1 The test results are evaluated against tolerances of ±10% for current levels and ±20% for power levels Additionally, the test specimens undergo storage conditions, either at 105°C for at least one day followed by two hours at room temperature, or at room temperature for one week During the measurement process, the temperature is maintained at a constant level, as specified in the relevant standard.
IEC 60444-1 and IEC 60444-5) c) Measurement at low drive level (10 àA): f r = f r1 , R r =R 11 d) Measurement at high drive level (1 mA): f r = f r2 , R r = R 12 e) Measurement at low drive level (10 àA): f r = f r3 , R r =R 13 f) Calculation of γ 12 = R 1 1 /R 1 2 The value of γ 12 shall be smaller than the maximum value of γ given by the line drawn in Figure 1 (abscissa =R 12 ) g) The tolerable frequency change f r2 −f r1 shall be 5 × 10 -6 × f r1 unless otherwise specified in the detail specification h) Calculation of γ 13 = R 11 /R 13 The value of γ 13 shall be smaller than (γ + 1)/2, where the value of γ is taken from Figure 1(abscissa = R 13 ) i) The tolerable frequency change f r3 − f r1 shall be 2,5 × 10 -6 ×f r1 , unless otherwise specified in the detail specification j) The resistance value shall not exceed the maximum value given by the detail specification at any drive levels.
Testing according to specification
Testing is conducted across a range of drive levels, from low to high and back to low, as outlined in section 5.1.1 The detail specification will define these levels, including their tolerances, permissible deviations in frequency and resistance, as well as the required storage conditions.
NOTE 1 The given γ -curve was verified by results obtained over many years of experience with crystal units for many oscillator types In most cases, there will be no trouble in start-up, but in critical oscillator configurations, problems may occur As it is not possible to manufacture crystal units, which have a constant resistance at any drive level, the proposed ϒ -curve gives tolerable relations
Definition of drive level values can be agreed between manufacturer and customer
Use the nominal drive level of the detail specification as value for the high drive level For measurement at very high drive levels an additional amplifier may be required
R es is tanc e rat io γ ( γ + 1)/ 2
Figure 1 – Maximum tolerable resistance ratio γ for the drive level dependence as a function of the resistances R r2 or R r3
NOTE 2 The equation for the recommended drive level (if not otherwise specified in the data sheet) is as follows
Details can be found in Annex A of IEC 60122-2-1:1991, Amendment 1:1993 f
I q is the recommended current for oscillating state; n is the overtone, fundamental vibration mode, n = 1;
A is the electrode size in mm 2 ; f is the frequency in MHz;
Method B (Multi-level reference measurement method)
Testing is conducted at both low and high drive levels, following the guidelines of Clause 3 and measuring resonance frequency and resistance in accordance with IEC 60444-5 The acceptable tolerances are ±10% for current levels and ±20% for power levels Additionally, samples must be stored for a minimum of one day at 105 °C, followed by at least two hours at room temperature, or alternatively, stored for one week at room temperature.
The customer and manufacturer may agree to a higher temperature and longer storage duration before DLD measurement if deemed necessary It is essential to maintain a constant temperature during the measurement in accordance with IEC 60444-5 Drive levels are applied using two types of measurement units, and the application should proceed sequentially from the lowest to the highest value and then return to the lowest A clear definition of the drive level units must be established between the crystal manufacturer and the user.
1) When the unit of a drive level is mA;
Measurement drives level: from 2 àA to nominal drive level in at least 7 levels which are logarithmically scaled (Refer to the equation given under line item f))
2) When the unit of a drive level is àW;
Measurement is essential for determining drive levels, which range from 2 nW to a nominal level across at least seven logarithmically scaled increments Additionally, the maximum frequency variation across all drive levels must adhere to specified limits.
The frequency measurement values must adhere to the condition \$f_s(i) < 0.5 \times f_{NOM}\$, where \$f_s(i),_{max}\$ represents the maximum frequency measurement for drive levels ranging from \$i = 1\$ to \$2 \cdot N - 1\$, and \$f_s(i),_{min}\$ denotes the minimum frequency measurement within the same range Additionally, \$f_{ADJ}\$ specifies the tolerance for frequency adjustment Furthermore, the specifications for the maximum ratio of resistance change and the maximum resistance across drive levels must be defined accordingly.
R 1 (i) ,max is the maximum value for resistance measurement values with i = 1 to 2⋅N-1 drive levels;
R 1 (i) ,min is the minimum value for resistance measurement values with i = 1 to 2⋅N-1 drive levels;
The maximum resistance, denoted as R 1,max, is defined by the detail specification, while γ represents the resistance ratio Additionally, the N drive levels must be logarithmically scaled, following the relationship DL N+1 = DL N × K The recommended drive level equation should be used unless otherwise indicated in the data sheet.
K DL g) A larger number of drive levels may be necessary in special applications, e.g those caused by mechanical coupling with nonlinear spurious resonances (dips) and for failure analysis
Relationship between electrical drive level and mechanical displacement of quartz crystal units
The power loss of a crystal unit in watts is given by:
I is the current through the crystal unit in amperes
R 1 is the motional resistance in ohms
The reactive power is given by:
B 2π where f is the resonance frequency in hertz
C 1 is the motional capacitance in farads
The electric energy in watt seconds is given by:
The mechanical energy of a crystal unit can be represented by the following terms:
= (acceleration work) ρ = 2 650 kg/m 3 (density) where
The volume of the oscillating area is denoted as \$V\$ in cubic meters (m³) The velocity, represented as \$v\$, is calculated as the derivative of displacement with respect to time, measured in meters per second (m/s) Additionally, \$c\$ refers to the modulus of elasticity associated with the mode of vibration, specifically for AT-cut crystal units.
=c c N/m 2 ) x is ∆l/l is the elongation; s is the excursion from rest position in meters; b = d 2 s/dt 2 is the acceleration in meters per square seconds (m/s 2 ); n is the overtone order
The volume V can be calculated from the electrode area F EL and the electrode spacing d
C =ε ⋅ε ⋅ F where ε r is the relative dielectric constant of AT-cut quartz material and is equal to 4,54; ε 0 is the electric field constant and is equal to 8,86×10 − 12 F/m;
N is the frequency constant equal to f ⋅(d/n) N = 1 665 Hz⋅m for AT-cut crystal units; n is the overtone order
= ε 0 ε and the maximum current from the maximum velocities, elongations, excursions or accelerations of the mechanical vibrations: max 1 0 1 max K n C C v
For non-convex AT-cut crystal units, the following also applies:
C =γ = ⋅ where n is the overtone order
The following is obtained with C o = 5 pF for the currents:
I max,1 = 50 mA I max,2 = 1 mA v 1 = 0,01 m/s v 2 = 0,2 m/s x 1 = 1,8 × 10 -6 x 2 = 3,6 × 10 -5 at f= 10 MHz: s 1 = 6,7 × 10 -11 m s 2 = 1,3 × 10 -9 m b 1 = 2,6 × 10 5 m/s 2 b 1 = 5,3 × 10 6 m/s 2 s 1 = 6,7 × 10 -12 m s 2 = 1,3 × 10 -10 m b 1 = 2,6 × 10 6 m/s 2 b 2 =5,3 × 10 7 m/s 2
The maximum currents or levels for each type of crystal unit are determined by factors such as frequency, quality factor, vibration mode, and the volume of the vibrating zone These limits must not be exceeded when using crystal units in oscillators and filters to ensure optimal performance and device reliability.
The maximum drive level shall be selected so that with a further increase of the drive level by
50 %, the resistance does not increase reversibly by more than 10 % or the frequency changes by more than 0,5 × 10 -6
Method C: DLD measurement with oscillation circuit
Detecting the DLD effect across the entire drive level range using the method outlined in section 5.1 is costly and unsuitable as a definitive go/no-go test The proposed method efficiently tests crystal units at their maximum R_r during startup, making it applicable for both 100% final inspections and 100% incoming inspections Additionally, this method serves as a reliable tool to assess whether the crystal unit meets the specified requirements for R_rmax outlined in the detailed specifications.
The crystal unit in the oscillator can be represented as indicated in Figure B.1
There will be no oscillation when the magnitude of the −R osc of the circuit is lower than R r of the crystal unit
During start-up, the R r of the crystal unit may behave as shown in Figure B.2
When measuring the crystal unit several times, the characteristic can shift slightly to the right or to the left or it can flatten
The ratio \$\gamma = \frac{R_{r2}}{R_{r1}}\$ can vary with each measurement, and reaching a specific value of \$\gamma\$ does not imply that the oscillator will cease to function The critical factor to consider is the safety margin between the maximum \$R_{r}\$ of the crystal unit and the \$R_{osc}\$ of the oscillator circuit.
It is recommended that the circuit should have a –R osc of ≥ 3R r max because in the temperature range, the R r max as well as –R osc can shift
During the start-up, the drive level will move from the low values (left side of the graphics in
Figure B.3) to the nominal drive level
The principle of measurement is presented in Figure B.4
The test setup features a precisely engineered crystal oscillator that functions as a true negative resistance across a broad frequency range It includes a feedback network that restricts power dissipation in the crystal unit to 1 mW, ensuring efficient operation Additionally, a detector circuit with an LED provides a clear visual indication, enhancing the system's reliability and ease of monitoring.
Figure B.1 – Insertion of a quartz crystal unit in an oscillator
– loop gain > 1, which means −R osc > R r
– feedback signal at oscillator input shall have correct phase
R es onanc e re si st anc e R r ( Ω )
Figure B.2 – Crystal unit loss resistance as a function of dissipated power
NOTE The ratio R r2 /R r1 is not a reproducible value since the crystal unit curve slightly shifts at different measurement cycles
The negative resistance and the DLD reject level of the oscillator can be adjusted by adding a positive resistor in series This allows for a range of values to be achieved.
0 Ω and 200 Ω may be selected Connecting a quartz crystal unit with a sufficiently low R r value between the test clamps, the oscillation will build up starting from the initial noise level
(approximately 10 -16 W to 10 -15 W) to its limiting point for 1 mW as shown in Figure B.5
During the start-up, the R r,max of the crystal unit is continuously compared with a Calibrated-
R osc and the result is detected and transferred into a go/no-go decision
If the tested crystal unit exhibits a specific level of DLD, the oscillation amplitude may fail to reach the 1 mW threshold (point B in Figure B.6) As illustrated in Figure B.6, the oscillation build-up ceases at a significantly lower drive level.
(point A) Normally in such cases, no oscillation is observed and only with very sensitive equipment can some oscillation be detected
When a crystal unit attains the 1 mW level (point B), the LED indicator activates, indicating that the resonance resistance of the quartz crystal unit remained below the DLD reject level during startup.
This measurement method offers several benefits, including speed, ease of calibration, affordability, and a straightforward setup A comprehensive electrical diagram is provided in Figure B.7, and the necessary equipment can be purchased commercially.
R es onanc e re si st anc e R r ( Ω )
Figure B.3 – Behaviour of the R r of a quartz crystal units
Figure B.4 – Block diagram of circuit system
O sc illa to r r es is ta nc e – R os c ( Ω )
Figure B.5 – Installed − R osc in scanned drive level range
R es onanc e re si st anc e R r ( Ω )
Figure B.6 – Drive level behavior of a quartz crystal unit if − R osc = 70 Ω is used as test limit in the “Annex B” test
27 kΩ 560 Ω 150 Ω 10 nF 10 nF 10 nF 10 nF 10 nF
Figure B.7 – Principal schematic diagram of the go/no-go test circuit
IEC 60122-2-1:1991 provides a comprehensive guide on the use of quartz crystal units specifically designed for frequency control and selection This standard focuses on quartz crystal units utilized as microprocessor clock supplies, ensuring precise and stable frequency generation essential for microprocessor operation It outlines the characteristics, performance requirements, and application guidelines to optimize the reliability and accuracy of quartz crystal units in electronic devices Adhering to IEC 60122-2-1:1991 helps manufacturers and engineers achieve consistent frequency control, enhancing the overall performance of microprocessor-based systems.
3.1 Changements réversibles de la fréquence et de la résistance 24
3.2 Changements irréversibles de la fréquence et de la résistance 24
3.3 Causes des effets de la DNE 25
4 Niveaux d'excitation pour la mesure de la DNE 25
5.1 Méthode A (méthode de mesure rapide normalisée) 26
5.2 Méthode B (méthode de mesure de référence à plusieurs niveaux) 28
Annexe A (normative) Relation entre le niveau d'excitation électrique et le déplacement mécanique des résonateurs à quartz 30
Annexe B (normative) Méthode C: Mesure de la DNE avec un circuit d'oscillation 33
Figure 1 – Rapport des résistances maximales tolérables γ pour la dépendance du niveau d'excitation en fonction des résistances Rr2 ou Rr3 27
Figure B.1 – Insertion d'un résonateur à quartz dans un oscillateur 33
Figure B.2 – Résistance de perte d'un résonateur en fonction de la puissance dissipée 34
Figure B.3 – Comportement de Rr d'un résonateur à quartz 35
Figure B.5 – −Rosc installée dans une gamme de niveaux d'excitation balayés 36
Figure B.6– Comportement du niveau d'excitation d'un résonateur à quartz si
−Rosc = 70 Ω est utilisée comme limite de l'essai de l'Annexe B 36
Figure B.7 – Schéma principal du circuit d'essai tout-ou-rien 37
MESURE DES PARAMÈTRES DES RÉSONATEURS À QUARTZ –
Partie 6: Mesure de la dépendance du niveau d'excitation (DNE)