AN1410/D Configuring and Applying the MC74HC4046A Phase-Locked Loop A versatile device for 0.1 to 16MHz frequency synchronization http://onsemi.com APPLICATION NOTE Prepared by: Cleon Petty, Gary Tharalson & Marten Smith Logic Application Engineers Abstract the following information is useful for approximating a design but, because of process, layout and other variables, there can be substantial deviation between theory and actual results. Therefore, it is highly recommended that prototypes be built and checked before committing a design to production. Typical applications for the HC4046A usually involve a configuration such as shown in Figure 1. The MC74HC4046A (hereafter designated HC4046A) phase–locked loop contains three phase comparators, a voltage–controlled oscillator (VCO) and an output amplifier. The user of this document should have a copy of the HC4046A data sheet in ON Semiconductor Data Book DL129 available for details of device operation and operating specifications. The user should also be aware that Phase Detector Ref Osc Low Pass Filter VCO fo Feedback ÷N Figure 1. Typical Phase–Locked Loop VCO/OUTPUT FREQUENCY and reworking it to obtain a formula that incorporates all the detail to fit the HC4046A. First, the charge time of the device for half–cycle time is obtained as follows: The output frequency, Fo, is calculated as a function of the Ref Osc input and the ÷N feedback counter: Fo = Ref Osc * N dt (1) The ability of the loop to emulate the above formula makes it ideal for multiplying an input frequency by any number up to the maximum of the VCO. The HC4046A VCO frequency is controlled by the equation: VCO freq = f(I * C) or, + c dV dt Semiconductor Components Industries, LLC, 2000 March, 2000 – Rev. 3 Fo + and Fo 1 2CdV I + 2dt1 I + 2CdV (4) where I and dV must be obtained for the HC4046A. There are two components that comprise the I charge for the HC4046A VCO, I1 and I2. I1 is the current that sets the frequency associated with the VCO input and is a function of R1, VCOin, and an internal current mirror that is ratioed at 120/5 ≈ 24, resulting in the equation: (2) where I is controlled by the external resistors R1 and R2 and C by external capacitor Cext . Frequency of oscillation is calculated by starting with the familiar equation: I + dV CI I1 in + VCO R1 ǒ Ǔ 120 5 (5) (3) 1 Publication Order Number: AN1410/D AN1410/D I2 is set by R2 and adds a constant current to limit the Fo min of the VCO and is a function of Vdd, R2, and an internal current mirror of ratio 23/5, resulting in the equation: I2 + ǒ Ǔǒ Ǔ 2Vdd 3R2 23 5 1. Iconstant ratio versus R1 R1 (KΩ) 3.0 5.1 9.1 12 15 30 40 51 110 300 (6) The dV of Equation ( 4 ) is determined by design to be ≈ 1/3 Vdd. Substituting this and I = I1 + I2 into Equation ( 4 ) results in: Fo ǒ Ǔ ) ǒ Ǔǒ Ǔ 2Vdd 3R2 + VCOin 120 5 R1 + VCOin (24) R1 + 3VCOin (24) R1 2Cext Vdd 3 )ǒ 2Vdd 3R2 23 5 Ǔ (4.6) Fo 2Vdd (4.6) R2 (7) 2Cext Vdd Fo + 2C ) For comparison, from Chart 14D in the HC4046A data sheet, the Fo based on measurements is approximately 270 Hz. Thus, the calculated and measured values are not too far apart taking into consideration such variables as process variation, temperature, and breadboard inaccuracies. The Cstray of a PCB layout will affect results if the Cext is not Cstray. So for Cext ≤ 1000pF, adding Cstray to the Cext fixed capacitance will result in better accuracy. The gain of a VCO is calculated by knowing fmax at VCOin max and fmin at VCOinmin and calculating the following equation: & 2Vdd (4.6) R2 ) 3 * undershoot) ) + 2C (V ext dd ) 3 * undershoot) ext (Vdd 3VCOin(Iconstant ratio) R1 9.2(Vdd) R2 (3)(1)(31) 300K + 2(0.1 * 10-6)(4.5 ) 2.1) + 235Hz It was found by experiment that when the Cext potential reaches threshold (at Vdd/3), the inversion of the charging voltage of Cext is forced below ground due to charge coupling. Therefore, the dV is not just Vdd/3 as expected and the charging time must start at a point below ground which affects t and thus, Fo. An undershoot voltage must be added to the equation for better accuracy in calculating t and Fo. This modifies Equation ( 7 ) as follows: 3VCOin (24) R1 13.5 17.5 21.5 23.0 24.0 26.5 27.0 28.5 29.0 31.0 The VCO calculation [Equation ( 8 )] becomes a bit more accurate by adjusting the VCOin and Iconstant ratio. For example, with R1 = 300K, R2 = ∞, Cext = 0.1µF, VCOin = 1.0V, Vdd = 4.5V, and Iconstant ratio = 31, Equation ( 8 ) yields: V 2Cext 3dd ) Iconstant ratio VCO gain (8) Equation ( 8 ) now contains all the factors to calculate an Fo for the HC4046A VCO. It was determined by experiment that the undershoot of the charging waveform is a function of Cext and an on–chip parasitic diode that clamps it at a maximum of –0.7V. The size of the Cext capacitor limits the voltage and was found to be near zero volts for Cstray ≈17pF ≤ Cext ≤ 30pF; the voltage increases at 6 mV/pF for a 30pF ≤ Cext ≤ 150pF range of Cext. The on–chip diode then takes over and limits the voltage to –0.7V. It was also found that the Iconstant ratio is a function of R1 and increases as R1 becomes larger. The change is attributed to saturation of the current mirror at lower value resistances, and to voltage divider problems at higher value resistances combined with the resistance of the small FET in the current mirror. Experimental data shows that Iconstant ratio follows Table 1 somewhat. The ratio goes to 25 somewhere between 9.1KΩ and 51KΩ, and for those limits, 25 should give reasonable results. In addition, these numbers seem to hold for a range of Vdd of 3.0V ≤ Vdd ≤ 6V. max * f min + VCO f max * VCOin min in + Dfreqńvolt (9) The gain of the VCO is needed to calculate a suitable loop filter for a PLL system. Fo is determined by VCOin and is clamped as a function of a % of Vdd. The clamp voltage generally follows the slope of 4%/V for Vdd changes from 3.5V ≤ Vdd ≤ 6V, starting at 56% at Vdd = 3.5V and going to 66% at Vdd = 6V. Knowing this limit point allows picking a VCOin max point a few hundred mV below it and keeps Fo in the linear range of operation. It also best to pick a VCOin min point at a level of a few hundred mV above 0V for the same reason given above. As an example, for a Cext =1100pF, R1 = 9.1K, R2 = ∞, Vdd =5.0V, and VCOin min = 0.25V, VCOin max can be determined and a gain calculated as follows. VCOin limit = (4%/V)(1.5V) + 56% = (62%)(Vdd ) = 3.1V. So, for sake of linearity, choose VCOin = 2.5V. Using Equation ( 8 ), VCOin min and VCOin max can be used to calculate Fo min and Fo max as follows: http://onsemi.com 2 AN1410/D Fo min Now choose a Cext of 200pF. Then, from above result, (3)(0.25)(21.5) 9.1K + 2(1100 * 10-12)(5 ) 2.1) + 113.4KHz Fo max R1 (3)(2.5)(21.5) 9.1K + 2(1100 * 10-12)(5 ) 2.1) + 1.3MHz This appears reasonable and there are standard values for Cext = 200pF and R1 = 27K. Using these values, Equation ( 8 ) can be adjusted according to the desired Fo min, Fo max, and Fo center. Then, using Equation ( 9 ), the VCO gain is: VCO gain 6–0.11 * 106 + 1.3 * 102.5–0.25 + 528.9KHzńV LOW PASS FILTER DESIGN The design of low pass filters is well known and the intent here is to simply show some typical examples. Reference should be made to the HC4046A Data Sheet and to Application Note AN535/D — “Phase–Locked Loop Fundamentals” (available through ON Semiconductor Literature Distribution). Some simple types of low pass filters are shown in Figure 2 and Figure 3. This gain factor will be known as Kvco in the loop filter equations. R2 is used in applications where a minimum output frequency is desired when VCOin is 0V. It is calculated at VCOin = 0V causing Equation ( 8 ) to become: Fo (Vdd) + 2C (R2) (V9.2 dd ) 3 * undershoot) The additional I2 current is a constant that adds to total charge current for Cext and increases the VCOin versus Fo curve by a theoretical constant amount. In reality, the amount of increase actually decreases at a slight rate as VCOin increases. The decrease is slight and the use of Equation ( 8 ) will give adequate accuracy for most applications. The Fmax of the HC4046A VCO was determined to be about 16MHz. Beyond 16MHz, the output logic swing tends to reduce and is therefore somewhat useless for driving a CMOS input. The VCO will operate at ≈ 28MHz but the output has a VOL≈ 2.0V and a VOH≈ 4.5V at Vdd = 5.0V. The following table was generated to make calculation of R1 and Cext a function of Fo with Vdd = 5V, VCOin = 1V, and room temperature. Use of the table allows a rough estimate of (R1)(Cext) for a given Fo. The final values can be adjusted by use of Equation ( 8 ), Table 1 for Iconstant ratio, rules for undershoot voltage, Vdd variations, and VCOin variations. The example below shows a typical calculation. ∅det Charge Pump Output Cext (pF) 3.0K ≤ R1 ≤ 9.0K ∅det Charge Pump Output 9.1K ≤ R1 ≤ 50K 50K ≤ R1 ≤ 900K 0 ≤ Cext ≤ 30 30 ≤ Cext ≤ 150 150 ≤ Cext ≤ ∞ 7.50/Fo 5.77/Fo 5.28/Fo 0 ≤ Cext ≤ 30 30 ≤ Cext ≤ 150 150 ≤ Cext ≤ ∞ 9.00/Fo 6.92/Fo 6.34/Fo VCOin C1 Figure 3. Simple Low Pass Filter B The equations for calculating loop natural frequency (wn) and damping factor (d) are as follows: For Filter A (Figure 2): d + Ǹ KøKVCO NC1R1 n + Kø0.5w KVCO where K∅ = phase detector gain, KVCO = VCO gain, and N = divide counter. For Filter B (Figure 3): wn Assume a desired value of Fo of 1MHz. From 2, choose an R1 range of 9.1K ≤ R1 ≤ 50K and a Cext range of > 150pF; this condition leads to (R1)(Cext) = 5.28/Fo. Thus, (R1) (Cext) R1 R2 (R1)(Cext) 5.40/Fo 4.15/Fo 3.80/Fo VCOin Figure 2. Simple Low Pass Filter A wn 0 ≤ Cext ≤ 30 30 ≤ Cext ≤ 150 150 ≤ Cext ≤ ∞ R1 C1 2. (R1)(Cext) versus Fo R1 (Ω) 5.28 * 10-6 + 26K + 200 * 10-12 d + 15.28 + 5.28 * 10-6 * 106 http://onsemi.com 3 + Ǹ KøKVCO NC1(R1 R2) ) + 0.5wn(R2C1 ) KøKNVCO) ( 10 ) AN1410/D Figure 4 shows an active filter using an op amp from Application Note AN535/D. R1 ∅det (K∅) Ref Osc Fref LP Filter wn KVCO fo C1 Feedback ÷N R1 ∅det Charge Pump Output OP AMP VCOin Figure 5. Parametized PLL To determine N, use equation (1) for Fo min = 200KHz, and Fo max = 2MHz resulting in the following: Figure 4. Op Amp Filter Ǹ N min = 200/100 = 2, and N max = 2000/100 = 20 The results so far indicate the following starting parameters: For Figure 4, the equations become: wn d + KøKVCO NC1R1 ( 11 ) R2 + Kø2wKVCO nNR1 A. A VCO with a 10:1 range is required B. wn = Fref/10 = 10KHz C. d = 0.707 D. R2 = ∞ E. Vdd = 5.0V ( 12 ) + wnC21R2 , where Op Amp gain is large The Fo center frequency ≈ F max From the above equations, it is possible to design a suitable filter to meet the needs of many PLL applications. The inclusion of R2 in the equations for Figure 3 and Figure 4 permits the capability to change wn and d separately while Figure 2 equations do not. Normally, a design is easier if wn and d can be chosen independently. Both factors affect the loop acquisition time and stability. A good starting value for d is 0.707 and Fref/10 for wn. Manipulation of the equations allows calculation of R1, R2, and C1 from the other measured, calculated, or picked parameters. For example, R1 ) R2 + KNCøKVCO w 2 ( 13 ) R2 * N + C2d 1wn C1(KøKVCO) ( 14 ) C1 + Nw K2ø(RKVCO )R C1 * N + R2d 2wn R2(KøKVCO) 2 1 2 Recalling that the clamp voltage % at Vdd = 5V is about 62, then Fmax VCOin limit = (0.62)(5) = 3.1V, but as described earlier, this needs to be reduced by a factor to bring it into linearity (≈ 350mV) so the final Fmax VCOin limit = 2.75V. For the Fmin VCOin limit pick 0.25V. This results in a center frequency VCOin of: Center freq VCOin + 2.75 *2 0.25 + 1.25V From 2, for picked values of 9.1K≤R1≤50K and 30≤Cext ≤150, obtain an estimate for (R1)(Cext) of 5.77/Fo. Thus, at the Fo center frequency, 1 n n ) F min + 2.0 ) 0.2 + 1.1MHz (R1)(Cext) + 1.15.77 + 5.245 * 10-6 * 106 Now, a reasonable starting point is established for setting the values of the loop filter and the VCO range. Choosing R1 = 9.1K, Cext becomes , or alternatively, 2) Cext + 5.2459.1K* 10-6 + 576pF WHOOPS! This value, 576pF, is outside of the original picked range for Cext; therefore, we need to go back and pick a larger value of R1, e.g., 42K should be sufficient. Then Cext becomes Usually, C1, wn, and d are picked and the remaining parameters calculated. DESIGN EXAMPLE Cext The goal is to design a phase–locked loop that has an Fref of 100KHz, an output Fo of 1MHz center frequency, and the ability to move from 200KHz to 2MHz in 100KHz steps. + 5.24542K* 10-6 + 125pF and now both R1 and Cext are within selected ranges. http://onsemi.com 4 AN1410/D The final values used for the desired frequency range are R1 = 42kΩ, Cext = 175pF, R2 = , VCOin max = 2.75V, and VCOin min = 0.25V. The next step is to determine the loop filter. Choosing a filter like the one in Figure 3, calculate the component as follows: Now calculate Fmax and Fmin using Equation ( 8 ) with R1 = 42kΩ, R2 = , Vdd = 5.0V, Iconstant ratio = 27 (from 1. and R1 = 42kΩ), Vundershoot = 0.57V (calculated from 6pF/mV (125pF–30pF) = 0.57V), VCOin min = 0.25V, and VCOin max = 2.75V: 1 1 0 Fo min ) R + (2)(125 * 10-12f) [5.0V ) 3(0.57V)] (3)(0.25)(27) 42K (9.2) (5.0) 20.25 + 70.455 + 287.4KHz * 10-6 Fo max (3)(2.75)(27) 42K ) 0 + 222.75 70.455 * 10-6 ) 3(0.57V)] Fo min + 10KHz * 2p + 62.83 * 103radńsec + 100KHz 10 where K∅ = phase detector gain Vdd = output swing Fmax is > the required 2.0MHz, but the Fmin is not low enough for required application. It is necessary to adjust either Cext or R1 to achieve required specification of 0.2 to 2.0MHz Fo. Since R1 = 42kΩ is a standard resistor value, try adjusting Cext to a higher value, such as 175pF. Because Cext is now > 150pF, the Vundershoot must be adjusted to 0.7V, as per earlier explanation: So, ) wn N = 2 to 20 + 3.16MHz Choose C1 to be 0.01µF, N = 10 for approximate mid–range Fo, and calculate R1 and R2 using Equations ( 13 ) and ( 14 ): R1 R2 ) + 2250.52–514.4 + 1736W 0 R (9.2) (5.0) (2)(175 * 10-12f) [5.0V Then, R1 = 4924.5 – 1736 = 3188.5Ω. Since N is changeable, it is a good idea to check min and max on wn and d. For more information on why, see Application Note AN535/D or the MC4044 Data Sheet in the MECL Data Book DL122/D. The following examples show sample calculations for N = 2 and 20. For N = 20, use Equation ( 10 ) to calculate wn and d: ) 3(0.7V)] 222.75 + 104.37 + 2.13MHz * 10-6 These values are adequate for the specified application. The next item to determine is the VCO gain factor, KVCO, using Equation ( 9 ): * f min * VCOin min KVCO + KVCO * 106 * 0.194 * 106 + 774.4KHzńV + 2.13 2.75V * 0.25V f max VCOin max + C2d * N 1wn C1(KøKVCO) 10 + (0.01 (2)(0.707) – * 10-6) (62830) (0.01 * 10-6)(0.4)(4.86 * 106) and (3)(2.75)(27) 42K 1 n 0 (9.2) (5.0) 20.25 + 104.37 + 194.02KHz * 10-6 Fo max (0.4)(4.86 * 106) ) R2 + KNCøKVCO + 2 w (10)(0.01 * 10-6)(62.83 * 103)2 * 106 + 4924.5W + 1.944 394.76 + (2)(175 * 10-12f) [5.0VR) 3(0.7V)] (3)(0.25)(27) 42K + V4ddp + 5.0 + 0.4Vńrad 4p d = 0.707 (for starters), and R (9.2) (5.0) (2)(125 * 10-12f) [5.0V Kø wn min + + Ǹ KøKVCO NC1(R1 R2) ) Ǹ (0.4)(4.86 * 106) (20)(0.01 * 10-6)(3188.5 ) 1736) + 44.43 * 103radńsec, or + 44.43 * 103radńsec [ 7KHz or in radians + (2p) (774.4 * 103) + 4.86 * 106RadńsecńV 2p http://onsemi.com 5 AN1410/D and + (0.5)(wn) d min + ƪ R2C1 ƪ ) KøKNVCO and ƫ d max (0.5)(44.43 * 103) * (1736)(0.01 * 10-6) 20 ) (0.4)(4.86 * 106) (1736)(0.01 * 10-6) ƫ wn max + Ǹ 2 ) (0.4)(4.86 * 106) ƫ + 1.292 This shows the effect of changing n on loop performance and for this application is adequate. If the components are not what is desired, choosing a different wn and/or d allows them to be modified. Alternatively, picking different C, R1 or R2 and recalculating the other parameters can be done. If the filter does not provide adequate performance, making wn smaller or d larger may improve stability. + 0.6144 For N = 2: ƪ + (0.5)(140.49 * 103) * (0.4)(4.86 * 106) (2)(0.01 * 10-6)(3188.5 1736) ) + 140.49 * 103radńsec, or + 140.49 * 103radńsec + 22.36KHz 2p http://onsemi.com 6 AN1410/D Notes http://onsemi.com 7 AN1410/D ON Semiconductor and are trademarks of Semiconductor Components Industries, LLC (SCILLC). SCILLC reserves the right to make changes without further notice to any products herein. SCILLC makes no warranty, representation or guarantee regarding the suitability of its products for any particular purpose, nor does SCILLC assume any liability arising out of the application or use of any product or circuit, and specifically disclaims any and all liability, including without limitation special, consequential or incidental damages. “Typical” parameters which may be provided in SCILLC data sheets and/or specifications can and do vary in different applications and actual performance may vary over time. All operating parameters, including “Typicals” must be validated for each customer application by customer’s technical experts. SCILLC does not convey any license under its patent rights nor the rights of others. SCILLC products are not designed, intended, or authorized for use as components in systems intended for surgical implant into the body, or other applications intended to support or sustain life, or for any other application in which the failure of the SCILLC product could create a situation where personal injury or death may occur. Should Buyer purchase or use SCILLC products for any such unintended or unauthorized application, Buyer shall indemnify and hold SCILLC and its officers, employees, subsidiaries, affiliates, and distributors harmless against all claims, costs, damages, and expenses, and reasonable attorney fees arising out of, directly or indirectly, any claim of personal injury or death associated with such unintended or unauthorized use, even if such claim alleges that SCILLC was negligent regarding the design or manufacture of the part. SCILLC is an Equal Opportunity/Affirmative Action Employer. PUBLICATION ORDERING INFORMATION NORTH AMERICA Literature Fulfillment: Literature Distribution Center for ON Semiconductor P.O. Box 5163, Denver, Colorado 80217 USA Phone: 303–675–2175 or 800–344–3860 Toll Free USA/Canada Fax: 303–675–2176 or 800–344–3867 Toll Free USA/Canada Email: [email protected] Fax Response Line: 303–675–2167 or 800–344–3810 Toll Free USA/Canada N. American Technical Support: 800–282–9855 Toll Free USA/Canada EUROPE: LDC for ON Semiconductor – European Support German Phone: (+1) 303–308–7140 (M–F 1:00pm to 5:00pm Munich Time) Email: ONlit–[email protected] French Phone: (+1) 303–308–7141 (M–F 1:00pm to 5:00pm Toulouse Time) Email: ONlit–[email protected] English Phone: (+1) 303–308–7142 (M–F 12:00pm to 5:00pm UK Time) Email: [email protected] EUROPEAN TOLL–FREE ACCESS*: 00–800–4422–3781 *Available from Germany, France, Italy, England, Ireland CENTRAL/SOUTH AMERICA: Spanish Phone: 303–308–7143 (Mon–Fri 8:00am to 5:00pm MST) Email: ONlit–[email protected] ASIA/PACIFIC: LDC for ON Semiconductor – Asia Support Phone: 303–675–2121 (Tue–Fri 9:00am to 1:00pm, Hong Kong Time) Toll Free from Hong Kong & Singapore: 001–800–4422–3781 Email: ONlit–[email protected] JAPAN: ON Semiconductor, Japan Customer Focus Center 4–32–1 Nishi–Gotanda, Shinagawa–ku, Tokyo, Japan 141–8549 Phone: 81–3–5740–2745 Email: [email protected] ON Semiconductor Website: http://onsemi.com For additional information, please contact your local Sales Representative. http://onsemi.com 8 AN1410/D