Design Note: HFDN-29.0 Rev.1; 04/08 Using the MAX3645 as a Pin-for-Pin Replacement of the Mindspeed MC2045-2/2Y Functional Diagrams Pin Configurations appear at end of data sheet. Functional Diagrams continued at end of data sheet. UCSP is a trademark of Maxim Integrated Products, Inc. LE AVAILAB Using the MAX3645 as a Pin-for-Pin Replacement of the Mindspeed MC2045-2/2Y The MAX3645 is a pin-for-pin replacement of the Mindspeed MC2045-2 and MC2045-2Y postamplifiers. Critical performance advantages of the MAX3645 include stable LOS (loss-of-signal) assert/deassert levels over the supply range, stable LOS hysteresis over the programmable levels, and better sensitivity. The MAX3645 does not require board layout modifications, but external component values need to be adjusted in order to match previous LOS assert/deassert levels and LOS assert/deassert times. This design note provides typical data that can be used to determine new external component values. 2 Replacement Design Recommended external component values are given in the MAX3645 datasheet but adjustments are required to match previous design settings. Typical LOS deassert levels of the MAX3645, MC2045-2, and MC2045-2Y with the supply set to 3.3V and 5V are given in Figures 1 and 2. LOSS-OF-SIGNAL DEASSERT LEVELS vs. RTH 70 65 DIFFERENTIAL INPUT VOLTAGE (mVpp) 1 Introduction 60 (155Mbps, 223-1 PRBS, 25ºC) MC2045-2 (3.3V) 55 50 45 40 35 30 25 MC2045-2 (5V) MAX3645 (3.3V AND 5V) 20 15 10 5 0 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 RTH (kΩ) Figure 1. MAX3645 vs. MC2045-2 LOS Deassert Levels 2.1 LOS Assert/Deassert Levels An external threshold resistor (RTH) connected from the TH pin to ground programs the LOS assert/deassert levels. The MAX3645 recommended range for RTH is 0Ω to 2kΩ, which provides a guaranteed LOS deassert range of 2mVpp to 20mVpp (differential). A big advantage of the MAX3645 is the ability to maintain the same assert/deassert levels over the supply range. The assert/deassert levels of the MC2045-2 and MC2045-2Y are also programmed with an external threshold resistor but the levels change significantly with supply voltage. Mindspeed is a trademark of Mindspeed Technologies, Inc. Design Note HFDN-29.0 (Rev.1; 04/08) Figure 2. MAX3645 vs. MC2045-2Y LOS Deassert Levels Maxim Integrated Page 2 of 4 The new RTH value that provides approximately the same deassert level with the MAX3645 as with the MC2045-2/2Y can be determined in two steps using Figures 1 and 2: 1. Use the RTH value of the previous design to find the corresponding deassert level on the appropriate Mindspeed curve. The appropriate curve depends on whether the previous design used the MC2045-2 or the MC2045-2Y, and whether the design assumed a 3.3V or 5V supply. 2. Find the same deassert level on the MAX3645 curve. Then scale down to the x-axis to find the new RTH value. 2.2 LOS Hysteresis The MAX3645 LOS hysteresis is much more stable than the MC2045-2/2Y over the programmable threshold range. The LOS hysteresis was calculated as 10log(VDEASSERT/VASSERT). Typical data points are given in Figure 3 over the RTH range of 0Ω to 2kΩ. LOSS-OF-SIGNAL HYSTERESIS vs. RTH 2.8 (155Mbps, 223-1 PRBS, 25ºC) 2.7 10LOG(VDEASSERT/VASSERT) (dB) 2.6 MC2045-2/2Y (3.3V) 2.5 2.4 2.3 MAX3645 (3.3V AND 5V) 2.2 2.1 2 1.9 1.8 1.7 MC2045-2/2Y (5V) 2.3 Input Sensitivity The MAX3645 has better sensitivity than the MC2045-2/2Y. The minimum differential input specification of the MAX3645 is 1.0mVpp whereas the MC2045-2/2Y is 1.6mVpp. Typically the MAX3645 has a BER<10-12 with input amplitudes down to 0.5mVpp (differential). The LOS assert/deassert levels of the MAX3645 can be programmed lower than the MC2045-2/2Y. For example, at the lowest possible threshold setting (RTH = 0Ω) the assert/deassert levels of the MAX3645 are approximately a factor of 2 lower than the assert/deassert levels of the MC2045-2/2Y. 2.4 LOS Assert/Deassert Times The external capacitor CSD determines the internal power detector time constant, which determines the LOS assert/deassert time. Table 1 provides typical assert/deassert times for the MAX3645 with three different capacitor values. A CSD of 1nF is recommended for the MAX3645. Table 1. Typical MAX3645 Assert/Deassert Time CSD Assert Time Deassert Time 0.1nF 3µs 3µs 1nF 21µs 21µs 10nF 222µs 222µs The MC2045-2/2Y datasheets recommend using a 10nF capacitor for CSD. A 10nF capacitor typically provides assert/deassert times of 5-6µs with the MC2045-2/2Y, but assert/deassert times of 222µs with the MAX3645. Therefore, to match the assert/deassert times of the MC2045-2/2Y, the MAX3645 requires a CSD value that is approximately 40 times smaller than the value used with the MC2045-2/2Y. 1.6 0.0 0.2 0.4 0.6 0.8 1.0 1.2 1.4 1.6 1.8 2.0 RTH (k Ω) Figure 3. LOS Hysteresis vs. RTH No significant change in hysteresis is apparent between the MC2045-2 and MC2045-2Y, but the hysteresis changes significantly with supply voltage and over the RTH range. The MAX3645 hysteresis remains stable over supply voltage with less than 0.15dB variation over the entire RTH range. Design Note HFDN-29.0 (Rev.1; 04/08) 2.5 Other External Components The external capacitor CAZ determines the time constant of the DC offset correction loop. A CAZ value of 0.1µF is recommended for the MAX3645. Typically 0.1µF capacitors are used to AC-couple the data inputs of the MAX3645. Maxim Integrated Page 3 of 4 2.6 Eye Diagrams The eye diagrams given in Table 2 were taken under typical conditions for an overall performance comparison. [VCC = 3.3V, 223-1 PRBS, VIN = 15mVpp (differential), RTH = 500Ω (MC2045-2) and 950Ω (MAX3645)]. Table 2. Eye Diagram Comparison Mindspeed MC2045-2 Maxim MAX3645 125Mbps 200Mbps Design Note HFDN-29.0 (Rev.1; 04/08) Maxim Integrated Page 4 of 4