View MC10H106_1085312.PDF datasheet online --- IC-ON-LINE

Datasheet File OCR Text:

? semiconductor components industries, llc, 2004 july, 2004 ? rev. 5 1 publication order number: and8020/d and8020/d termination of ecl logic devices with ef (emitter follower) output structure prepared by: paul shockman on semiconductor logic applications engineering contents of application note introduction ? dc termination analysis vt r t r t r t r t r t r t vt1 vt2 external internal near (standard pair) far (standard pair) far (standard pair) v ee r t r t r t r t r t vt1 vt2 v ee v to (open) v ee (shorted) v tt near (standard pair) v tt v tt r e v ee section 2. parallel termination ? external and internal section 1. unterminated lines r r r r section 3. thevenin equivalent/parallel termination r r section 4. series (back) termination v bb v bb driver receiver *all media * * d1 d2 d1 d2 section 5. diode termination r r r r r v cc v bb section 6. capacitive coupling r e r e r e r e r e r e r e r e application note http://onsemi.com
and8020/d http://onsemi.com 2 introduction static dc termination analysis a standard emitter coupled logic (ecl) output driver typically uses a current switching differential with an emitter follower for level shifting the output and the internal cml levels to familiar ecl levels. this output driver architecture presents about 6?8 internal impedance in both low and high states when properly current biased. this results in a typical v pp signal of 800 mv pp (measured single?endedly on each line) swinging around a dc voltage point of v cc ? 1.3 v when properly terminated and operating correctly as shown in figure 1. q v ee v cc d v cc v ee driver receiver 8 internal output impedance v ee figure 1. typical ecl output with emitter follower output structure, typical termination, and typical ecl input interconnect q d r e r e for proper static and dynamic operation, the output emitter follower transistor must remain in the active region of operation which requires an external resistive path be provided from the output pin to a voltage more negative than worst case v ol , such as v ee . the resistor, r e , is considered a current bias for the emitter follower output structure. when properly terminated and current biased (loaded), the outputs will generate both: (1) static state voltage levels v ol (low) or v oh (high) and (2) a dynamic transition edge (tr or tf) between state levels. static state voltage levels figure 2 illustrates the typical relationship of static signal levels and dynamic transition edges between an output driver signal and a receiver input signal. both outputs of a differential driver should always be terminated and loaded as identically as possible to preserve minimum skew and jitter operation of the device. v ih v oh v il v cc v ol v ee figure 2. state levels v oh , v ol , and dynamic transitions at q or q and d or d v cc ?1.3 v tr tf v ih v oh v il v cc v ol v ee v cc ?1.3 v tr tf output driver signal input receiver signal
and8020/d http://onsemi.com 3 output open, short, and safe dc current left open, an output will only swing a few millivolts due to parasitic aminimum currento leakage paths. shorted to v ee , a maximum current will develop, limited only by the output transistor 8 impedance, and may cause damage to the output. worst case short circuit current risks destruction of the devices . i sc v oh r int 4v 8 = 500 ma! (eq. 1) where: v oh = 4.0 v v cc = 5.0 v v ee = 0.0 v r int =8 the continuous safe output current, i out (continuous), maximum limit is 50 ma under all spec operating conditions. the continuous safe repetitive surge, i out (surge), maximum current limit is 100 ma for 10 milliseconds per second duty cycle, provided the device's total thermal limits are observed. output current polarity will always be sinking into the termination scheme during proper operation. static analysis of termination resistor r e the output continuous safe current limit, i out (cont), determines r e minimum dc termination scheme resistance to v ee although this will not provide a practical ac signal termination as shown in table a: minimum r e values. r e v oh i max (eq. 2) table a. minimum r e values line v oh r e (min) pecl 4.0 v 80 lvpecl 2.4 v 48 lvep pecl 1.6 v 32 a dc terminating resistor minimum, r e (min), of 80 , while sufficiently limiting the output load current to v ee , may generate insufficient pecl output low and high state transitions. the r e maximum is effectively determined by the application load capacitance, c l , since an rc network is formed by r e and c l which limits the signal fall time, discharging the line to the low state voltage level. a sufficiently high value r e or c l can cause the signal fall time to the v ol level to violate specification limits. designed r e or c l values may selectively eliminate undesirable noise. dynamic analysis of termination resistor r e the dynamic function of the termination resistor, r e is to develop the voltage change, v, during a high?to?low or low?to?high transition and present this to the transmission medium such as coax, twisted pair, microstrip or stripline. the v signal propagates to the receiver and is either reflected, dissipated, or a combination. since the reflection coefficient at the load is of opposite polarity to that of the source, a reflection will travel back and forth over the transmission changing polarity after each reflection until critically damped by line impedance. thus, steps may appear in the signal v at the receiving gate input due to impedance mismatch and consequent partial reflections. when r e is too large, steps appear in the trailing edge of the propagating signal, v, at the input to the receiving gate, slowing the edge speed and increasing the net propagation delay. a reasonable negative?going signal swing at the input of the receiving gate results when the value of r e is selected to produce an initial step of 75% of the expected v, or a 600 mv step for an 800 mv signal at the driving gate. for a rsecl expected v swing of 400, a 300 mv initial step is desired. hence for a 600 mv initial step: (v oh v ee ) (r t z 0 ) *z 0 0.6 i(init) * z 0 0.6 (eq. 3) the value for r e is found in table b: recommended values of r e in dynamic functional application. this table lists recommended r e values for the various ecl devices by family series according to the equation above. the table assumes operation with various data sheet v oh values and various v cc values driving a z 0 = 50 line. lowering the value of r e will increase the voltage change, v, launched into the transmission media. raising the value of r e will decrease the voltage change, v, launched into the transmission media. table b. recommended nominal values of r e in dynamic functional application series |v cc ?v ee | r e ( nb 2.5 140 nb 3.3 250 10/100lvep 2.5 50 10/100ep, 100lvel 3.3 120 10/100el, 10/100e 5.0 235
and8020/d http://onsemi.com 4 section 1. unterminated lines r from transmission line theory, when the driver r e develops a v swing, the signal propagates from point a arriving at point b at time td later as shown in figure 3. this configuration is also referred to as a stub or an open line. figure 3. unterminated transmission line stub v ee t?line z 0 ab td r e at point b, the signal is reflected as a function of l . if the input impedance of the receiving gate is large relative to the line characteristic impedance, according to equation 4: l (r l z 0 ) (r l z 0 ) (eq. 4) where: l = load reflection coefficient r l = load impedance z 0 = line characteristic impedance a large positive reflection occurs resulting in overshoot. the reflected signal reaches point a at time 2td , and a large negative reflection results because the output impedance of the driver gate is much less than the line characteristic impedance (i.e. r o << z 0 ). when the reflected signal arrives at the source it is reflected back toward the load with a magnitude dictated by the source reflection coefficient: s (r s z 0 ) (r s z 0 ) (eq. 5) where: s = source reflection coefficient r l = source impedance z 0 = line characteristic impedance the reflected signal continues to be reflected by the source and load impedances and is attenuated with each passage over the transmission line. the output response appears as a damped oscillation asymptotically approaching a steady state value. this phenomena is often referred to as aringing.o the importance of minimizing the reflected signals lies in their adverse affect on noise margin and the potential for driving the input transistors of the succeeding stage into saturation. both of these phenomena can lead to less than ideal system performance. to maximize signal integrity on transmission lines, four basic techniques are available: 1. minimizing interconnect line lengths (section 1) 2. parallel termination (sections 2 and 3) 3. series termination (section 4) 4. diode termination (section 5) interconnect line lengths the output signal waveform rise (tr) and fall (tf) time are measured from the 20% and 80% levels of the static signal levels. this edge rate represents the waveforms highest harmonic and determines the maximum unterminated open line trace length, l max , permissible without sustaining signal reflections. the impetus in restricting interconnect lengths, l, is to mitigate the effects of overshoot and undershoot. a handy rule of thumb is that the undershoot can be limited to less than 15% of the logic swing if the two way line delay is less than the rise time of the pulse. with an undershoot of <15%, the physics of the situation will result in an overshoot which will not cause saturation problems at the receiving input. thus, the maximum line length can be determined: l max t r 2*t pd (eq. 6) where: l max = maximum open line length t r = signal rise time t pd = length pulse delay per unit length further, the propagation delay increases with gate loading; thus, the effective delay per unit length (t pdeff ) is given as: t pdeff t pd 1 c d l*c o (eq. 7) where: t pd = length pulse delay per unit length c d = distributed capacitance c o = capacitance per unit length (foot) l = line length using the effective delay per unit length, tpd eff, yields: t r (2) (l) (t pd ) 1 c d l*c o (eq. 8) solving for l max line length produces: l max 0.5
c d c o 2
t r t pd 2 c d c o (eq. 9) where: l max = line length maximum c d = distributed capacitance c o = capacitance per unit length (foot) t pd = length pulse delay per unit length assuming a worst case capacitance of 2 pf and a rise time of 100 ps for ep gives a value of 0.03 inch for the maximum open line length. maximum open line lengths derived from spice simulations for single and double gate loads, a maximum overshoot of 40% and undershoot of 20% was assumed. the simulation results indicate that for a 50 line, a stub length of 0.03 inches will limit the overshoot to less than 40%, and the undershoot to within 20% of the logic swing. signal traces will most assuredly be larger than 0.03 inch for most practical applications. therefore, it will be necessary to use controlled impedance environments for ep devices in general and devices with faster edges.
and8020/d http://onsemi.com 5 termination of ecl logic devices section 2. parallel termination ? external and internal r t r t external internal near (standard pair) far (standard pair) near (standard pair) far (standard pair) v ee r t v tt vt r t r t r t r t vt1 vt2 r t r t r t r t vt1 vt2 v ee v to (open) v ee (shorted) v tt v tt r e r e r e r e r e r e parallel termination advantages: ? method of choice for best circuit performance ? particularly excellent for driving distributed loads ? undistorted waveform along the full length of the line ? decreased power consumption. far dc current return ? v tt a parallel terminated line is one in which the receiving end is signal terminated internally or externally (usually to a voltage v tt ) through a resistor (r t ) with a value equal to the line characteristic impedance (figure 4). this line also carries the biasing current for the drivers output far from the driver. output current and power dissipation is decreased due to use of a v tt termination supply. the v tt supply must sustain the emitter follower output transistor in its active operating region under all operating conditions. a minimum continuous current occurs for the most negative v ol , therefore the v tt supply must remain more negative than the worst case v olmin and always sink current. standard v tt is 2.0 v below v cc supply. a parallel resistor, r t , matching the controlled impedance transmission line, z 0 , connects the signal to the v tt supply. the parallel termination to v tt is shown in figure 4. the termination resistors may be internal or external and either ganged into a combo pin or offered as singulated pins. some devices may have each internal resistors independently pinned out, allowing further termination versatility. figure 4. parallel termination to v tt ? differential and single?ended with combo or singulated vt pins (far return) v tt r t = z 0 v tt = v cc ?2.0 v t?line z 0 v tt (*or twisted pair) driver receiver *t?line z 0 *t?line z 0 v tt r r (*or twisted pair) driver receiver *t?line z 0 *t?line z 0 vt driver receiver external (far, diff.) internal termination combo pin (far, diff.) v tt r r (*or twisted pair) driver receiver *t?line z 0 *t?line z 0 vt1 internal termination singulated pins (far, diff.) r t r t r t vt2 external (far, s.e.)
and8020/d http://onsemi.com 6 internal termination resistors internal termination conveniently uses 50 values for r t , with the most popular being z 0 . note the internal termination allows the combo pin node, v t , from the internal resistors to be connected to an external v tt supply, typically at v cc ? 2.0 v, as shown in figure 5. alternatively, this combo pin may be pulled to v ee through an external resistor to form a ayo type termination variant, as shown in figure 5. see the ay varianceo topic and the ay term tableo for r t3 resistor values. figure 5. combo pin v tt or ayo connection with internal parallel termination v tt r r (*or twisted pair) driver receiver *t?line z 0 *t?line z 0 v tt connection v ee r r (*or twisted pair) driver receiver *t?line z 0 *t?line z 0 y connection r t3 a. b. example calculations ideally, v tt supply tracks 1:1 with v cc ; however, supply tolerances need to be considered. assume for instance a mc10ep16, +85 c, nominal +3.3 v cc , terminated 50 (r t ) to v tt , where v tt is v cc ? 2.0 v, or 1.3 v: i ohmax of (v cc ) 0.885 v i olmin of (v cc ) 1.685 v resulting in the nominal case: i ohmax (v ohmax v tt ) r t (3.3 0.885) 1.3 50 22.3 ma i olmin (v olmin v tt ) r t (3.3 1.685) 1.3 50 6.3 ma if +5% tolerances are assumed, two worst case conditions result. case #1: v cc min = v cc ? 5%, v tt max = v tt + 5% ((3.135 0.885) 1.365) 50 17.7 ma i ohmax (v ohmax v tt ) r t i olmin (v olmin v tt) r t ((3.135 1.685) 1.365) 50 1.7 ma case #2: v ccmin + 5%, v ttmax ? 5% ((3.465 0.885) 1.235) 50 26.9 ma i ohmax (v ohmax v tt) r t ((3.465 1.685) 1.235) 50 1.09 ma i olmin (v olmin v tt) r t y variance the ayo termination for a differential pair may be preferred when avoiding the use of a v tt supply. the design is shown in figure 6 and utilizes the following formulas for calculating resistor values which are found in the y term table. the voltage at the node where r t1 , r t2 , and r t3 connect remains at a static v tt voltage of v cc ? 2.0 v, or 1.3 v. r t3 r t1
v tt v ee v oh v ol 2v tt r t1 r t2 z 0 (eq. 10) (eq. 11) v tt r t3 (v oh v ol ) (r t1 *v ee ) r t1 2r t3 (eq. 12) figure 6. ayo variance driver receiver *t?line z 0 *t?line z 0 v cc * or twisted pair r t1 r t2 r t3 c 1 0.1?0.01 f
and8020/d http://onsemi.com 7 table c. y term table |v cc ?v ee | = 5.0 v |v cc ?v ee | = 3.3 v |v cc ?v ee | = 2.5 v z 0 r t1 r t2 r t3 z 0 r t1 r t2 r t3 z 0 r t1 r t2 r t3 50 50 50 112 50 50 50 46 50 50 50 21.2 70 70 70 156 70 70 70 64 70 70 70 29.7 75 75 75 166 75 75 75 68 75 75 75 31.8 80 80 80 179 80 80 80 72 80 80 80 33.9 90 90 90 201 90 90 90 82 90 90 90 38.1 100 100 100 223 100 100 100 91 100 100 100 42.4 120 120 120 268 120 120 120 109 120 120 120 50.8 150 150 150 335 150 150 150 136 150 150 150 63.6 v ee r e r e figure 7. standard pair with external parallel r t (*or twisted pair) *t?line z 0 *t?line z 0 near dc current return ? standard pair termination the standard pair termination scheme uses a pull?down resistor, r e , located at each driver pin to return the output transistor bias current near the driver, and an impedance matching parallel resistor, r t , located at the receiver input pins (see figure 7, standard pair with external parallel, and figure 8, standard pair termination with internal termination, and figure 9, standard pair termination with singulated internal termination resistors). the impedance matching parallel resistor may be internal or external depending on the receiver device. if internal to the receiver, the resistor may be singulated or combined (acomboo) for external pinout. the diagram of figure 7 shows a standard pair termination with an r e resistor for dc output current bias located nearby each driver pin: refer to table b, for values of r e . the differential transmission line ac impedance matching resistance, r t , is located externally near the receiver input pins. as a variation of a standard pair termination, a receiver may provide the differential transmission line ac impedance matching resistance, r t , internally. this internal impedance matching termination may be pinned out either combined into a combo v t pin or each resistor may be singulated and pinned out, such as v t1 and v t2 . when left open, the combo pin still provides a passive 100 termination across the nearby receiver's differential signal line pair. this can compliment a pull?down resistor, r e , located on each line of a differential at the driver pins. this is illustrated in figure 8. figure 8. standard pair termination with internal termination open v t pin r r (*or twisted pair) driver receiver *t?line z 0 *t?line z 0 internal termination combo pin v ee r e r e when the internal termination resistors are singulated, the two v t pins must be shorted to create the 100 value as shown in figure 9. figure 9. standard pair termination with singulated internal termination resistors internal termination singulated pins r r (*or twisted pair) driver receiver *t?line z 0 *t?line z 0 v ee vt1 vt2 shorted r e r e
and8020/d http://onsemi.com 8 internal 100 termination (lvds) for some technologies, such as lvds, this passive 100 internal termination can provide sufficient termination for the driver as shown in figure 10. devices with a combo pin will require this pin to remain open, while devices with singulated internal resistors require the two pinned out v t nodes for a differential pair to be shorted together to provide the 100 termination. figure 10. lvds interconnect with internal termination open v t pin r r (*or twisted pair) lvds driver receiver *t?line z 0 *t?line z 0 internal termination combo pin internal termination singulated pins r r (*or twisted pair) lvds driver receiver *t?line z 0 *t?line z 0 vt1 vt2 shorted differential ecl outputs can be terminated as independent complimentary single?ended lines. both sides of any differential pair must be terminated as identically as possible to minimize phase error and pulse width duty cycle skew. the i oh currents in these two cases will vary the dc v oh levels by 40 mv. however in the vast majority of cases, dc levels are well centered in their specification windows, thus this variation will simply move the level within the valid specification window and no loss of worst case noise margin will be seen. the i ol situation on the other hand does pose a potential ac problem. in the worst case #1 i olmin situation, the output emitter follower could move into the cutoff state (0 ma). the output emitter followers of ecl devices are designed to be in the conducting, active region of operation at all times. when forced into cutoff, the delay of the device will be increased due to the extra time required to pull the output emitter follower out of the cutoff state. again, this situation will arise only under a number of simultaneous worst case situations and therefore, is highly unlikely to occur. but, because of the potential, it should not be overlooked. output drive characteristics figure 11 shows the nominal output characteristics for ecl devices operating in negative ecl mode, driving various load impedances (including the standard 50 ) returned to a negative two volt supply. the output resistances, rh (high state output resistance) and rl (low state output resistance), are obtained from the reciprocal of the slope at the desired operating point. many applications require loads other than 50 ? the resulting v oh and v ol levels can be estimated using the following technique. v oh ?2.0 0 ?5 ?10 ?15 ?0.75 ?1.0 ?1.25 ?20 ?25 ?30 ?35 ?40 ?1.75 ?0.5 ?0.25 0 output voltage (v) output current (ma) slope = 6 ? 8 ?1.5 figure 11. normal output levels driving various load impedances v ol t a =25 c 25 to ? 2.0 v 50 to ? 2.0 v 150 to ? 2.0 v 100 to ? 2.0 v
and8020/d http://onsemi.com 9 section 3. thevenin equivalent parallel termination r r r r although the single resistor termination to v tt conserves power, it requires an additional supply voltage. an alternate approach to using a v tt power supply is to use a resistor divider network as shown in figure 12 to develop a thevenin voltage, v tt , and provide a parallel impedance matching ac termination, the thevenin parallel termination. figure 12. thevenin equivalent parallel termination v ee t?line z 0 driver r2 r1 v cc receiver r2 r2 r1 r1 driver receiver t?line z 0 t?line z 0 v ee v cc or twisted pair v tt v cc 2.0v v cc
r2 r1 r2 * * * * * * * (eq. 13) differential ecl outputs can be terminated as independent complimentary single?ended lines. both sides of a differential pair must be terminated. balanced, symmetrical loading of each line must be preserved. while a thevenin parallel technique dissipates more termination power, it does not require the additional v tt supply. this additional power is consumed entirely in the external resistor divider network and thus will not change the current being sourced by the device, hence it does not alter the ic reliability or lifetime. as with standard parallel termination, variance of v tt and v cc supplies must be considered. the thevenin equivalent of the two resistors needs to be equal to the characteristic impedance of the signal transmission line. calculated values for resistors r1 and r2 may be obtained from the following relationships. r2 z 0
v cc v ee v cc v tt (eq. 14) r1 r2
v cc v tt v tt v ee (eq. 15) where: v tt =v cc ? 2.0 v z 0 = characteristic impedance of the signal transmission line for a typical v cc = 5.0 v pecl scheme, where v ee = gnd, v tt = 3.0 v, and z 0 = 50 : r2 50
5 0 5 3 125 (eq. 16) r1 125
5 3 3 0 83.3 (eq. 17) and cross?checking for v tt : v tt 5
125 125 83.3 3.0 v (eq. 18) v tt v cc 2.0 v 3.0 v (eq. 19) for the typical v cc = 3.3 v lvpecl scheme, where v ee = gnd, v tt = 1.3 v, and z 0 = 50 : r2 50
3.3 0 3.3 1.3 82.5 (eq. 20) r1 82.5
3.3 1.3 1.3 0 126 (eq. 21) and cross?checking for v tt : v tt 3.3
82.5 126 82.5 1.3 v (eq. 22) v tt v cc 2.0 v 1.3 v (eq. 23) table d. thevenin term table |v cc ?v ee | = 5.0 v |v cc ?v ee | = 3.3 v |v cc ?v ee | = 2.5 v z 0 r1 r2 z 0 r1 r2 z 0 r1 r2 50 83 125 50 127 83 50 250 62.5 70 117 175 70 178 115 70 350 87.5 75 125 188 75 190 123 75 375 93.8 80 133 200 80 203 132 80 400 100 90 150 225 90 229 149 90 450 112.5 100 167 250 100 253 165 100 500 125.5 120 200 300 120 305 198 120 600 150 150 250 375 150 381 248 150 750 187.5
and8020/d http://onsemi.com 10 because the resistor divider network of r1 and r2 is used to generate v tt , the variation in v tt will be intimately tied to the variation in v cc . differentiating the equation for v tt with respect to v cc yields: v tt v cc r2 (r1 r2 ) v cc (eq. 24) for the nominal case, this equation reduces to: v tt 0.6 v cc (eq. 25) if v cc = 5% = 0.25 v, then v tt = 0.15 v. as mentioned previously, the real potential for problems will be if the v ol level can potentially put the output emitter follower out of the active operating region and into cutoff. because of the relationship between the v cc and v tt levels, the only cutoff risk condition occurs at v ccmin , the lowest value of v cc . applying the equation for i olmin under this ?5% v cc condition yields: i olmin (v olmin v tt ) r t (eq. 26) i olmin (4.75 1.85) 2.85 50 1.0 ma (eq. 27) the results of this cutoff risk analysis show there is no potential for the output emitter follower to be in cutoff. this would indicate a thevenin equivalent termination scheme is more robust to variation in v cc . since the designer has the flexibility of choosing the v tt level via the selection of the r1 and r2 resistors, the following procedure can be used. at ?5% minimal variation case for v cc : v cc = 4.75 v v tt = v cc ? 2.0 v = 2.75 v r2 = 119 r1 = 86 thus: i ohmax = 23 ma i olmin = 3.0 ma at +5% minimal variation case for v cc : v cc = 5.25 v v tt = 3.05 v thus: i ohmax = 28 ma i olmin = 5.2 ma although the output currents are slightly higher than nominal, the elimination of emitter follower cutoff risk is well justified. when the equivalent termination resistance matches the line impedance, no reflection occurs because all the energy in the signal is dissipated by the termination. hence, in comparing properly terminated schemes parallel and thevenin, a primary consideration is the power supply requirements. as mentioned earlier, the parallel v tt scheme requires an extra power supply; however, the thevenin termination dissipates 10 times more dc power. fortunately, this extra power dissipation cannot be seen on the die; therefore, either technique results in similar die junction temperatures.
and8020/d http://onsemi.com 11 section 4. series termination r r r r series damping is a technique in which a termination resistance is placed between the driver and the transmission line with no termination resistance placed at the receiving end of the line (figure 13). *t?line z 0 *t?line z 0 *t?line z 0 driver r t v ee r s r s * optional receiver driver v ee r s * optional receiver figure 13. series termination or twisted pair r t r t differential ecl outputs can be terminated as independent complimentary single?ended lines. both sides of any differential pair must be terminated as identically as possible to minimize phase error and pulse width duty cycle skew. series termination is a special case of series damping in which the sum of the termination resistor (r s ) and the output impedance of the driver gate (r o ) is equal to the line characteristic impedance (figure 14). r s r o z 0 (eq. 28) where: r s = series termination resistor r o = output impedance z 0 = line characteristic impedance *t?line z 0 v ee r s receiver figure 14. series termination driver v o ab r o r t series termination techniques are useful when the interconnect lengths are long or impedance discontinuities exist on the line. additionally, the signal travels down the line at half amplitude minimizing problems associated with crosstalk. unfortunately, a drawback with this technique is the possibility of a two?step signal appearing when the driven inputs are far from the end of the transmission line. to avoid this problem, the distance between the end of the transmission line and input gates should adhere to the guidelines specified from the section on unterminated lines. series termination theory when the output of the series terminated driver gate switches levels, this driver output voltage change, v o , is impressed on the input to the transmission line (point a) as a change in voltage ( v a ) and propagates to the receiver at the output of the transmission line (point b) as a change in voltage ( v b ) in figure 14. v a v o *
z 0 r s r o z 0 (eq. 29) where: v a = input to the transmission line voltage change v b = receiver input voltage change v o = driver output voltage change z 0 = line characteristic impedance r o = output impedance of the driver gate r s = termination resistance since z 0 = r s + r o , substitution into the above equations yields: v a v o 2 (eq. 30)
and8020/d http://onsemi.com 12 from this relationship, v a = v o / 2, an incident wave of half amplitude propagates down the transmission line. at the receivers input point b, typically high impedance, the transmission line sees an unterminated open line and the signal reflection coefficient at the receiver load is approximately unity. the reflection causes the voltage to double at the receiving end. when the reflected wave arrives back at the source end, its energy is dissipated by the series resistor. when the sum of the source and series impedance is equal to the characteristic impedance of the line, no further reflections occur. calculation of r t the emitter pull?down resistor, r t , functions to establish v oh and v ol levels. voltage transitions imposed on r t propagate through r s and z 0 to a receiver. negative voltage transition are current limited by r t , r s , and z 0 when the driver output switches to the low state. the r t value must maximize the negative voltage transition and prevent the output transistor from entering the cutoff operating region in a low state (figure 15). *t?line z 0 v ee r s receiver figure 15. equivalent circuit for re determination driver r o r t the worst case scenario occurs when the driver output emitter follower enters into cutoff during a negative going transition. when this happens, the driver can be considered opened and, at the instant it opens, the line characteristic impedance behaves as a linear resistor returned to v oh . the model becomes a simple series resistive network as shown in figure 16. *t?line z 0 v ee r s figure 16. equivalent circuit with output cutoff v ee v oh r t the maximum current, i max , occurs at the instant the switch opens and is calculated by: i max (v oh v ee ) (r t r s z 0 ) (eq. 31) an initial current, i init , must be sufficient to generate a transient voltage equal to half of the logic swing since the voltage at the receiver will double due the reflection coefficient approaching 1.0 for series termination. to accommodate reflections caused by discontinuities and load capacitances the transient voltage should be increased by 25%. thus, i init is defined as: i init
1.25 * v pp 2 z 0 (eq. 32) to satisfy the initial constraints of i max > i init : ( v oh v ee ) ( r t r s z 0 )
1.25 * vswing 2 z 0 (eq. 33) solving for r t , gives the inequality: r t (kz 0 )z 0 r s (eq. 34) where: z 0 = line characteristic impedance r o = output impedance of the driver gate r s = termination resistance kz 0 = coefficient to z 0 for various series, the coefficient to z 0 , kz 0 , is presented in table e: coefficient to z 0 . table e. coefficient to z 0 series kz 0 10ep 4.0 100lvel 4.01 10el 5.99 10e 7.10 100e 6.57 for the 10ep series (lvpecl mode operation), where v oh = 2.4 v, v swing = 0.8 v, and v ee = 0.0 v: ( 2.4 0.0 ) ( r t r s z 0 ) 0.5 z 0 *z 0 r s r t 4.0 (eq. 35) for the 100lvel series (lvpecl mode operation), where: v oh = 2.345 v, v swing = 0.750 v, v ee = 0.0 v: ( 2.345 0.0 ) ( r t r s z 0 ) 0.468 z 0 *z 0 r s r t 4.01 (eq. 36) for the 10el series (pecl mode operation), where: v oh = 4.185 v, v swing = 0.958 v, v ee = 0.0 v: ( 4.185 0.0 ) ( r t r s z 0 ) 0.599 z 0 *z 0 r s r t 5.99 (eq. 37)
and8020/d http://onsemi.com 13 for the 10e series (ecl mode operation), where: v oh = ?0.9 v, v swing = 0.85 v, v ee = ?5.2 v: ( 0.9) ( 5.2) r t r s z 0 0.531 z 0 *z 0 r s r t 7.10 (eq. 38) for the 100e series (ecl mode operation), where: v oh = ?0.955 v, v swing = 0.75 v, v ee = ?4.5 v: ( 0.955) ( 4.5) r t r s z 0 0.468 z 0 *z 0 r s r t 6.57 (eq. 39) parallel fanout of series termination an extension of the series termination technique, using parallel fanout, eliminates the problem of lumped loading at the expense of extra transmission lines (figure 17). *t?line z 0 r s n receiver n figure 17. parallel fanout using series termination *t?line z 0 r s 1 receiver 1 driver r t v ee n number of lines figure 17 shows a modification of the series termination scheme in which several series terminated lines in parallel fanout are driven using a single ecl gate. the principle concern when applying this technique is to maintain the current in the output emitter follower below the maximum rated value. the value for r t can be calculated by viewing the circuit in terms of conductances. g output g1 g2 g n (eq. 40) from table b, for each of the series: 1 ( r t ) 1 ( *z 0 1 r s 1 ) kz 0 1 ( *z 0 2 r s 2 ) kz 0 1 ( *z 0 r s ) n kz 0 (eq. 41) where: n = number of parallel circuits when: z 0 1 z 0 2 z 0 , and r s 1 r s 2 r s nn (eq. 42) then r t is calculated as: r t ( kz 0 *z 0 r s ) r n (eq. 43) when a single series terminated line is driving more than a single receiver, the maximum number of loads must be addressed. the factor limiting the number of loads is the dc voltage drop across the series termination resistor caused by the summary input currents i t during the receivers quiescent high state. noise margin loss, nm loss , will probably determine the acceptable dc voltage drop limit across r s . nm loss i t *(r s r o ) (eq. 44) where: i t = sum of iinh currents r o = output impedance of the driver gate r s = termination resistance figure 18. noise margin loss example *t?line z 0 r s receiver 1 driver r t v ee r o receiver 2 receiver n i t for the majority of ecl devices typical maximum value for quiescent high state input current is 150 ua. thus, for the circuit shown in figure 18, in which three gate loads are present in a 50 environment, the loss in high state noise margin is calculated as: nm loss 3* 150 ma * 50 22.5 mv (eq. 45) this represents a potential shift in the v oh level of ?22.5 mv.
and8020/d http://onsemi.com 14 section 5. diode termination v bb v bb driver receiver *all media * * d1 d2 d1 d2 alternative to the resistor network termination schemes, a diode method shown in figure 19 may afford certain advantages when a design has the following constraints. 1. impedance controlled media line is not required (coax, twisted pair, striplines, etc.) 2. impedance matched termination network is not required. 3. overshoot and noise need to be clamped to logic high/low levels. figure 19. diode termination v bb v bb driver receiver *all media * * d1 d2 d1 d2 v bb driver receiver *all media * d1 d2 d1 and d2 diodes may be an mbd301, mmbd301, mbd701, mmbd701lt1, or a dual package mmbd452lt1. diode forward voltage curves from a data sheet, such as shown in figure 20, will determine specific current and voltage operation range. frequency limitations may be a consideration when selecting the diodes. the silicon hot?carrier schottky barrier diode mbd701, for example, displays a forward v f of about 0.55 v and an i f of about 11 ma (at 25 c) to match a 50 impedance line. at higher temperatures, the current decreases. figure 20. mmbd701 diode forward voltage 0.1 1.0 10 100 0 0.2 0.4 0.8 1.2 1.6 2.0 i f , forward current (ma) v f , forward voltage (v) t a = 85 c t a =?40 c t a = 25 c
and8020/d http://onsemi.com 15 section 6. capacitive coupling r r r r r v cc v bb although not strictly a termination, ac or capacitive coupling is often used to provide features in conjunction with proper termination. such capabilities as hot swapping capability, dc isolation to a receiver, and level shifting are possible with capacitive coupling. data stream characteristics may impose restriction on both termination and capacitive coupling. ac coupled signals have the line dc blocked and will require a dc restoration voltage, v bias , for the receiver input. data in unencoded non?return?to?zero (nrz) format will require dc restoration prior to ac coupling into a ecl receiver input. a sinusoidal waveform clock signal may be cap coupled for conversion to a square wave with 50% duty cycle and sharp rise and fall edges. the capacitor used to couple the signal must have a impedance rating of < 50 over the frequency range of the input signal. because large capacitors appear somewhat inductive at high frequencies, it may be necessary to use a small capacitor in parallel with a larger one to achieve satisfactory operation. a coupling capacitor and the signal load impedance form an rc network which will boundary the duration of a pulse. values for the r (load and leakage total resistance) and c (coupling capacitor) should be selected to provide a time constant, t c , of at least 10x the pulse width. data streams may require larger t c values to retain logic levels. hot swapping the desire often arises to remove or install a receiver or daughter card without powering down the driver or motherboard. this is termed ahot swappingo. powered driver and an unpowered receiver damage risk hot swapping presents a potential risk to an unpowered or powered down ecl device receiver and driver in either the negative or positive mode when driven by a typical signal level. when a receiver pecl receiver v cc is off or powered down, the v cc power supply typically appears as a low impedance source at 0.0 v capable of sinking considerable current. typical driver signal levels present voltages that forward bias the input esd protection diode structure and the input base collector junction. potentially lethal current paths may develop through forwarded junctions to v cc . there is also a risk for a powered down or off necl or lvnecl receiver and driver. a v ee supply will typically appear as a low impedance path to 0.0 v (gnd). typical negative levels present signal voltages that will forward bias the input esd protection diode structure and the input base collector junction to this low impedance path. potentially lethal current paths may develop through the forwarded junctions and v ee to 0.0 v. powered down receiver risk may be managed in several ways. 1. physical sequencing ? the supplies for v ee (ground) and v cc (power) may be physically connected prior to signal lines by altering the daughter board edge connection geometry, making v ee and v cc connectors protrude and engage or sequence first. v ee connectors could even be sequenced prior to v cc . this insures the supplies are powered prior to input signal voltages. 2. switching ? a relay (or analog switch) could be used to open or close the supply lines insuring the power supply line is opened when powered off. 3. cap coupling ? dc isolation of potentially damaging current. 4. series r ? an additional series impedance matching resistor, r s , will act as power splitter with an existing parallel termination resistor, r t , to accomplish some current limiting to help manage the risk. this will also attenuate the amplitude 50%, easily tolerated by most high gain, high input sensitivity devices. using v bb pin for v bias some devices provide a convenient v bb pin for use as a vbias reference supply to rebias a dc level. a dc rebias level must be at the common mode voltage of the input signal to properly preserve a 50% output duty cycle (see and8066). a package v bb pin may provide an internally generated dc switching reference voltage for the device inputs, and is available only to the package input pins. do not port one package v bb pin directly to another device without current amplification. when used, decouple v bb to v cc (or v tt ) via a 0.01 to 0.001 f capacitor to suppress noise injection. limit current to less than 0.5 ma (absolute maximum rating source or sink) as shown in figure 21. when not used, v bb should be left open.
and8020/d http://onsemi.com 16 figure 21. differential and single?ended ac configurations using v bb reference v cc or v tt v cc rpu 25 k to 100 k receiver out out in inb r t v bb v tt 1 k * 0.001 f 0.001 f v cc or v tt v cc rpu 25 k to 100 k receiver out out in v bb v tt 1 k * 0.001 f 0.01 ? 0.001 f r t r t r t * high voltage cap may be needed 0.01 ? 0.001 f a. differential b. single?ended in figure 21a, the in line has a 1 k resistor to v bb , presenting a 1 k impedance across the differential signal lines. this assumes the signal impedance matching has been accomplished prior to the cap coupling, on the driver side of cap. locate the coupling capacitor as physically close to the input pin as possible to minimize the trace length and diminish potential reflections due to the impedance mismatch. if signal impedance matching has not been accomplished prior to the cap coupling, then a characteristic impedance resistor, 2z 0 , would be used across the input lines, on the receiver side of the cap. the value of the rpu resistor would be adjusted to produce an acceptable null signal default voltage drop. auto?oscillation suppression with v bb if the differential inputs to the ac coupled device are left open or if the driving signals are lost, both receiver input pin voltages converge toward the vbias reference voltage v bb value. sustained oscillation may autonomously result from a combination of ambient environmental noise, the device small signal gain, and feedback from the output to the input through parasitic capacitive and inductive paths. as a differential receiver input voltage diverges, the output responds by transitioning toward a state voltage. a sufficient voltage across the receiver inputs will force the output to state level. depending on conditions, about 10 to 50 mv is sufficient to suppress instability oscillation and force a determined state on the output. for the configuration using the v bb reference, figure 21, this input voltage may be accomplished by injecting a minimum current from v cc through an external pullup resistor, rpu, on one input line. the value of rpu could range from 25 k to 100 k . as rpu increases, the phase error is diminished and the susceptibility to oscillation increases. generally, an internal pull?down resistor ranging in value from 52 k to 75 k is deployed on an input pin. on some d?bar (invert) input pins an additional 36 k to 75 k resistor is deployed to suppress oscillation by forcing a determined state on the output under open input or null voltage conditions. a minimum input voltage of 20 to 30 mv may be effective depending on noise, gain, and layout. generating v bb for vbias when v bb voltages are desired, but not available within a device, the reference level may be ported from a generator as illustrated in figure 22. any of the a16o type buffers are recommended for use in a high current gain v bb generator buffer. for example, the e416, el16, lvel16, ep16, lvep16, el17, lvel17, etc. type devices have a v bb pin available for constructing a v bb generator buffer. figure 22. v bb voltage reference generator v bb(out) 16 rt v tt v cc or v tt v bb 0.01 f 1 k non?v bb biasing alternative to a device supplied v bb , any voltage source may be supplied to bias receiver inputs to provide an acceptable v ihcmr (voltage input high common mode range) dc reference to the receiver (see specific device data sheet). signal impedance matching may be accomplished prior to cap coupling, allowing a wide range values for a rebiasing resistor network.
and8020/d http://onsemi.com 17 when the coupling capacitor is physically located near enough to the receiver input pins to prohibit reflections on the connecting trace length or signal impedance matching has been accomplished prior to cap coupling, then a simple high value resistor divider network from v cc to v ee is recommended as shown in figure 23. differential and single?ended ac configurations using non?v bb biasing (a and b). this network total resistance may be from 1 k to 10 k . for 50 impedance traces, the typical value for the voltage divider resistors are given in table f. typical rebias and impedance matching resistor network values @ z 0 = 50. note the impedance presented to a signal is 5 k table f. typical rebias and impedance matching resistor network values @ z 0 = 50 resistor |v cc ?v ee | = 5.0 v |v cc ?v ee | = 3.3 v |v cc ?v ee | = 2.5 v units r1 (r1 ) 4 4 4 k r2 (r2 ) 6 6 6 k v rebias 3.3 2.2 1.7 v when the coupling capacitor is physically located at a distance from receiver over a trace or cable length capable of sustaining reflections, a thevenin parallel network matching the line of impedance is recommended for their suppression. this is shown in figure 23. differential and single?ended ac configurations using non?v bb biasing (a and b). the rebias voltage may always be safely set at v cc ?1.3. for 50 impedance traces, the typical value for the voltage divider resistors are given in table g. typical rebias and impedance matching resistor network values @ z 0 = 50. table g. typical rebias and impedance matching resistor network values @ z 0 = 50 resistor |v cc ?v ee | = 5.0 v |v cc ?v ee | = 3.3 v |v cc ?v ee | = 2.5 v units r1 (r1 ) 68 83 96.15 r2 (r2 ) 192 127 104.16 v rebias 3.7 2.0 1.2 v figure 23. differential and single?ended ac configurations using non?v bb biasing v cc receiver out outb in inb r t r t v tt 0.001 f 0.001 f v ee r2 r2 r1 r1 v cc receiver out outb in v tt 0.001 f v ee r2 r2 r1 r1 r t r t a. differential b. single?ended the characterized v bb reference voltage bias, vbias, is v cc ? 1.33 v, but a device is not restricted to this vbias value. the vbias range is determined by the vpp amplitude and the signal high level, v ih . input high level, v ih , is constrained by the data sheet specification of common mode range, v ihcmr or v cmr . thus, the vbias range is constrained: vbias max v ihcmrmax (0.5)(vpp) vbias min v ihcmrmin (0.5)(vpp) a single?ended source into a differential type input signal amplitude swing, v pp , is typically constrained from v ppmin = 300 mv to v ppmax = 1000 mv. an input signal must swing symmetrically above and below vbias to preserve a 50% duty cycle out of the receiver. differential signals must have identical crosspoint voltages to preserve minimum phase error and duty cycle error. crosspoint voltages are determined by the matched precision of the resistor divider network from v cc to v ee . auto?oscillation suppression without v bb for a configuration without a v bb reference pin, such as illustrated in figure 23, the resistor network may be modified to have an input voltage of 20 to 30 mv offset between the input pins. either a high resistor value divider or a thevenin parallel network may be modified to accomplish this input voltage . this is accomplished by altering the values of r1, r1 , r2, and r2 .
and8020/d http://onsemi.com 18 combining the high impedance and impedance matching networks results in an input voltage scheme shown in figure 24. this creates the proper input voltage , vbias, using fewer components. v tt v cc figure 24. vbias and auto?oscillation suppression with thevenin parallel network r1 receiver out outb in r t r t v ee 0.001 f r z r2 inb 0.001 f for a 3.3 v v cc , the values of r1 and r2 provide a thevenin parallel network divider voltage with v ih in the v ihcmr of the receiver. current through the divider develops the default offset across r z and can be adjusted as needed. for example, in z 0 = 50 traces, a 30 mv default offset difference will be created if v cc = 3.3 v and the dc bias voltage is 2.0 v (typical v bb ) when: r1 4.22 k r2 6.34 k r z 100 the 0.001 coupling cap may need to be adjusted to frequency and vpp amplitude of the receiver input signal. a similar single?ended network may be used with only one coupling cap and sufficient bypass capacitance on the non?driven resistor to preserve a dc level. output level shifting receiver inputs may be level shifted using capacitive coupling and adjusting vbias within the acceptable common mode range for v ih . output levels may also be changed independent of input levels. the driver device may be operated with both v cc and v ee at shifted values. this is used at the factory to evaluate devices and conveniently port signals directly into standard 50 impedance equipment modules. the v cc is fixed to +2.0 v above test system chassis ground and the test equipment internal 50 impedance constitutes a proper signal termination. thus, the split v ee supply is adjusted to a negative value. |v cc ? v ee | split v cc split v ee unit 3.0 +2.0 ?1.0 v 3.3 +2.0 ?1.3 v 5.0 +2.0 ?3.0 v 5.5 +2.0 ?3.5 v output levels may be shifted to symmetrically cross 0.0 v by a similar method although the advantage of conveniently directly connecting into standard test equipment is no longer available. on semiconductor and are registered trademarks of semiconductor components industries, llc (scillc). scillc reserves the right to mak e changes without further notice to any products herein. scillc makes no warranty, representation or guarantee regarding the suitability of its products for an y 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 wi thout limitation special, consequential or incidental damages. atypicalo parameters which may be provided in scillc data sheets and/or specifications can and do vary in different application s and actual performance may vary over time. all operating parameters, including atypicalso 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 sit uation 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, af filiates, and distributors harmless against all claims, costs, damages, and expenses, and reasonable attorney fees arising out of, direct ly 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. this literature is subject to all applicable copyright laws and is not for resale in any manner. publication ordering information n. american technical support : 800?282?9855 toll free usa/canada japan : on semiconductor, japan customer focus center 2?9?1 kamimeguro, meguro?ku, tokyo, japan 153?0051 phone : 81?3?5773?3850 and8020/d literature fulfillment : literature distribution center for on semiconductor p.o. box 61312, phoenix, arizona 85082?1312 usa phone : 480?829?7710 or 800?344?3860 toll free usa/canada fax : 480?829?7709 or 800?344?3867 toll free usa/canada email : orderlit@onsemi.com on semiconductor website : http://onsemi.com order literature : http://www.onsemi.com/litorder for additional information, please contact your local sales representative.

▲Up To Search▲

Price & Availability of MC10H106

	To Download MC10H106 Datasheet File
If you can't view the Datasheet, Please click here to try to view without PDF Reader .