On-chip termination circuit

[0002] This application claims priority from U.S. Provisional Application No. 60/194,674, filed Apr. 4, 2000, which is incorporated herein by reference.

[0003] 1. Field of the Invention

[0004] This invention relates to transmission of digital image data sent from a computing device to a display device and, more particularly, to an on-chip termination circuit for terminating a line transmitting the digital image data from the computing device to the display device for eventual visual display.

[0005] 2. Description of the Prior Art

[0006] A personal computing device transmits digital image data to an attached display device through a transmission line for eventual visual display. An ideal transmission line will deliver the image data signal to the display without any loss or corruption. Practical transmission lines, however, are not ideal.

[0007] To minimize distortion of the transmitted digital data, the transmission line requires proper transmission line termination. If the transmission line is not terminated with its characteristic impedance, the input impedance may become complex and exhibit inductive or capacitive characteristics depending on the nature of the load, its electrical distance from the source, or the measure of termination resistance.

[0008] If a mismatch occurs between the line's characteristic impedance and the termination resistance as the image data signal travels down the transmission line, a portion of the image data signal is reflected toward the computing device. The reflected wave combines with the forward wave and alters it according to its relative phase and amplitude, corrupting the image data. To avoid impedance mismatches, transmission lines must be terminated commonly with precise termination resistances having tight tolerances typically around 5% or better.

[0009] Transmission line termination is typically achieved through fixed, discrete, off-chip, precision resistors that accurately match the transmission line characteristic impedance. Off-chip precision resistors having tight tolerances are readily available. But they require significant circuit board space increasing circuit board layout complexity and cost.

[0010] Transmission line termination is particularly challenging when working with resistances built into semiconductor integrated circuits (on-chip). On-chip termination resistances are difficult to manufacture with tight tolerances. On-chip resistors typically show a variation as much as 15-20% or worse over process and temperature. Precision on-chip resistors are difficult to achieve with high volume semiconductor processing methods, for example, because of variations in sheet resistance and line width control. Furthermore, many applications use high-density digital Complementary Metal Oxide Semiconductor (CMOS) technology for high speed. As speed performance improves, process elements such as poly and moat resistors are often unavailable to circuit designers because of the use of silicide to minimize sheet resistances. Thus, to create resistors, additional mask levels, or the use of analog specific processes, are required.

[0011] An alternative technique for implementing termination resistors on a CMOS integrated circuit is to bias a Metal Oxide Semiconductor Field Effect Transistor (MOSFET) into the triode region. FIG. 1 is a typical output characteristic of a MOSFET device in which two regions of operation are shown: the triode region and the saturation region. A MOSFET device biased in the triode region operates similarly to an ideal resistor as shown in FIG. 1 where the drain current is approximately linearly proportional to the drain-source voltage. As the drain source voltage increases, however, the MOSFET device's resistance exhibits non-linearity with respect to drain-source voltage. This non-linearity results in operation that deviates from the ideal resistor. More specifically, the differential or small signal MOSFET resistance given by dV_DS/dI_D, changes as the voltage across the drain and source of the MOSFET changes. This change in differential resistance is large enough to be problematic in many applications, including image data transmission, where it can cause impedance mismatch resulting in bit errors.

[0012] Accordingly, a need remains for an inexpensive and effective on-chip termination circuit that prevents impedance mismatches in transmission line circuits by exhibiting superior linearity and tight manufacturing tolerances.

[0013] The foregoing and other objects, features, and advantages of the invention will become more readily apparent from the following detailed description of a preferred embodiment that proceeds with reference to the following drawings.

[0014] FIG. 1 is graph of the output characteristics of a MOSFET device.

[0015] FIG. 2 is a block diagram of a termination circuit of the present invention.

[0016] FIG. 3 is a schematic diagram of the termination circuit of the present invention.

[0017] FIG. 4 is a graph of output characteristic curves comparing the termination resistance of the termination circuit of the present invention with a termination resistance comprising a single MOSFET transistor operating in the triode region and with a termination resistance comprising an ideal resistor.

[0018] FIG. 5 is graph of the differential resistance of the termination circuit of the present invention, the differential resistance of a MOSFET transistor operating in the triode region, and the differential resistance of an ideal resistor.

[0019] FIG. 6 is a graph of termination resistance variations over worst-case process and temperature extremes for the termination circuit of the present invention.

[0020] The following description of the preferred embodiment of the present invention refers to specific technical requirements of the Digital Display Working Group's (DDWG) Digital Visual Interface (DVI) standard, revision 1.0, published Apr. 2, 1999. The DVI is incorporated herein by reference. The termination circuit of the present invention, however, is not limited to the DVI application. Rather, the termination circuit of the present invention applies generally to transmission lines and twisted-pair transmission lines in particular.

[0021] Those of skill in the art will appreciate that the invented apparatus and method described and illustrated herein may be implemented in hardware, software, firmware, or any suitable combination thereof. Preferably, the apparatus and method are implemented in hardware integrated into a semiconductor device and, more particularly, are implemented integrated into a semiconductor digital display controller for purposes of driving a visual display inexpensively and flexibly. Those of skill in the art will appreciate that the apparatus and method of the present invention can be implemented in software or firmware by a computer or microprocessor process in which instructions are executed. The instructions are stored for execution on a computer-readable medium and executed by any suitable instruction processor. Alternative embodiments are contemplated, however, and are within the spirit and scope of the invention.

[0022] The present invention is designed to address resistance variability by making a portion of the termination resistance a MOSFET transistor operating in the triode region and replicating the termination resistance inside of a feedback loop to control the resistance of the termination resistance through the replication resistance. The MOSFET transistor's non-linear operation is mitigated by the addition of a linear resistor in series with the MOSFET transistor. By doing so, a substantial portion of the voltage across the composite termination resistance is across the linear termination resistor thereby significantly reducing the non-linear effects of the MOSFET transistor.

[0023] Referring to FIG. 2, an image data signal 12 is transmitted through a twisted pair differential transmission line (not shown) from a computing device (not shown) to a termination circuit 10 included in a semiconductor display device controller (not shown) that is, in turn, included in a display device (not shown). A termination resistance 14 receives the image data signal 12 and is coupled to a replication resistance 16. The replication resistance 16 replicates or duplicates the termination resistance 14. A reference voltage generating circuit 20 generates a reference voltage V_REFand a reference current I_REF. The reference voltage generating circuit 20 is coupled to the replication resistance 16 and the feedback circuit 18. The reference voltage generating circuit 20 includes a reference circuit 24 coupled to a bandgap reference voltage generator 22 and a reference resistor 26. The reference resistor 26 is preferably not integrated into the display device controller (not shown). That is, the reference resistor 26 is preferably off-chip so that an accurate enough resistance is easily achievable. The bandgap reference voltage generator 22 is preferably integrated into the display device controller (not shown) primarily because the termination circuit 10 is insensitive to the accuracy of the bandgap reference voltage V_BGas is explained in more detail below. The reference current Iref flows through the replication resistance 16. A feedback circuit 18 receives the reference voltage V_REFand maintains the reference voltage V_REFacross the replication resistance 16. Since the reference current I_REFand the reference voltage V_REFare maintained substantially constant due to the operation of the reference resistor 26 and the feedback circuit 18, nonlinear operation associated with the termination resistance 14 substantially disappears as described in more detail below.

[0024] FIG. 3 is a detailed schematic of the termination circuit 10 of the present invention. Referring to FIG. 3, the termination circuit 10 includes a termination resistance 14 for terminating each of the twisted pair transmission lines RX+ and RX− as required by the DVI standard. The transmission line RX+ is terminated with termination resistance 14A comprising MOSFET transistor MR2 serially connected to a termination resistance TR2. Likewise, the transmission line RX− is terminated with termination resistance 14B comprising MOSFET transistor MR3 serially connected to a termination resistance TR3. The MOSFET transistors MR2 and MR3 have corresponding gates, sources, and drains. The sources of transistors MR2 and MR3 are commonly connected to a power supply V_DD. The drain of MOSFET MR2 is connected to a first end of the resistor TR2 and the drain of MOSFET MR3 is connected to a first end of the resistor TR3. A second end of the resistor TR2 receives the image data signal 12 transmitted on the transmission line RX+. A second end of the resistor TR3 receives the image data signal 12 transmitted on the transmission line RX−.

[0025] The gates of transistors MR2 and MR3 are commonly connected to the replication resistance 16. The replication resistance 16, in turn, comprises a MOSFET transistor MR1 serially connected to a replication resistor RR1. The MOSFET transistor MR1 includes a gate commonly connected to the gates of transistors MR2 and MR3 and to the feedback circuit 18, a source connected to the power supply V_DD, and a drain connected to a first end of the replication resistor RR1.

[0026] The second end of the replication resistor RR1 is coupled to the reference voltage generating circuit 20. The reference voltage generating circuit 20 includes a bandgap voltage reference generator 22 adapted to generate a bandgap voltage V_BGand supply the bandgap voltage V_BGto a positive input of amplifiers A1 and A2. The amplifier A1 includes a negative input connected to a first end of the reference resistor R_REF(block 26 in FIG. 2). The second end of the reference resistor R_REFis connected to ground. The amplifier A1 includes an output connected to a gate of a MOSFET transistor M2. A source of the MOSFET transistor M2 is connected to the first end of the resistor R_REF.

[0027] The bandgap voltage reference generator 22 is preferably integrated into the semiconductor display device controller (not shown). The reference resistor R_REFis external to the display device controller (not shown) or off-chip. The reference resistor R_REFis preferably an inexpensive precision (e.g., 1 percent or better) metal film resistor. Both the bandgap reference voltage generator 22 and the reference resistor R_REFare common in mixed signal integrated circuits and thus, do not add significant cost to the termination circuit 10.

[0028] The amplifier A2 includes a negative input connected to a first end of a resistor R1. The amplifier A2 includes an output connected to a gate of a MOSFET transistor M1. A source of the MOSFET transistor M1 is connected to the first end of a resistor R1. A second end of the resistor R1 is connected to ground. A drain of the MOSFET transistor M1 is connected to a first end of a resistor R2. A second end of the resistor R2 is connected to the power supply V_DD.

[0029] The feedback circuit 18 comprises an Operation Transconductance Amplifier (OTA) and a capacitor C1. The OTA has a negative input connected to the first end of the resistor R2 and the drain of the MOSFET transistor M1. An output of the OTA is connected to a first end of the capacitor C1 and to the gates of MOSFET transistors MR1, MR2, and MR3. The second end of the capacitor C1 is connected to ground. The positive input of the OTA is connected to the second end of the replication resistor RR1 to thereby form a feedback loop with the replica resistance 16 and the capacitor C1. The capacitor C1 is a compensation capacitor that ensures stability of the OTA's feedback loop.

[0030] The termination circuit 10 operates as follows. The bandgap reference voltage generator 22 provides the bandgap voltage V_BGto the positive input of the amplifiers A1 and A2. The feedback loop of amplifier A1 forces the bandgap voltage V_BGacross the first end of the reference resistor R_REFgenerating the reference current I_REF. The reference current I_REFflows through the replica resistance 16 because substantially no current flows through the negative input of amplifier A1 or the gate of transistor M2 due to their correspondingly high input resistance. The reference current I_REFis given by V_BG/R_REF.

[0031] The feedback loop of amplifier A2 forces the bandgap voltage V_BGacross the first end of resistor R1. The current I_R1flows through the resistor R1 and through the resistor R2 because substantially no current flows through the negative input of amplifier A2, the gate of transistor M1, or the negative input of the OTA due to their correspondingly high input resistance. The current I_R1is given by V_BG/R1. The reference voltage V_REF1across the resistor R2 is given by V_REF1=R*I_R1. Since I_R1=V_BG/R1, then V_REF1=R2/R1*V_BG.

[0032] Note that in the above-described equation, the reference voltage is proportional to a ratio of resistors R2 and R1. Both of these resistors are constructed of identical material, e.g., non-salicide polysilicon, on a common monolithic integrated circuit (display device controller (not shown)). Therefore, the ratio of R2 to R1 is precisely known despite variations in absolute resistance due to temperature and process variations. A person skilled in the art should recognize that the resistors R2 and R1 might be made of a variety of materials including non-salicide polysilicon.

[0033] The replication resistance 16 is implicitly known by so establishing a reference current I_REFthrough and a reference voltage V_REF1across the replication resistance 16. Amplifier A2, resistors R1 and R2, and transistor M1 generate the reference voltage V_REF1responsive to the bandgap voltage V_BG. The action of the feedback circuit 18 is to ensure that the reference voltage V_REFis maintained substantially constant across the replication resistance 16. Amplifier A1, reference resistor R_REF, transistor M2 generate the reference current I_REFresponsive to the bandgap voltage V_BG.

[0034] The OTA compares the voltage across the replication resistance 16 to the reference voltage V_REF1and acts to drive the voltage across the replication resistance 16 to the reference voltage V_REF1. For example, if the voltage at the second end of the replication resistor RR1 is lower than the reference voltage V_REF1, the OTA sinks current from the capacitor C1 causing the voltage at the gate of transistor MR1 to decrease. This gate voltage decrease increases the enhancement of the transistor MR1, in turn, increasing the voltage at the second end of the replication resistor RR1.

[0035] The value of the replication resistance 16 is found from application of Ohm's Law: $\begin{array}{c}\mathrm{Replication}\mathrm{Resistance}=V_{\mathrm{ACROSS}\mathrm{REPLICATION}\mathrm{RESISTANCE}}/\\ I_{\mathrm{THROUGH}\mathrm{REPLICATION}\mathrm{RESISTANCE}}\\ =\left(\left(\left(\mathrm{R2}/\mathrm{R1}\right)*\mathrm{Vbg}\right)/\left(\mathrm{Vbg}/R_{\mathrm{REF}}\right)\right)\\ =\left(\mathrm{R2}/\mathrm{R1}\right)*R_{\mathrm{REF}}\end{array}$

[0036] Thus, the replication resistance 16, and the termination resistance 14 by association, is dependent on the ratio of resistors R2 and R1 and on the external reference resistor R_REF. As mentioned above, the ratio of R2 and R1 is precisely controlled on the display device controller (not shown). The reference resistor Rref is preferably chosen as an inexpensive precision (e.g., 1 percent or better) metal film resistor. Importantly, the replication resistance 16 is not dependent on the bandgap voltage V_BG. Thus, any convenient reference voltage can be used regardless of whether its magnitude is precisely controlled.

[0037] The termination resistance 14 can be similarly precisely controlled provided that the same gate bias is applied to the termination transistors MR2 and MR3 as is applied to the replication transistor MR1 and provided that the termination transistors MR2 and MR3 are built using the same processes as the replication transistor MR1 on the same monolithic integrated circuit such as the display device controller mentioned above. The termination transistors MR2 and MR3 receive a gate bias from the feedback circuit 18, including the replication transistor MR1. In the preferred embodiment of the present invention, the replication transistors MR2 and MR3 are built by connecting N number of devices identical to the replication transistor MR1 in parallel. This results in a drain/source resistance of the termination transistors MR2 and MR3 that is 1/N times that of the replication transistor MR1.

[0038] Similarly, resistors TR2 and TR3 are made by paralleling N number of resistors identical to the replication resistor RR1 resulting in a resistance equal to RR1/N. That is, the value of resistor TR2 is RR1/N and the value of resistor TR3 is RR1/N. Thus, the termination resistance 14 is 1/N times that of the replication resistance 16. By so scaling the termination resistance 14, the replication resistance 16 can operate at a higher resistance and a consequent lower current, resulting in an overall power savings.

[0039] For example, the DVI standard requires that the termination resistances 14A and 14B be connected to corresponding transmission lines RX+ and RX− on a first end and to a power supply V_DDon a second end, the power supply V_DDbeing nominally equal to +3.3 V. The DVI further requires that the termination resistances 14A and 14B are equal to 50 Ω. In the preferred embodiment, the termination resistors TR2 and TR3 are non-salicide polysilicon resistors having excellent linearity at the expense of a large tolerance. The value of each of resistors TR2 and TR3 is selected such that each can never exceed the desired 50 Ω resistance. Let resistors TR2 and TR3 have a nominal value of 37.5 Ω. Assuming a 15% tolerance, the range of values for resistors TR2 and TR3 is 31.9 to 43.1 Ω. Thus, transistors MR2 and MR3 must take on a possible range of resistance values from about 6.9 to 18.1 Ω. The size of transistors MR2 and MR3 is selected such that this range of drain/source resistance is achieved when the transistors MR2 and MR3 operate in the triode region. Assuming N is 10, the replication resistance 16 is 10 times that of the termination resistances 14A or 14B. Therefore, the replication resistance equals 500 Ω. The value of the replication resistor RR1 can be selected in similar fashion to that described above in relation to termination resistors TR2 and TR3.

[0040] FIG. 4 is a graph of output characteristic curves comparing the termination resistance of the present invention with a termination resistance comprising a single MOSFET transistor operating in the triode region and with a termination resistance comprising an ideal resistor. The termination resistance comprising a MOSFET transistor is shown as the topmost curve. As shown, the MOSFET exhibits a nonlinear performance over the range of currents indicated with a significant nonlinearity at currents greater than 10 mA. The MOSFET transistor matches the present invention and the ideal resistor in large signal resistance at slightly below 10 mA. In contrast, the present invention substantially tracks the performance of the ideal resistor over the broad range of currents indicated. The ideal resistor is represented by the lower most curve on the graph. The performance of the termination circuit of the present invention is represented by the middle of the three curves shown in FIG. 4.

[0041] FIG. 5 highlights the nonlinear operation of the termination resistance comprising a single MOSFET transistor by showing the curved differential resistance of the MOSFET transistor (line X). In contrast, the differential resistance of the present invention is fairly fiat (line O). The differential resistance of an ideal 50 Ω resistor is 50 Ω. The present invention maintains a differential resistance very close to 50 Ω over the 0-12 mA range. The MOSFET transistor operating in the triode region varies from less than 40 Ω to over 100 Ω over the same range, a clearly unsuitable termination resistance where 5-10 percent tolerances are required.

[0042] FIG. 6 is a graph of variations in termination resistance of the present invention over worst-case process and temperature extremes. Each of the curves shown in FIG. 6 shows the differential termination resistance of the present invention versus current at a different process corner. The temperatures shown are 0, 62.5, and 125 degrees and the five process comers shown are TT (typical N and P processes), FF (fast N and P processes), SS (slow N and P processes), FS (fast N and slow P processes), and SF (slow N and fast P processes). As is shown in FIG. 6, the differential termination resistance of the present invention remains within acceptable tolerances at all extremes.

[0043] Having illustrated and described the principles of my invention in a preferred embodiment thereof, it should be readily apparent to those skilled in the art that the invention can be modified in arrangement and detail without departing from such principles. I claim all modifications coming within the spirit and scope of the accompanying claims.