■WITH PIPELINE DELAY CIRCUIT

STATEMENT OF GOVERNMENT INTERES

[0001] The invention was made with United States Government support under contract .DAAD17-Q2-C-01 15 awarded by the Army, and the United States

Government has certain rights in this Invention.

FIELD OF THE DISCLOSURE

j0002j This disclosure relate to communications, and more particularly, to ROM- based direct digital synthesizer architectures.

BACKGROUND

10003] Direct digital synthesizers (DDS) include a phase accumulator, a phase-to- amplitude converter, and a digitai-to-anaio converter (DAG). In liigh-speed applications, these functional blocks are generally clocked with pipeline registers internal to each block. The phase-to-amplitude converte can be implemented by various techniques,, including using a read onl memory (ROM), logic gates, or by a combination of logic gates and sine-weighted DAC. ROMs, however, are typically not used in high-speed DDS circuits because propagation delays in the ROM circuit make it either impossible to pipeline at high-speed, or are too costl to pipeline in terms of power consumption. As such, instead of a ROM, high-speed DDS circuits typicall implement the phase-to-amplitude converter as a combination of logic gates and a sine- weighted DAC". Such conventional approaches suffer from limited spurious-free dynamic range (SFDR), typically in the range of 30 dBc to 40 dBe (i.e., with respect to carrier .frequency amplitude) at best.

SUMMARY

[0004] One. embodiment of the present invention provides a direct digital synthesis (DPS), system. The system includes a read only memory (ROM) for storing phase-to- amplitude conversion data and generating digital amplitude values corresponding to respective digital phase values. The system further includes delay circuitry operativel coupled to the ROM and for adjusting timing of data output by the ROM to compensate for propagation dela of the DPS system, the delay circuitry including a plurality of delay elements that can he selected alone or in combination to adjust the timing of the data output by the ROM, The system may include a phase accumulator for computing the phase value based on a frequency control word _* The system may include a■digital- to-analog converte (DAC) tor converting the digital amplitude values int corresponding analog amplitude values. In one such configuration, the delay circuitry may also be operatively coupled to the DAC and for adjusting timing of data output by the DAC to compensate for propagation dela of the DDS system. In another such configuration, the system may include additional delay cireiiitry operatively coupled to the DAC and for adjusting timing of data output by the DAC to compensate for propagation dela of the DDS system, the additional delay circuitry including a plurality of delay elements that can be selected alone of in combination to adjust the timing of the data output by the DAC. In one particular case, the delay circuitry may be responsive to a control signal that can change _: automatically during operation of the DDS system in response: to a change in a DDS operating parameter. In one such configuration, the DDS operating parameter may include, for example, a frequency control word input to DDS system, a clock signal for clocking one or more pipeline registers of the ROM, and/or an environmental variable, in anoihe such configuration, the system may further include a processor for generating the control signal. The system .may be included, for example, in a system-on-chi configuration, and may also be implemented with discrete components.

[0005] A number of variations will be apparent in light of this disclosure.: For instance, another example embodiment of the present invention provides a DDS system that includes a phase -accumulator for computing the phase values based on a frequenc control word, The system further includes a ROM fo storing phase-to-ampiitude conversion data and generating digital amplitude values corresponding to respective digital phase values provided by the phase accumulator. The system further include delay circuitry operatively coupled to the ROM and for adjusting timing of data output by the ROM to compensate for propagation delay of the DDS system, the delay circuitry including a pluralit of dela elements that can he selected: alone o in combination to adjust the timing of the data output by the ROM, and the delay circuitr adjusts th timing of the data output by the RO by adjusting delay of a clock; signal that clocks one or more pipeline registers of the ROM. The system further includes a DAC for converting the digital amplitude values into corresponding analog ampiitude values. In one particular such case, the dela circuitry may also be operatively coupled to the DAC and for adjusting timing of data output by the DAC to compensate for propagation dela of the DDS system, in another particular case, the system may. further include additional delay circuitry operatively coupled to the DAC and for adjusting; timing of data output by the DAC to compensate for propagation delay of the DDS system, the additional delay circuitry including a plurality of delay elements that can be selected alone or in combination to adjust the timing of the data output by the DAC. in another particular case, the delay circuitry may be responsive to a control signal tha can change automatically during operation of the DDS system in response to a change in a DDS operating parameter. In one such case,, th DDS operating parameter may include at least one of the frequency control word, the clock signal, and/or an environmental variable. In another such particula case, the system ma include a processor for generating the control signal. The system may be included, for example, in, a system-on-chip or chip set configuration.

| f006) Another example embodiment of the present invention provides a DDS system that includes: a phase accumulator for computing the phase values based: on a frequency control word, and a ROM for storing phase4O-amplitu.de conversion data and generating digital amplitude values corresponding to respective digital phase values provided by the phase accumulator _* The system further includes dela circuitry operatively coupled to. the ROM and for adjusting timing, of data output b the ROM to compensate fo propagation delay of the DDS system,, the delay circuitry including a plurality of delay elements that can be selected alone or in combination to adjust the timing of the data output by the ROM, and the delay circuitr adjusts the timing of the data output by the ROM b adjusting, delay of a clock signal that clocks: one or more pipeline registers of the ROM. The system further includes a DAG for converting the digital amplitude values into corresponding analog amplitude values. The system further includes a processor for generating a control signal to which the delay circuitryis responsive, and changing the control signal in response to a change in a DDS operating parameter, the DDS operating parameter including at least one of the frequency control word, the clock signal, and/or an environmental variable. In one such case, the dela -circuitry is also operative])' coupled to the DAC and for adjusting timing of data output by the DAC to compensate for propagation delay of- the DDS system, in anothe such case, the system includes additional delay circuitry operatively coupled to the DAC and for adjusting timing of data output by the DAC to compensate for propagation delay of the DDS system,, the additional delay circuitry including a plurality of delay elements that can b selected alone or in combination, to adjust the timing of the data output by the DAC. and the additional delay circuitry adjusts the timing of the data output by the DAC by adjusting dela of a clock signal that clocks one or more pipeline registers of the DAC, The system may be included, for example, in a system-on-chip configuration (or in a chi set or with discrete components).

[0007] The features and advantages described herein are not all-inclusive and, in particular, many additional features and advantages will be apparent to one of ordinary skill in the art. hi view of the drawings, specification, and claims. Moreover, it should be noted that the language used in the specification has bee principally selected for readability and instructional purposes, and not to limit the scope of the inventive subject matter,

BRIEF DESCRIPTION OF THE DRA INGS

[0008] Figure 1 is a block diagram, of a DDS system having a ROM-based phase-to- ampiimde converter, configured in accordance with one embodiment of the present invention.

[0009] Figure 2a is a block diagram showing further details of a ROM-based phase- to-ampiitude converter, configured in accordance with one embodiment of the present invention. |0010] Figure 2b is a block diagram showing further details of a ROM-based phase- io-amplitudc converter, configured in accordance with another embodiment of the present invention.

[0011] Figure 3a is a block diagram showing a circuit configured for generating a dela select control signal, in accordance with one embodiment of the present invention.

100123 Figure 3b illustrates a method for generating a delay select control signal, in accordance with one embodiment of the present invention,

|0013] Figure 4a is a block diagram of an adjustable clock delay circuit for use in a ROM-based phase-to-amplitude converter of a DDS system, configured in accordance wi th one embodiment of the present invention .

[0014] Figure 4b is a block diagram of an adjustable clock delay circuit for use in a ROM-based phase-to-arnpiitude converter of a DDS system, configured in accordance with another embodiment of the present invention.

DETAILED DES CRIPTION

[0015] A ROM-based DDS system is disclosed that is configured to provide a variable clock delay that allows timing of data coming out of the ROM to be matched to the timing of the DA.C. This variable clock delay approach further allows compensation for variations in propagation delays due to process variations and/or environmental conditions. The disclosed techniques can be impiexnented in an number of configurations, including integrated system-on-chip or chip set configurations as well as on printed circuit boards configured with various discrete DDS systenx components. The speed of the DDS system (e.g.. below 1 GHz, or in the range of 1 to 20 GHz, or in excess of 20 (il l/) iil depend on the manufacturing processes employed as well as the particular application at hand, as- will be appreciated in light of this disclosure.

General Overview

[00161 As previously explained, ROMs are typically not used to implement the phase o-anipli†u.de converter of a high-speed DDS circuit because propagation delays in the ROM circuit make it either impossible to pipeiine at high-speed, or are too costly to pipeline in terms of power consumption. The ROM, however, can allow for improved SFDR over other approaches, Overcoming the limitations associated with a ROM-based DDS to exploit the improved SFDR is: not a trivia! task.

f0017J Thus, and in accordance with an embodiment of .the -present invention,- a DDS system is provided that is configured with a phase accumulator, a RDM-based phase- to~amplitude converter, and a DAC. The output bit of the ROM generally have similar propagation delays. An adjustable clock delay circuit is used to take advantage of this approximately uniform (although larger than a clock cycle) propagation delay of the ROM bits, so the ROM can be pipelined and used in high-speed DDS circuits to enable improved SFDR,

[0Θ18] n one specific example case, the adjustable clock dela circuit is used for the pipeiine registers at the output of the ROM and in the DAC. to match the delayed pipeline timing. The variable clock delay allows the digital timing of the data comin out of the ROM to be matched to the timing of the DAC. Ideally, one might, consider determining and implementing an appropriate fixed clock delay that would accomplish this tmiing match between the ROM and DAG. However, there are a number of issues associated with using a fixed delay. [0019] For instance, and assuming modeling would be used in determining the appropriate fixed clock delay, the device models and parasitic extraction are not always sufficiently accurate.. As such, there is some degree of ambiguity in simulations of a fixed clock delay circuit. In addition, there are: variations in the devices and metallization from circuit to circuit within a single wafer, as well as variation between lots of wafers, that impact how well the simulation models match what is actually implemented as circuits which can be measured. In short., depending on how tightl the: process tor making the devices is controlled, variations in actual circuits versus models can be relatively high. Moreover, even if all the models were perfect, the performance of the devices varies, for instance, with temperature and supply voltages. For example, the timing see at 0°C will be different than the timing seen at I00°C. This is also the case, for example, for -10% of nominal voltage supplies: versus +10% of nominal voltage supplies. So while it may be possible to design and model a correct timing relationship at one operating condition, the design may not work over a wide range of operating conditions.

100201 The adjustable clock delay provided in accordance with an embodiment of the present invention allows for different timing settings to be used depending on variables such as the process variations, temperature, and supply voltages, and also allows for flexibility if there is ambiguity in the modeling. In addition, if there are any variations in the expected timing of a simulated DDS circuit compared to the actual timing of a fabricated, measured DDS circuit, different delay settings can b used to ensure that timing of ROM and DAC meet. This flexibility improves circuit yield and leads to a better chance of circuits working the first time they are fabricated. [0021] The techniques provided herein are particularly suitable for high-speed DDSs, such as those -operating In. the 20 to 40: GHz range. These DDSs can be used, for example, as fast, agile local oscillator generators and jammers for system-on-chi applications in commercial and military platforms. Note, however, that the techniques here may also be used in lower speed .DPS applications, if so desired. Numerous other applications will be apparent in light of this disclosure.

System Architecture

[0022] Figure 1 is a block diagram of a D S system having a ROM -based phase-to- ampliaide converter _* configured in accordance with one embodiment of the present invention. The system can. be implemented, for example, as an integrated system-on- chip configuration, but other embodiments can be implemented with chip sets and/or discrete components:.

[0023] As can be seen, the DPS system includes a phase accumulator, ROM-based phase o -amplitude converter, and a DAC, each of which operates in conjunction with a clock signal. The phase accumulator further receives a frequency control word as input, as will be explained in turn. In operation, the phase accumulator computes phase values representing phase angles, and the phase-to-amplitude converter outputs digital amplitude values corresponding to the phase, values (e.g., a sine of the phase angle). The DAC converts the digital amplitude values into a. corresponding- analog _, amplitude (e.g., _: voltage or current), thereby producing an analog sinusoidal output signal (e.g., sine wave, square wave, triangle wave, etc).

[0024J Any number of suitable process technologies and material systems can be used in fabricating the. DPS system. For example, for higher speed applications (e.g., in excess of 10 GHz), the DDS system can be implemented, for instance, in hetero junction bipolar transistor (HBT) and/or high electron mobility transistor (HEMT) technology. However, if such high speeds are not necessary .(e.g., less than 2 GHz), then other technologies such as conrplementar metal oxide semiconductor (CMOS) technology can he utilized. Example materials from which the system can be implemented include III- V -material systems such as gallium arsenide (GaAs) or indium phosphide (IhP), or other suitable semiconductor material systems

[O025J The phase accumulator can be mplemented ^' as. typically done, and -determines the frequency range and resolution of the DDS system. In particular, the clock speed at which the phase accumulator operates determines frequency range and the bit-width of the frequency control word determines the resolution. The phase accumulator essentially is a counter that increments its count for each pulse of the clock received. The magnitude of the increment is determined by the .frequenc control word (FCW).. which, is binary-coded. This FCW sets the phase step size between the clock updates. In general, the larger the phase ste size, the faster the phase accumulator reaches full count and completes the corresponding sinusoidal cycle. Once overflowed, the counter effectively wraps around and repeats the counting process.

[0026] In operation, for instance, an 8-bit wide phase accumulator Would provide a DDS frequency resolution of 1/256 of the clock frequency , with 128 steps of frequency control, and a 12-bit wide phase accumulator would provide a. DDS frequency resolution of 1/4096 of the clock frequency, with 2048 steps of frequency control. For an 8-bit wide phase accumulator, n FCW of 1 (1)0000.001) would result in the phase accumulator overflowing after 2 ^s clock cycles (256 increments), and an FCW of 128 (011 1 1 1 1 1 ) would result in the phase accumulator overflowing after two clock cycles, which is the minimum per Nyquist sampling theory. The bit width of the phase accumulator can vary from one embodiment to the next (e.g... from 4 to 48 bits), and the claimed invention is not intended to be limited to an particular phase accumulator configuration,

[0027J As is known, the basic tuning equation for a DDS architecture, which is: f _ou

~ [ (FCW x F _c ) ./ 2 ⁿ ]. where ¾ _u{ is the output frequency of the DDS system, _. FCW is the binary frequency control word, F _c is the frequency of the clock signal, and n is the bit-length of the phase -accumulator; Changes to the value of FCW result in immediate and phase-continuous changes in the output frequency f ₀ ut- ^s ^out ^{i 'S} increased, the number of samples pe cycle decreases. Since sampling theory dictates that at least two samples per cycle are required to reconstruct the output waveform, the maximum fundamental output frequency f _0U is one half the clock frequency F _c . For instance, and assuming the DDS has a 12-bit phase accumulator (n=12) and a clock frequency F _c of 24 GHz,, the DDS would be capable of synthesizing output frequencies up to 12 GHz in steps that are 1/4096 of the 24 GHz clock rate. ^" Note,, however, the output frequency may be limited to slightl less than that, giving consideration to real life factors: such, as- improving the quality of the reconstructed waveform and output filtering. When generating a constant frequency , the output of the phase accumulator increases linearly, so the waveform generated hy the phase accumulator is effectively- ramp (in the digital domain).

J0028] The digital phase value output of the phase accumulator is applied to the ROM-based phase-to-amplitude .converter, which includes one or more ROMs for Storing phase-to-aniplittide conversion data and generating the corresponding; digital amplitude values for each phase value received from the phase accumulator.. As previously explained, ROM technology can limit the speed of a DDS, To this end, the overall speed of the DBS can be increased by reducing the bit access time of the ROM. For instance, and in one example embodiment, the ROM included in the ROM-based phase-to -amplitude converter can be implemented with a 16 by 6-bit, 36 GHz clock frequency ROM fabricated in any ^' suitable process (e.g., InP DHBT process). Any number of fast, access ROMs can be used here, as will be appreciated. Othe embodiments not concerned -with speed may of course use slo wer ROM technology .

[0029] The input bit width of the ROM-based phase-to-amplitude converter can vary from one embodiment to the next (e.g., from 4 to 24 bits), and the claimed invention is not intended to be limited to any particular input bit width range. Likewise,, the DDS system clock frequency can vary as well, and in some embodiments ranges from to 20 GHz to 40 GHz, but other embodiments may have a lower or higher clock frequency, depending on the given application.

[0930] In addition to the ROM technology, the phase-to-amplitude converter is further configured to adjust the digital timing (delay) of the data coming out of the. ROM with respect to the timing of the DAC The adjustment can be made based on the dela select control signal, as will be explained in turn. Among other things, the adjustable delay allows for different timing settings to be used depending on factors such as the proces variations, environmental conditions, and suppl voltages (e.g., emitter-coupled logic ^' or trimsistor-transistor logic supply voltages.). For instance, the adjustable delay can be used to compensate for variations in propagation delays through the ROM and/o DAC circuitry due to process vaiiations associated with HBT implementations and/or temperature in which the DDS is operating. [0031] The adjustable delay in this example embodiment is shown as being integrated into the ROM-based pbase-to~amplitude converter circuitry. However, in other embodiments, it may be implemented in a separate dedicated component that operatively couples with the ROM-based phase-to-amplitude converter. Moreover, the adjustable delay can be implemented For both the ROM and DAC, Alternatively, each of the ROM and DAC can have its own independent adjustable delay. Numerous configurations will be apparent in light of this disclosure. The ROM-based phase-to- ampiiiude converter will be discussed in more detail with reference to Figures 2a-b, 3, and 4

[0032j The digital amplitude valu output of the ROM-based phase-to-amplitude converter is applied to the DAC, which can be implemented as conventionally done, to convert the digital amplitude signal into its analog equivalent. The resolution, .of the DAC can var from one embodiment to the next, but in some example embodiments is in the range of 4 to 12 bits. The claimed invention is not intended, to be limited to any particular DAC configuration. Just as wit the ROM-based phase-to-ampliiude converter, the DAC: is referenced to the . clock signal and a clock delay associated with, the DAC can be adjusted, in accordance with an embodiment of the present invention.. Among other standard DAC componentr (e.g., resistive ladder network, current sources and switches, bias circuit, reference current, etc), the. DAC include one or more pipeline registers clocked b the clock signal, as adjusted by an adjustable clock delay circuit, in accordance with one example embodimen

[0033] ny number of variations on the DDS shown in Figure 1 will be apparent in light of this disclosure. For instance, other embodiments may include other componentry, such as a buffer circuit on the .output of the DAC to drive a 50-ohrn differential load ^■ off-chip or to some other target destination. n addition, the components may be integrated into a signal circuit (e.g., system on chip configuration), or may be implemented as separate modules that operatively couple to provide the desired DDS functionality. Other circuitry, such as processors or control circuitry for providing control and clock signals,, power supply circuitry, littering and/or signal shaping circuitry may also be provided as desired, depending on the application.

ROM-Based Phase-to-Amplitude Converter

[0034] Figure 2a i a block diagram showing further details of a ROM-based phase- to-amplitude converter, co figured in accordance with one -embodiment of the present invention. As can be seen, the converter in this embodiment includes a ROM that is operatively coupled to an adjustable clock delay circuit, which is configured to receive the DDS system clock signal as well as a delay select control signal, and to output a dela adjusted clock signal. In . this example case, the adjustable clock delay circuit output is also provided to the DAC clock input.

[0035) In an alternative embodiment, shown in Figure 2b. each of the ROM and DAC' is coupled to _. its own adjustable clock dela circuit. In particular, adjustable clock delay circuit I receives the clock signal and ROM delay select control signal, and Outputs the delay adjusted clock signal to the ROM. in a similar fashion, adjustable clock delay circuit 2 receives the clock signal and DA.C delay select control signal, and outputs the delay adjusted clock signal to. the DAC. As previously explained, the adjustable clock signal delays aliow for compensation of propagation delays associated with the corresponding ROM and DAC circuits and/or operating conditions. Using separate adjustable clock delay circuit allows for independent delay adjustment to ROM and DAC. J0036] The clock delay control signal is a digital signal tha -can be generated off-chip or by on-chip (or otherwise external of .internal to the DDS circuitry), for example, by a processor, and essentiall configures the adjustable clock delay circuit to provide the desired amount of delay. The number of available clock delay control signals can be set to 2^, where is the number of bit in the clock delay control signal in general, some of the clock delay control signals will result i poor o otherwise sub-optimal DDS performance where the timing relationship between the ROM and DAC is incorrect, and as a result digital glitches/wrong data are processed through the DAC (thereby causing, erroneous DDS outputs), and some of the clock delay control signals will result in good or otherwise sufficient DDS performance where the RO /DAC data is properl aligned and the correct data is processed through the DAC (thereby causing correct DDS outputs). Thus, in accordance with one embodiment of the present invention, a calibration routine can be■run where each of the available clock delay control signals are sequentiall applied in conjunction with, a known FCW, and the DAC output can be monitored for correct signal value. For any gi en set of test parameters (e.g., FCW, operating temp, clock frequency), at least one of the clock delay control signals will cause a best or otherwise optimal DAC output relative to DAC outputs caused b the other clock dela control, signals. Such a calibration routine can be carried out, for example, either manually as part of an initial .cheek of the DDS. and/or as part of an automated calibration that can be run periodicall or a requested.

[0037] The phase accumulator and DAC can be implemented as previousl discussed with reference to Figure 1, and that disclosure is equally applicable here. The ROM can be; implemented as conventionally done, and in one embodiment includes an address decoder, row and column select circuits, an array of memory cells, sense amplifiers, and one or more pipeline registers ( ^' clocked by trie clock signal outpu b the adjustable clock, delay circuit). As will be appreciated, the input bits from which the address of interest is decoded can b set based on th coding scheme and output bits of th phase accumulaior, and the output of the ROM can be configured to provide th number of bits desired, and may be. for example, thermometer-coded and/or binary coded. Example ROM configurations, may range from 4 to. 48 input bits and 4 to 64 output bits, but any other suitable configuration can be used. Note that the least significant bits of any stage digital output can be truncated or otherwise ignored, if so desired, t reduce the number of output bits to suit the input bit range of the next stage.

Delay Select Control Signal fDSCS)

[0038] Figure 3a is a block diagram showing a. circuit configured for generating a delay select control signal, in accordance, with one embodiment of the present invention,. As can. be seen, a processor is: configured to receive the FCW, clock, and DAG output, as well as input from one or more environmental sensors, such as temperature, humidity, altitude, and/or other input characterizing environmenta! variables thai have an impact on DDS performance and a change in which can be sensed. The processor furthe includes a look-up table (LUX) that includes a plurality of entries, each entry defining a DDS operatin parameter set (e.g., FC W, clock, and/o one or more -environmental variables) along, with a corresponding delay select control signal (DSGS) that provides optimal timing for tire ROM and DAC operation of the given DDS under that operating parameter set. Table 1 illustrates an example LUT. As will be appreciated, the LUT may include an number of entries to provide the resolution needed for the desired degree of accuracy for a given DDS application.

}0Θ39] The LUT entries can be. for example, empirically, developed by manually or automatically testing the given DDS for proper operation over a range of DDS operating parameter sets and testing each avaiiabie corresponding delay select control signal for each DDS operating parameter set, to identify the best performing delayselect control signal for a given parameter set. For instance, and with reference to the example of Table 1, if the DDS clock is set to 40 GH¾ FCW is set to 001 (assuming a 3-bit wide phase accumulator), and the operating temperature of the environment in which the DDS is operating is in the range of 35 to 45. °G, then a delay select control signal of 011 (assuming a 3-bit DSCS) is the one that provided the most accurate DAC output (for the known set of stimulus) of all the available delay select control signals. f0040j The processor can be implemented with any suitable processor: technology, such as a programmable microprocessor or a micrQeontro.il er programmed or otherwise, configured to carry- out the methodologies described herein. For example, a .microcontroller configured with a number of input/output ports and a number of embedded routines for -interrogating the current DDS operating parameter set and identifying -a suitable DSCS as described herein. The input ports can each receive signals indicative of a given, operating parameter, and an output port can provide the desired DSCS. The LUT can be stored in .a memory on-board or otherwise accessible by the microcontroller. The -microcontroller- may further include memory (e.g., ROM or flash) for storing th various routines as well as workspace {e.g., random access memory) for executing those routines. Any number of processo configurations can be used.

(004.1] Figure 3b illustrates a method for generating a delay select control signal, in accordance with one embodiment of the present invention. This method can be carried out, for example, by the processor of Figure 3 a. As can be seen, the method includes a mode for normal DDS operation, and a mode for DDS calibration where the delay select control signal is updated based on changes in the DDS operating parameter set. The method includes monitoring 301 DDS operating parameters of interest, such as FCW, clock speed, and temperature of environment. The method continues with

I S detenranitig; 303 if an parameter has changed. In one such example, each parameter can he associated with a pre-defined tolerance range, and any detection that a parameter i outside its tolerance range can be used to indicate a. parameter change. Slight parameter changes thai occur, but do not cause an out of tolerance reading can be treated as a non-change. If no parameter change is detected, then the method continues ^• with maintaining 305 the .current delay select control signal (DSCS). However, if a parameter change is detected, then the method continues with searching 307 a looku table for an entry associated with the new DDS operating parameter set. and outputting 309 a corresponding delay select: control signal associated with the new DDS operating parameter set,

|0042] Variations on th example embodiments shown in Figures 3a and. 3b will be apparent in light of this disclosure. For instance, in another embodiment, rather titan using a look-up table populated with ^• empirically developed data, the processor can be programmed to monitor for changes in the DDS operating parameter set, and when such a change is detected, the calibration mode can be entered where known stimulus {clock, f CW, etc) are input to the DDS (in the given environment of interest). During calibration mode, the DAC output can be monitored in real-time for accuracy as the available delay select control signals are sequentially cycled through. After all available dela selec control, signals are tested, the one providing the DAC output closest in value to the desired DAC output value can be used fo normal mode operation.

Adjustable Clock Delay Circuit

[0:043] Figure. 4a is a block diagram of an adjustable clock delay circui for use in a ROM-based phase-to-amplitude converter of a DDS system, configured in accordance with one embodiment of the present invention. As can be seen, the example adjustable clock delay circuit includes seven buffers (A through G) and three multiplexers (Mux 0, Mux 1. and Mux 2). The delay select control signal (DSCS) is 3-bits in this example, such that each of the three muxes is effectively controlled by one bit (i.e., Mux 0 by bit 0, Mux I by bit L and Mux 2 by bit 2).

[0044} This configuration allow for the selection of discrete delay increments, as provided by the buffers A-G. In general, each of the buffers A-G is associated with a fixed propagation delay, which can be exploited to delay the clock signal input by a known amount. The buffers can be implemented- with any suitable buffer technology, and in one example ease are implemented with unity gain buffers (voltage followers). Other embodiments may utilize buffers having amplifying or attenuating capability, if so desired. The multiplexers Mux 0, Mux L and Mux 2 can be implemented, for example, with a conventional high-speed multiplexer circuit design (e.g., such as the SN54/74LS 157), and are used to select the desired amount of delay.

[0045] In. operation, this example adjustable clock dela circuit allows for selection of 0 to 6 buffer delays, not counting buffer A (the delay of which is. constantly applied). Such an implementation can be used, for instance, to provide full clock coverage for DO systems operating up to 30 GHz. Note, however, that higher operating frequencies can be used by scaling the number of buffers and multiplexers accordingly. Table 2 demonstrates the operation of the adjustable clock delay circuit.

Tabk 2: Adjustable Delay

0046| As can be seen, if the DSCS bit is logic (},. then the corresponding multiplexer switches the L input to the multiplexer output, which is effectively a non-delay feed through path (i.e., no buffers on path). However, if the DSCS bit is logic 1 , then the corresponding multiplexer switches the H input to the multiplexer output, which will provide a unit of delay for each buffer on that path, clock delay of 1 unit (buffer A) is provided for all DSCS. Note that for a delay of 4 units, there are two possible buffer delay paths (generall designated as 4a or 4b in Table 2, where buffers A,B.C,D or A _s E,F,G, respectively, are switched into the delay path).

[0047] Figure 4b is a block diagram of an adjustable clock delay circuit for use in a ROM-based phase~to~anip!itude converter of a DOS system, configured in accordance with another embodiment of the present invention. In this example configuration, the adjustable clock delay circuit includes eight buffers (A through H) and seven multiplexers (Mux. 0.. Mux 1, Mux 2, Mux 3, Mux 4. Mux 5. Mux 6). The delay select control signal (DSCS) is 7-bits in this example, such that each of the seven muxes is effectively controlled by a corresponding one of the seven DSCS bits.

[0048] This configuration allows for the selection of discrete delay increments, as provided by the buffers A-H, each of which is associated with a fixed propagation delay that can be connected into the delay path by a combination of the multiplexers to delay the clock signal input by a known amount. The previous disclosure with respect to the buffers and multiplexers of Figure 4a is equally applicable here. This configuration is also scalable to cover additional .delays. Table 3. demonstrates the operation of the alternate adjustable clock delay circuit configuration of Figure 4b.

Table 2: Adjustable Delay

[0049] Just as with the embodiment of Figure 4a,. if the DSCS bit is logic 0 _? then the corresponding multiplexer switches the L input to the multiplexer output. However, if the DSCS bit is logic 1, then the- corresponding. -multiplexer .switches the H input to the multiplexer output. Note that a clock delay of 1 unit (buffer A) is provided fo all DSCS. Further note the layout symmetry- and duplicate conductor length for each of buffers B~H and muxes 0.-6, such that the delay adjustment is predominantly provided b the buffers (rather than the conductor coupling the buffers and: muxes from the clock input to the clock output).

[0050] The foregoing description of the embodiments of the invention has been presented for the purposes of illustration and description. It is not intended to be exhaustive or to limit the invention to the precise form disclosed. Many modifications and variations are -possible in light of this disclosure, it is intended that the scope of the invention be limited not by this detailed description, but rather by the claims appended hereto.