A Simplified GPS-Derived Frequency Standard - …

1 14 Sep/Oct 2006242 Robert-Martial StGatineau, Qu bec J9J Simplified GPS-DerivedFrequency StandardBertrand Zauhar, VE2 ZAZHere is a simple and modern approach to a 10-MHz many reasons, an accurate frequencystandard at an Amateur Radio stationis desirable. For weak-signal operationsuch as EME (moonbounce) at microwavefrequencies, you must be transmitting and lis-tening exactly at the right Frequency ; other-wise the narrow receive filters used will makeyou miss that weak signal. With such a setup,a 10-MHz Frequency reference feeds the fre-quency synthesizer that generates the radio soperating Frequency . An accurate 10-MHzreference is also useful for test equipmentadjustment.

2 With an accurate Standard , youcan put Frequency counters and signal gen-erators on advent of the Global Positioning Sys-tem (GPS) has allowed a Simplified approachto time and Frequency accuracy. Several com-mercially available GPS receiving units pro-vide a 1 pulse-per-second (pps) signal. Thissignal typically exhibits a short-term accu-racy of 1 microsecond (1 ppm or 1 10 6).Figure 1 Block diagram of the GPS-Derived Frequency with permission; copyright ARRL Sep/Oct 2006 15By averaging it over a long period, a muchbetter accuracy can be achieved. This is whatthis project does: it locks an external 10-MHzvoltage-controlled signal source to the 1 ppsGPS good work of Brooks Shera1 has gen-erated a lot of interest within a broad commu-nity of experimenters who want to increase thelevel of Frequency accuracy available to themat low cost.

3 His system uses a PLL techniqueto lock an external oscillator to a GPS receiverand obtain an accurate Frequency project I present here provides asimpler and more modern approach to aGPS- derived 10-MHz Frequency found in today s technologyoffer the following benefits: solid perfor-mance, more features and a reduction in thenumber of design differentiates itself from otherpreviously published designs because: It uses a simpler Frequency measurementtechnique, as opposed to phase measure-ment. It provides on-board reference buffering andfan-out with 50- output impedance. It provides the three most common refer-ence frequencies of 10 MHz, 5 MHz and1 MHz. It provides full software control of the fre-quency acquisition and control processes,without DIP switches.

4 It has fewer components and does not re-quire an external DAC or external inputcounter chips. It runs off only one supply voltage: +5 V dc(excluding the VCXO supplies).Tests have shown that this system consis-tently produces a short-term reference accu-racy in the 1 10 10 range. This is derived us-ing Standard automotive-grade GPS range of accuracy does not rival cesium-based references. It is much better than mostof the Standard built-in, free-running oscilla-1 Notes appear on page 2 Schematic diagram of the GPS-Derived Frequency Standard Sep/Oct 2006tors seen in commercial test instruments, how-ever. Just to give you an idea of the type ofaccuracy, one part in 1010 represents an errorof one hertz on a 10 GHz signal!

5 System DescriptionFigure 1 shows a block diagram of myGPS- derived Frequency Standard . The systemoperates a hardware/firmware Frequency -locked loop (FLL). In essence, the system com-pares a local Frequency source (an externaloscillator) to a GPS-Derived reference. It willadjust the local 10-MHz variable source tomatch the GPS-Derived 1 pps reference. The10-MHz source is kept aligned with respectthe 1-Hz GPS reference on a real-time basisby the firmware. The resulting 10-MHzreference is fanned out, Frequency -divided andprovided to the user for high-accuracy appli-cations. System control and monitoring areachieved using a bicolor LED and a serial portconnected to a terminal (PC).

6 Hardware DescriptionFigure 2 shows the system circuit sche-matic. The main operation consists of count-ing the number of rising edges produced bythe 10-MHz voltage-controlled crystal oscil-lator (VCXO) signal over a 16 s period (16 GPS pulses). If the GPS and the VCXO areat the same Frequency , exactly 160,000,000pulses will be counted ( 1 pulse, inherent tocounter technology).Prior to entering the microcontroller, the10-MHz VCXO signal is buffered and am-plified by U1, a receiver chip. An optionalinput termination resistor R1 can be added ifthe VCXO s output circuit calls for one. Thebuffered 10-MHz reference signal is fannedout to several locations on the , the Microchip PIC18F2220 micro-controller, has a built-in 16-bit counterincremented by an external source, theVCXO.

7 The counter value is latched by an-other external signal rising edge, the GPS1 pps signal in our application. This processis totally autonomous and independent fromfirmware. The microcontroller s task in thisprocess is to analyze the results and adjustthe VCXO Frequency Frequency ControlThe Microchip PIC18F2220 microcon-troller does not have an integrated digital-to-analog converter (DAC). To produce anadjustable voltage source to vary the VCXO Frequency , the built-in 10-bit pulse-widthmodulator (PWM) is used instead. A continu-ous rectangular-wave output is produced by thePWM. A downstream external 1-Hz, two-stagelow-pass filter (U5A, U5B and discrete com-ponents) is used to recover the average dcvalue of the PWM output.

8 By varying the dutycycle of the PWM, it is possible to producean accurate analog dc voltage with 210 or 1024steps over the DAC of 14-bit resolution is achievedby precisely controlling the duty cycle of thePWM output. This translates to a tuninggranularity of 6 10 5 Hz for a VCXO thathas a 1-Hz tuning range. Achieving a 14-bitDAC using a 10-bit PWM requires additionalfirmware processing. The idea is to dither the pulse width within a 16-cycle 16 cycles translate into an additional4-bit resolution. For example, increasing the14-bit DAC output by one step involves in-creasing the 10-bit PWM output width by oneincrement on one of the 16 pulses. Increas-ing the DAC by two steps means increasingthe 10-bit PWM output width by one incre-ment on two of the 16 pulses, and so add flexibility for interfacing with vari-ous VCXOs, the filtering stages have a supplybypass feature that allows you to feed theoperational amplifiers with different upperand lower rail voltages.

9 This is done byreconfiguring JP2 and C17. Remember,though, that the maximum voltage differencebetween upper and lower rails must be kept to12 V or less. Another feature, the second stageof low-pass filtering, allows for additional gainusing R7/R8 if the VCXO operates on a largertuning voltage range than the more standard5-V range. A 5-V offset can also be added tothe second stage using jumper JP3. This pro-vides support for VCXOs that have a 5 V to+5 V tuning 1 lists some of the possible configu-rations on the filtering stages for variousVCXO tuning ranges. Finally, the tuningslope sign can be set in firmware to accom-modate both types of ReferencesThe system provides up to four 10-MHzreference signals.

10 In addition, it provides upto two references with a selectable frequencyof either 5 MHz or 1 MHz. These sub-ratesare produced by U2, a synchronous active sub-rate is selected with an on-board jumper, JP1. All references are of50- output impedance and provide an am-plitude of greater than 1 V pk-pk with asquare wave shape. These signals are pro-vided by U3, a line driver the firmware feature is enabled, thereference outputs are inhibited if the FLLgoes into its unlocked state. That protectionensures that the user does not use a referenceof unknown quality. LED D2 provides anindication of the reference output Status LEDThe system provides basic FLL status andalarm conditions with a single bicolor LED(D1).