| Magazines: AutoSpeed | Rallysport News | V8X | Silicon Chip  | Shopping: Property | Cars | Car Accessories | Electronics/IT |
![]() |
![]() |
| Current Articles | | | Archived Articles | | | Search | | | Free Preview | | | Register |
![]()
Frequency meters are used in virtually all areas of electronics and are invaluable for servicing and diagnostics. Among other things, they are ideal for checking the operation of oscillators, counters and signal generators. They can also be used for servicing RF equipment or to simply provide an accurate frequency readout for a function generator. This new 50MHz Frequency Meter is autoranging and displays the frequency in either Hz, kHz or MHz. This makes the unit easy to read, as it automatically selects the correct range for any frequency between 0.1Hz and 50MHz and inserts the decimal point in the correct place for each reading. The design is easy to build too, since it uses a programmed PIC microcontroller to do all the clever stuff. Apart from that, there's an LCD readout, a couple of low-cost ICs, two transistors, a 3-terminal regulator and a few sundry bits and pieces to complete the design. Note that although we have specified this Frequency Meter at 50MHz maximum, most units will be capable of measuring frequencies somewhat higher than this. In fact, our proto-type meter was capable of making frequency measurements to above 64MHz.
LCD readoutA feature of this unit is the use of a 2-line 16-character Liquid Crystal Display (LCD) to show the frequency reading. This has several advantages over LED displays, including much lower current consumption. This allows the unit to be operated from batteries if required. In addition, the LCD can show all the units without resorting to the use of separate annunciators, as would be required with a LED display. Resolution modesTwo resolution modes are available: (1) a low-resolution mode which has fast updates and is suitable for most measurements; and (2) a high-resolution mode which can be selected when greater precision is required. In the low-resolution mode, the resolution is 1Hz for frequencies from 1-999Hz and 10Hz for frequencies above this. The corresponding display updates time are 1s from 1-999Hz and 200ms from 1kHz-50MHz. By contrast, the high-resolution mode provides 1Hz resolution for frequencies from 150Hz-16MHz. Above 16MHz, the resolution reverts to 10Hz. The display update time is 1s. Below 150Hz in the high-resolution mode, the display has 0.1Hz resolution and a nominal 1s update time for frequencies above 10Hz. This 0.1Hz resolution makes the unit ideal for testing loudspeakers, where the resonance frequency needs to be accurately measured. Note, however, that the update time takes longer than 1s for frequencies below 10Hz.
This view shows the completed PC board for the 50MHz frequency Meter (DSE version). Note that the BNC input socket is shown soldered directly to the signal input PC stakes in this photo but this was only done for test purposes. In reality, this socket is mounted on the side of the case and connected to the signal input PC stakes on the underside of the board via a short length of 75-ohm coax.
The two resolution modes are toggled from one to the other by pressing the Resolution switch. The meter then displays either "Resolution LOW" or "Resolution HIGH" to indicate which mode is currently selected. In addition, the selected resolution mode is stored in memory and is automatically selected if the meter is switched off and on again. In the low-resolution mode, the display will show 0Hz if the frequency is below 1Hz. By contrast, in the high-resolution mode, the display will show "No Signal" for frequencies below 0.1Hz. If the frequency is below 0.5Hz, the display will initially show an "Await Signal" indication before displaying the frequency. If there is no signal, the display will then show "No Signal" after about 16.6s. The 0.1Hz resolution mode for frequencies below 150Hz operates in a different manner to those measurements made at 1Hz and 10Hz resolution. Obtaining 0.1Hz resolution in a conventional frequency meter normally means measuring the test frequency over a 10s period. And that means that the update time is slightly longer than 10s. This 10s update time is a very long time to wait if you are adjusting a signal generator to a precise frequency. However, in this frequency meter, the display update period is 1s for frequencies above 10.0Hz, increasing gradually to 10s for frequencies down to 0.1Hz. So for normal audio frequencies, the display will update at 1s intervals. Just how this is achieved is explained below, when we describe the block diagrams for the unit. PresentationAs shown in the photos, the 50MHz Frequency Meter is presented as a "standalone" unit that's housed in a small plastic case. As mentioned, it can be powered using either a 9-12V DC plugpack or a 9V battery. There are just two controls on the front panel: an on/off switch and the "Resolution" pushbutton. In addition, a DC input socket is mounted at one end of the box, while the signal input connects to a panel-mounted BNC socket on one side. Alternatively, the unit could be added to an existing piece of equipment to provide accurate frequency readout. Its low current requirements mean that it can usually be connected to an existing supply rail inside the equipment. Block diagramsFig.1 shows the general arrangement of the frequency meter. It's based mainly on the microcontroller (IC3). In operation, the input signal is processed and applied directly to a divide-by-256 prescaler that's internal to IC3. The divided signal then clocks timer TMR0 which counts up to 256 before clocking Register A.
Fig.1: the block diagram of the 50MHz Frequency Meter for "normal" frequency measurements. The incoming signal is first amplified, then fed through a gating circuit to clocking stage IC2b. This then drives a divide-by-256 prescaler inside microcontroller IC3. (ie, at the RA4 input).
Register A is an 8-bit register which counts up to 256 before returning to zero. Combining all three counters (the prescaler, TMR0 and register A) allows the circuit to count up to 24 bits, or a total of 16,777,216 counts. By counting over a 1s period, it follows that the unit can make readings up to about 16.7MHz. However, if the frequency is counted over a 100ms period, the theoretical maximum that can be measured is just over 167MHz. As shown in Fig.1, the input signal is first boosted using an amplifier to a level sufficient to drive gating stage IC2a. This, in turn, drives clocking stage IC2b which is controlled by IC3's RA3 output. Normally, IC2b allows the signal to pass through to the prescaler at IC3's RA4 input. IC3's RB2 output controls gating stage IC2a so that signal passes through for either a 100ms period or a 1s period. During the selected period, the signal frequency is counted using the prescaler, timer TMR0 and register A. Initially, the prescaler, the timer and register A are all cleared to 0 and the RB2 output is then set to allow the input signal to pass through to the prescaler for the gating period (ie, for 100ms or 1s). During this period, the prescaler counts the incoming signal applied to RA4. Each time its count overflows from 255 to 0, it automatically clocks timer TMR0 by one count. Similarly, when ever the timer output overflows from 255 to 0, it sets a Timer Overflow Interrupt Flag (TOIF) which in turn clocks Register A. At the end of the gating period, IC3's RB2 output is cleared, thus stopping any further signal from passing through to the prescaler. The value of the count in TMR0 is now transferred to Register B. Unfortunately, the value in the prescaler cannot be directly read by IC3 and so we need to derive the value. This is done by first presetting register C with a count of 255. That done, the RA3 output is taken low to clock the prescaler and timer TMR0 checked to see if it's count has changed. If TMR0 hasn't changed, the prescaler is clocked again with RA3. During this process, register C is decreased by 1 each time the prescaler is clocked. The process continues, with RA3 clocking the prescaler until timer TMR0 changes by one count. When this happens, it indicates that the prescaler has reached its maximum count. The value in Register C will now be the value that was in the prescaler at the end of the counting period. The processing block now reads the values in registers A, B and C. Based on this information, it then decides where to place the decimal point and whether to show Hz, kHz or MHz. The required value is then written to the LCD via the data and control lines (RB4-RB7 and (RA0-RA2). Alternative configuration
Fig.2: this is the alternative configuration for making high-resolution (ie, to 0.1Hz) measurements below 150Hz. In this case, the input signal is applied to the RA4 input as before. However, the prescaler is no longer clocked by the RA4 input but by an internal 1MHz clock instead.
If the input signal frequency is greater than 16MHz and the gating period is 1s, register A will initially have overflowed. In this case, the gating period is automatically changed to 100ms. Alternatively, if the high-resolution mode is selected and the frequency is below 150Hz, the frequency meter changes its configuration to that shown in Fig.2. In this case, the input signal is applied to the RA4 input as before. However, the prescaler is no longer clocked by the RA4 input but by an internal 1MHz clock instead. Basically, what happens is that the RA4 input is monitored for a change in state - ie, from a low voltage to a high voltage - which indicates a signal at the input. When this happens, the prescaler is cleared and begins counting the 1MHz internal clock signal. The overflows from the prescaler and timer TMR0 are carried to Register A as before. Counting continues until the input signal goes low and then high again, at which point counting stops. If the counting causes register A to overflow, then the display will show no signal (this will happen after 16.7s if the signal does not go low and high again). Conversely, if the counting is within range, the prescaler value is determined by clocking IC2b using the RA3 output as before.
The PC board is secured by plugging it into the matching header pins on the LCD module and installing four screws to fasten it to the spacers. Note the mounting method for VC1 (circled in red).
From this, it follows that if the input frequency is 1Hz (ie, a 1s period), the value in the A, B and C registers will be 1,000,000. That's because the prescaler is clocked at 1MHz for 1s. Similarly, the count will be 100,000 for a 10Hz signal and 10,000 for a 100Hz input signal. Finally, the value in the registers is divided into 10,000,000 and the decimal point placed immediately to the left of the righthand digit. This gives a direct readout in Hz with 0.1Hz resolution on the LCD. Note, however, that this technique can not be used for measuring very high frequencies. That's because the value in the counter becomes smaller as the frequency increases and so we begin to lose accuracy. For example, at 500Hz, the counted value would be 2000 and at 500.1Hz the counted value would be 1999. The result of the division of 1999 into 10,000,000 would be 500.2 instead of the 500.1 required. The 0.1Hz resolution has therefore been restricted to a maximum of 150Hz to ensure accuracy of the calculation. Circuit detailsRefer now to Fig.3 for the full circuit details. As shown, the input signal is AC-coupled to the unit via a 470nF capacitor to remove any DC component. This signal is then clipped to about 0.6V peak-to-peak using diodes D1 & D2, with current limiting provided by the 100kΩ series resistor. The 22pF capacitor across the 100kΩ resistor compensates for the capacitive load of the diodes.
Fig.3 (left): the circuit is based on microcontroller IC3. This processes the signals from the preceding amplifier stages and drives the LCD. Power comes either from a 9-12V DC plugpack or from a 9V battery.
From there, the signal is fed to the gate of Q1, a 2N5485 JFET. This transistor provides a high input impedance, which is necessary to ensure a wide frequency response. Q1 is self-biased using a 910kΩ resistor from gate to ground and a 470Ω source resistor. It operates with a voltage gain of about 0.7, which means that the signal is slightly attenuated at the source. This loss is more than compensated for in the following amplifier stages. Next, the signal is AC-coupled to pin 4 of amplifier stage IC1a via a 100μF electrolytic capacitor and a parallel 10nF capacitor. The 100μF capacitor is sufficiently large to allow for a low frequency response of less than 1Hz. However, this capacitor loses its effectiveness at higher frequencies due to its high internal inductance and the signal is coupled via the 10nF capacitor instead. IC1a is one of three differential line receivers in a single MC10116N IC package. It's biased via the DC output at pin 11 and this is decoupled using a 10μF electrolytic capacitor and a paralleled 10nF ceramic capacitor. The voltage is then applied to the wiper of trimpot VR1 (Offset Adjust) and this allows adjustment of the input bias voltage. In operation, IC1a is run open loop (ie, without feedback) so that it provides as much gain as possible. Even so, it only operates with a voltage gain of about seven times. It's differential output signals appear at pins 2 & 3 - ie, one output is opposite in phase to the other. These outputs are in turn applied to the differential inputs (pins 12 & 13) of IC1b. Note that the differential outputs have 470Ω pulldown resistors, as they are open emitters. In fact, the MC10116 IC is an emitter-coupled logic (ECL) device. Unlike IC1a, IC1b has negative feedback and this is provided by the two associated 100Ω resistors. This reduces the gain of this stage to just under two. The third stage using IC1c differs in that it employs positive feedback and so it functions as a Schmitt trigger rather than as an amplifier. Its hysteresis is around 450mV which means that the signal swing on its differential inputs must be greater than this in order for this stage to provide an output. In operation, the output swing at pins 6 & 7 is from 4.3V when high to 3.4V when low. This needs to be level-shifted to provide for normal CMOS input levels to the gating circuit (IC2a) and this is done using PNP transistor Q2. It works like this: when pin 6 is high at 4.3V, Q2's base is also at 4.3V, which is just 0.7V below the +5V supply rail. However, Q2 must have a base voltage that's at least 1.2V below the +5V rail in order to switch on - ie, to overcome the 0.6V "diode-drop" across D3 plus a 0.6V base-emitter voltage. As a result, when pin 6 if IC1c is high, Q2 is off and the 330Ω resistor at Q2's collector holds the output low. Conversely, when pin 6 of IC1c goes low (3.4V), transistor Q2 turns on and pulls pin 1 of IC2a high. IC2a is a Schmitt NAND gate. It inverts the signal on its pin 1 input when pin 2 is held at +5V by IC3's RB2 output (ie, the signal passes through to the pin 3 output but is inverted). Conversely, when RB2 is at 0V, IC2a's pin 3 output remains high and the input signal is blocked. So, in summary, the signal is allowed through to IC2b when RB2 is high and is blocked when RB2 is low, as described previously. IC2b normally has its pin 5 input held high via IC3's RA3 output, so that the signal from IC2a is again inverted at pin 6. When RB2 is brought low, pin 3 of IC2a remains high and so pin 4 of IC2b is also high. This allows RA3 to clock the RA4 input via IC2b. Driving the LCDIC3's RA0-RA2 outputs drive the control inputs to the LCD module and select the line and the position of the character to be displayed. Similarly, RB4-RB7 drive the data inputs (DB4-DB7) on the LCD module. A 470pF capacitor on the E-bar (enable control line) is included to slow down the rise and fall times of the square wave from IC3, which are nominally too fast for the LCD module to handle - particularly when the ambient temperature is well below 25°C. A 4MHz crystal connected between pins 15 & 16 of IC3 provides the clock signals for IC3. The recommended crystal has low drift but a standard 4MHz crystal could be used if accuracy is not critical. The capacitors at pins 15 & 16 provide the necessary loading for the crystal so that runs at the correct frequency, while VC1 also allows the clock frequency to be "tweaked" slightly to provide calibration. Power supplyPower for the circuit is derived from either a 9-12V DC plugpack or a 9V battery (but not both). Diode D4 protects the circuit against reverse polarity protection when using a plugpack supply, while regulator REG1 provides a +5V supply rail to power the circuit. If a 9V battery is used, it connects to the cathode side of D4; ie, it bypasses the reverse polarity protection. This means that D4 can be left out of circuit (along with the DC socket) if the unit is to be battery powered. ConstructionThe SILICON CHIP 50MHz Frequency Meter can be made in one of three versions, depending on where you buy the kit. That's because the LCD modules available from Dick Smith Electronics (DSE), Altronics and Jaycar are all different and so a different PC board has been designed to suit each module. These boards are coded 04108031 (DSE), 04108032 (Altronics) and 04108033 (Jaycar).
Fig.4: three different PC boards have been designed to suit the different LCD modules that are available from DSE, Altronics and Jaycar. Just follow the parts layout that's applicable to your version.
Each LCD plugs directly into its intended PC board, which means that there are no external wiring connections except to the BNC input socket. And in case you are wondering, there are no performance differences between the three versions. The unit is housed in a plastic case measuring 130 x 67 x 44mm, with the LCD module protruding through a cutout in the front panel. The Dick Smith version has the power switch on the righthand side and the signal input applied to the socket at the top left of the box. By contrast, both the Altronics and the Jaycar versions have the power switch at the top left, while the input socket is mounted on the lower right of the box. This difference comes about because the display readout for the DSE LCD module is upside down compared to the other two modules in relation to the input terminals. The unit shown in the photos is for the DSE version but both the Altronics and Jaycar modules were fully tested. Fig.4 shows the PC board layout for each of the three versions. Begin by checking that you have the correct PC board for the LCD module you are using. That done, check the mounting holes for the LCD module against those on the PC board (the holes must be 3mm in diameter). Check also that holes are large enough to mount switch S2 and the DC input socket. Next, install all the wire links and resistors, using the accompanying resistor colour code table as a guide to selecting each value. It's also a good idea to check the resistors with a digital multimeter just to make sure. IC1 and IC2 can go in next, taking care to ensure that they are correctly oriented. That done, install a socket for IC3 but don't install the microcontroller just yet. The diodes and capacitors can now all be installed, followed by REG1 and transistors Q1 & Q2. Note that the 100μF and 10μF capacitors in the Altronics version must be installed with their bodies parallel to the PC board, so that they don't later foul the LCD module. It's just a matter of bending their leads at right angles before installing them on the board. Similarly, the top of transistor Q2 must be no higher than 10mm above the PC board to prevent it from interfering with the LCD module (all versions). The next step is to install the socket for the LCD module. Both the DSE and Altronics versions use a 28-pin DIL IC socket which is cut in half to obtain a 14-way strip socket which is then soldered in place. By contrast, the Jaycar version uses a 14-pin IC socket which is cut into two 7-way strips which are then installed side-by-side.
Fig.5: this diagram shows how the unit is installed inside the case. Be sure to use Nylon spacers where indicated.
Once the sockets are in, install PC stakes for the "+" and "-" supply connections (near D4) and for the signal input and GND connections. These PC stakes should all be installed from the copper side of the board. PC stakes are also used to mount switch S1. These should be trimmed so that when the switch is mounted, its top face is 20mm above the top surface of the PC board. Be sure to orient S1 with its flat section facing towards the right, as shown in Fig.4. The remaining parts can now be installed on the board. These parts include switch S2, the DC socket, trimpots VR1 & VR2, crystal X1 and trimmer capacitor VC1. Note that VC1 is mounted on the underside of the PC board, so that it can be adjusted without having to remove the LCD module. Front panelThe front panel (ie, the case lid) must be drilled and a cutout made to accommodate the two switches and the display. However, if you have purchased a kit, then you probably won't have to worry about this. If you're preparing the case yourself, you can use one of the front panel artworks as a drilling template (see Figs.6 & 7). You can make the display cutout by first drilling a series of holes around the inside perimeter of the rectangle, then knocking out the centre piece and filing the job to a smooth finish. It will also be necessary to drill the mounting holes for the LCD module. Note that these should be countersunk so that the intended screws sit flush with the surface of the lid - see Fig.5. That done, the adhesive label can be attached to the panel and the cutouts made using a utility knife Kit versions will probably be supplied with screen-printed labelling. In that case, countersunk screws will no longer be necessary. Checkout timeNow for an initial smoke test - ie, before IC3 or the LCD are plugged in. First, apply power and check that there is +5V on pin 16 of IC1, pin 14 of IC2 and pins 4 & 14 of IC3. If this is correct, disconnect power and install IC3 in its socket, taking care to ensure it goes in the right way around. That done, plug the LCD module into its matching socket and temporarily fit a couple of 10mm tapped Nylon spacers to support it on the PC board. Next, reapply the power again and check that the display shows either 1Hz or 0Hz. If not, adjust VR1 so that the display shows 0Hz when the signal input terminals are shorted. VR2 can then be adjust for best display contrast. Now press the Resolution switch - the display should show "Resolution HIGH". It should then show "Await Signal" when the switch is released. If the switch is then pressed again, the display should show "Resolution LOW". Final assemblyThe Frequency Meter can now be installed in its case.
The LCD module is secured to the lid of the case using four M3 x 6mm cheesehead screws, four M3 nuts and four M3 x 10mm tapped Nylon spacers.
Fig.5 shows the assembly details. As shown the LCD module, is secured to the case lid using four M3 x 10mm CSK screws, four M3 nuts (used as spacers) and four 10mm-long tapped Nylon spacers. The PC board is then secured to the bottom ends of the four spacers. You will have to drill a 9mm-diameter hole in one side of the box to provide access to the DC socket if you are powering the unit from a plugpack. This hole should be positioned midway along one side and about 6mm down from the top edge of the case. Conversely, if the unit is to be battery powered, you will need to solder a battery clip lead to the supply PC stakes on the underside of the board. The battery can be secured to the bottom of the case by mounting it in a suitable holder. Alternatively, you could dispense with the holder and simply wrap the battery in some insulating material and wedge it between the PC board and the bottom of the case. The BNC input socket is mounted on one side of the case towards the base and wired using 75Ω cable to the two signal input PC stakes on the underside of the PC board. CalibrationThe completed 50MHz Frequency Meter can be calibrated against the 15.625kHz line oscillator frequency in a colour TV set. Fortunately, you don't need to remove the back of the set to do this. Instead, all you have to do is connect a long insulated wire lead to the input socket and dangle it near the back of the TV set. It's then just a matter or adjusting VC1 so that the meter reads 15.625kHz when the resolution is set to "High" mode. Note: the TV must be showing a PAL program, not NTSC (15.750kHz). If there is insufficient adjustment on VC1 to allow calibration, the 33pF capacitor at pin 15 of IC3 can be altered. Use a smaller value if the frequency reading is too high and a larger value if the frequency reading is too low. Usually, the next value up or down from 33pF will be sufficient - ie, use either 27pF or 39pF. If you require greater accuracy, the unit can be calibrated against the standard 4.43MHz colour burst frequency that's transmitted with TV signals. The best place to access this frequency is right at the colour burst crystal inside a colour TV set. This crystal will usually operate at 8.8672375MHz (ie, twice the colour burst frequency), although some sets use a 4.43361875MHz crystal. Be warned though: the inside of a colour TV set is dangerous, so don't attempt to do this unless you are an experienced technician. There are lots of high voltages floating around inside a colour TV set and you could easily electrocute yourself if you don't know what you are doing.
This photo clearly shows the location of the access hole for the DC input socket for the DSE version. It's located on the opposite side of the case for the Altronics and Jaycar versions.
In particular, note that much of the circuitry in a switchmode power supply circuit (as used in virtually all late-model TV sets) operates at mains potential (ie, many of the parts operate at 240VAC). In addition, the line output stages in some TV sets also operate at mains potential - and that's in addition to the lethal EHT voltages that are always present in such stages. Note too that some TV sets (particularly older European models) even have a "live" chassis, in which all the circuitry (including the chassis itself) operates at mains potential. Usually, there will be a label on the back of the set advising of this but don't take it for granted. Don't even think of messing about with this type of set. In short, don't attempt the following calibration procedure unless you are experienced and know exactly what you are doing. OK, assuming that you know what you are doing (and the set has a grounded chassis), you will need to make up an insulated probe with a 10MOmega; resistor in series with the input plus a ground lead. This probe can then be connected to one side of the colour burst crystal and VC1 adjusted so that the meter reads either 8.867237MHz or 4.433618MHz (resolution set to high mode). Make sure that the probe has no affect on the colour on the TV screen when it is connected to the colour burst crystal. If it does, it means that the probe is loading the crystal and altering its frequency. In that case, try connecting the probe to the other terminal of the crystal. That's it - your new 50MHz frequency Meter is now calibrated and ready for action.
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|
|
|||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||||
|