Fullscreen HTML5Med Res (??MB)HTML4

Design by EUGENE W. VAHLE JR.* Digital electrolytic capacitance meter Do you need to check large value electrolytic capacitors? Unfortunately, you can’t do it with the capacitance ranges on your digital multimeter or even with most capacitance meters. You need this special purpose instrument which can measure electrolytic capacitors ranging from around 10mF up to as high as 999,900mF – yep, almost 1 Farad. The problem with electrolytic capacitors is threefold. First is the sheer value of capacitance which typically ranges from around 1µF to many thousands of microfarads. Normal measur­ ing techniques which essentially measure the capacitor’s im­ pedance at a particular frequency just don’t work. Because the capacitance is so high, the impedance is just too low to measure unless you use quite low test frequencies or resort to special circuit techniques. Second, electrolytic capacitors need to be charged (or polarised) to present a reliable and consistent capacitance value. Third, compared with every other type of capacitor, elec­ trolytics can have quite a high leakage current and this can confuse a normal capacitance measuring instrument. So how does this instrument get around these problems? Instead of trying to measure impedance with a test frequency, this circuit measures the time taken to charge the capacitor to a particular voltage. It is based on the following formula for capacitance: C = Q/V where C is capacitance in Farads, V is the voltage applied to the capacitor and Q stands for charge in Coulombs. Without getting too technical, if we pump charge into a capacitor at a September 1999  63 CLOCK IC2 CONSTANT CURRENT SOURCE Q9 4-DIGIT COUNTER IC1 7-SEGMENT LED DISPLAYS COMPARATOR Q5, Q6 Constant-current source CUT Fig.1: the block diagram of the Electrolytic Capacitance Meter shows a constant current source to the charge the capacitor under test (CUT), a comparator and the 4-digit counter. The meter measures the time period to charge the test capacitor to 4V. known rate, the time taken to reach a certain voltage is directly proportional to the capacitance. We could write this as an equation too but suffice to say that pumping charge into a capacitor at a known or fixed rate is exactly the same as charging it with a constant current source. And that is exactly what this circuit does, as shown in the block diagram of Fig.1. The constant current source charges the capacitor under test (CUT) while the counter is clocked. When the capacitor reaches a particular voltage, a comparator stops the counter and the displayed value is the capacitance. Pretty simple, eh? Our Electrolytic Capacitance Meter is quite simple. It has a 4-digit 7-segment LED display, a 4-position range switch, a toggle switch with Test and Discharge positions and the terminals for the capacitor. So let’s test a capacitor. First, turn the unit on, connect a capacitor to the terminals, making sure that the negative lead goes to the black terminal and flip the toggle switch to the discharge setting. If the capacitor has some charge in it, the red LED will come on briefly and then go out, to signify that the capacitor is now discharged. Now flip the toggle switch to the test position and the 4-digit display will start counting up from zero. Depending on the value of the capacitor, the count will stop after a few seconds and the value shown is the capacitance in microfarads. What about the range switch? It has four settings: x0.1, x1, x10 and 100. So if the displayed value is 1500, for example, and the range switch is set to x1, then the value is 1500µF. Most electrolytic capacitors have large tolerances, ranging from -20% 64  Silicon Chip DC (Vcc). This voltage is applied to the two 5V regulators (REG1 & REG2). REG1 is wired in convention­al fashion and produces a +5V output. REG2, on the other hand, is wired in an unconventional manner which we’ll explain shortly. to +80%, meaning that a capacitor specified as 1000µF might have an actual capacitance of as little as 800µF (-20%) or as much as 1800µF (+80%). Other capacitor types have a lower tolerance (±10%). For capacitors with substantial series resistance (such as double layer capacitors used in memory backups), the formulas provided later on in Table 1 can be used to find the actual capacitance and series resistance. About the circuit Refer now to Fig.2 for the complete circuit of the Electro­lytic Capacitance Meter. As shown, it uses two integrated cir­cuits: IC1, a 74C926 4-digit counter/multiplexed 7-segment dis­ play driver and IC2, a 7555 CMOS oscillator/timer. In addition, there are 10 transistors (Q1-Q10), two 5V regulators (REG1 & REG2), a bridge rectifier (D1-D4), two light-emitting diodes (LED1 & LED2) and four 7-segment LED displays. Power for the device comes from a 12V AC plugpack. Its output is rectified by diodes D1-D4 and filtered by a 470µF capacitor to give about 18V Special Notice *This project and article has been adapted with permission from an article in the May 1999 issue of the American magazine Popular Electronics. The original design did not have a PC board and this has been produced by SILICON CHIP staff. The Popular Electronics design was also based on the 74C925 instead of the 74C926 used here since it is more readily available. The output from the bridge rectifier is also applied to a range-select resistor network via S2a and then to the emitter of transistor Q9. These components, in company with REG2, form a rather odd-looking constant-current source. Let’s see how it works. REG2 is a 7905 -5V regulator. Usually, the GND terminal of a 3-terminal regulator is referenced to GND or the 0V rail in a circuit but in this case, the input (IN) is grounded while the GND terminal is jacked up to +18V by connecting it to the bridge rectifier output. Because the regulator delivers a -5V rail with respect to the GND terminal, this means that the output will be at +13V (ie, Vcc - 5V). As a result, Q9’s base is held at a constant 5V below the Vcc rail and so its emitter maintains a constant voltage across the selected range resistor. This causes Q9 to function as a constant current source. After subtracting the 0.6V developed across the base-emit­ter junction of Q9, the voltage across the selected range resis­tor will be approximately 4.4V. This means that the current through the selected range resistor will be 44µA for the x0.1 range, 440µA for the x1 range, 4.4mA for the x10 range and 44mA for the x100 range. A 1µF capacitor is included to filter the output of REG2, while the parallel 2.2kΩ resistor sets the minimum load on REG2’s output. This is done because on the 44µA (x0.1) range, the base current needed for Q9 is very small (around 0.4µA). Comparator stage Let’s now take a look at the comparator circuit which is based on Q5 & Q6. First, a +4.5V reference voltage is derived from a 220Ω/2kΩ voltage-divider network across the output of REG1. This reference voltage is applied to the base of Q10 (which means that Q10’s emitter will be at +5.1V) and also to the emit­ter of Q6. The comparator is used to halt the count when the voltage across the test capacitor reaches 4V. It works as follows: when S1 is in the discharge position, +5.1V is applied to the base of Q6 via a 2.2MΩ resistor. Since the emitter of PNP transistor Q6 is at +4.5V, it is biased off and it removes base drive to Q5 so Q5 is off as well. With Q6 turned off, pin 5 of IC1 is pulled low via a 1MΩ resistor (between Q6’s collector and ground), thus latching the count into IC1 and transferring the latched data to the display. At the same time, with Q5 off, pin 13 goes high and resets IC1’s internal counter to 0000 (resetting the counter has no effect on the latched data). If S1 is now set to the TEST position, the base of Q6 is pulled low via the test capacitor (which initially acts as a short-circuit), thus turning it on. This pulls pin 5 of IC1 high, turning the internal latch off. At the same time, Q5 turns on and a low is applied to pin 13 of IC1 to release the reset on the counter. The test capacitor now charges via constant current source Q9. The rate at which it charges is determined by the selected range resistor and during this time, IC1 is clocked by IC2. When the voltage across the test capacitor reaches about 4V, Q6 turns off again, latching the final count into the display and reset­ting the counter again. The charge on the test capacitor then continues to increase until it reaches a level that’s sufficient to forward-bias Q10, at which point Q10 turns on and clamps the voltage to about 5.1V. The .01µF capacitor at Q6’s collector is included to prev­ ent the 2-transistor comparator from false triggering, while the 0.1µF capacitor at pin 13 of IC1 ensures that there is a short delay between the latching and resetting operations. The 1µF capacitor at Q9’s collector is also necessary to prevent false triggering of the comparator. Discharge indicator When S1 is subsequently switched to the DISCHARGE position, the test Fig.2 (left): the circuit shows a bridge rectifier at the power input so a 12V AC or DC plugpack can be used. Don’t try using a 555 for IC2 in­stead of the 7555 specified because it won’t work as well. September 1999  65 capacitor discharges through the parallel-connected 100Ω resistor. As the capacitor discharges, the voltage across this resistor turns on transistor Q8. This turns on Q7 and lights LED2. When the voltage across the 100Ω resistor drops below 0.6V, Q8 & Q7 turn off and LED2 extinguishes, indicating that the unknown capacitor has been safely discharged. Counter circuit There’s plenty of room left inside the case, since most of the circuitry is on the vertically-mounted PC board. Note how the 7-segment LED displays are mounted – see text. 66  Silicon Chip IC1 is a 74C926 4-digit counter/ display driver and is clocked by IC2, a 7555 CMOS oscillator/timer. The reason that the CMOS version of the 555 was chosen was because it has a cleaner output than the standard 555. IC2 is wired in astable mode and has an output frequency of 105Hz, as set by the RC timing compon­ents on pins 6 & 7. VR1 allows the output frequency to be adjust­ed so that the unit can be calibrated. The output from IC2 clocks pin 12 of IC1 and it does this while the test capacitor charges to 4V. When IC1’s latch enable (LE) pin is subsequently pulled low, the value in the counter is latched and transferred to the segment driver outputs. The digit driver outputs of IC1 are at pins 7, 8, 10 & 11. These multiplex the common-cathode displays via driver transis­ tors Q1-Q4 at a rate determined by IC1’s internal clock. While that’s going on, IC1’s segment-driver outputs (pins 1-4 and 15-17) activate the appropriate display segments. The 47Ω resistors connected in series with IC1’s segment-driver outputs provide current-limiting, while the 390Ω resistor in series with S2b’s wiper limits the current to the selected decimal point. The decimal points are controlled via S2b (part of the range switch). When S2b is in the x1 position, DISP4’s decimal point turns on. Similarly, when S2b is in the x0.1 position, DISP3’s decimal point lights. The other two displays do not need a decimal point. Construction All the components for the Electrolytic Capacitance Meter are assembled on one PC board, with the Range selector switch (S2) and the 7-segment LED displays mounted on the copper side. This allows access to the components and to the frequency preset trimpot (VR1) when the PC board is mounted on the front panel. The first step, as always, is to check the board for un­ drilled holes and etching faults. While these are uncommon, it is far easier to check for them before beginning the assembly, rather than getting half-way though and then finding it necessary to drill a hole. Fig.3 shows the assembly details. The 23 links are best fitted first, although if you use resistor pigtails as jumpers you will naturally have to fit them before the links. The diodes and preset potentiometer come next, followed by the eight PC stakes (these mount at the external wiring positions). Note that two of these PC stakes are inserted directly adjacent to the wiper pads for switch S2. Fig.4 shows the loca­tion of these two stakes. Next, install the transistors, diodes, capacitors (includ­ing the electrolytics) and the two 3-terminal regulators. Lie the 470µF electrolytic capacitor (adjacent to the 7805 5V regulator) flat against the PC board to keep it away from the regulator’s heatsink. Note that the four diodes have their cathode bands all facing in the same direction. Parts List 1 PC board, code 04109991, 195 x 62mm 1 plastic case, 200 x 70 x 160mm, Jaycar HB5912 or equivalent 1 Perspex window, red or smoked grey, 57 x 23mm 1 12V AC or DC plugpack 1 panel-mount socket to suit plugpack 1 2-pole 6-position PC-mounting rotary switch with indexing lug, 2 nuts & toothed washer 1 SPDT toggle switch 1 binding post terminal (red) 1 binding post terminal (black) 1 TO-220 heatsink 2 20-way pin strips, Jaycar PI6743 or equivalent 1 knob to suit rotary switch 9 PC board stakes 1 3mm x 18mm threaded spacer 1 3mm x 6mm CSK head machine screw 1 3mm x 6mm machine screw 2 3mm solder lugs 1 20kΩ horizontal PC-mount trimpot (VR1) Semiconductors 1 74C926 4-digit counter/display driver (IC1) 1 7555 CMOS timer (IC2) 1 7805 5V regulator (REG1) 1 7905 -5V regulator (REG2) Take care to ensure that the electrolytic capacitors are all correctly oriented. Also, be sure to use the 7805 device for REG1 and the 7905 for REG2 (don’t get them mixed up). NOTE: the PC board has been laid out to suit 2N2222 tran­sistors in the TO-18 metal can package but they are also avail­able as TO-92 plastic packs. Unfortunately, the pinouts for the two packages are different (see Fig.2). If you have TO-92 tran­sistors, the trick is to bend the base (centre) lead of each transistor towards the flat on its body. The transistor will then slot straight into the board. Make sure that the transistor lead connections are correct; the circuit won’t work if you get them mixed up. Rotary switch mounting The rotary switch is inserted from the copper side of the PC board. Before 4 2N2222 NPN transistors (Q1-Q4) 2 2N3904 NPN transistors (Q5,Q8) 2 2N3906 PNP transistors (Q6,Q7) 2 2N2905/2N2907 PNP transistors (Q9,Q10) 4 1N4001/4004 1A diodes (D1-D4) 4 7-segment common cathode displays (DISP1-DISP4) 1 5mm green LED (LED1) 1 5mm red LED (LED2) Capacitors 1 470µF 25VW PC electrolytic 3 100µF 16VW PC electrolytic 2 1µF 50V PC electrolytic 2 0.1µF MKT polyester 2 0.1µF monolithic ceramic 2 .01µF MKT polyester Resistors (0.25W, 1%) 1 2.2MΩ 1 2.2kΩ 2 1MΩ 1 2kΩ 1 100kΩ 2 1kΩ 1 47kΩ 1 390Ω 1 39kΩ 1 220Ω 1 22kΩ 1 150Ω 1 15kΩ 2 100Ω 1 10kΩ 7 47Ω 1 4.7kΩ Miscellaneous Tinned copper wire for links, light duty hook-up wire mounting it, solder 25mm lengths of tinned wire to the two common pins. This done, push the switch pins and wires through the board holes until the 12 outside pins are just pro­truding through on the component side. The outside pins can now all be soldered on the copper side of the board. You will need a soldering iron with a small tip for this job. It’s best to solder a couple of diagonally opposite pins first, as this will make it easier to ensure that the switch is square with the board. Once the switch soldering has been completed, connect the wire leads from the common pins to the adjacent PC stakes (see Fig.4). Regulator REG1 must be fitted with a small finned heatsink to keep it cool. You will need to drill another hole in this heatsink, towards one side, so that it can be offset to clear the lid of the September 1999  67 Fig.4: two short wires from PC stakes are used to connect the wipers of the rotary switch. 7-segment LED displays can be made by first drilling a series of small holes around the inside perimet­ er, then knocking out the centre piece and filing the cutout to a smooth finish. The indicator LEDs, toggle switch (S1) and the two test terminals (red to the wiper of S1, black for earth) can now be installed. The terminals used in the prototype had locating lugs which meant that matching holes had to be filed in the front panel after the centre holes were drilled. These locating lugs stop the terminals from rotating when the binding posts are tightened or undone. Once the terminals are mounted on the front panel and the solder lugs fitted, the excess lengths must be cut off so that they don’t foul the PC board. As long as they are shorter than the switch terminals, they will be OK. Final assembly Fig.3(a): take care when installing the 2N2222 transistors because their pin­outs are different depending on whether you have the metal TO-18 type or the plastic TO-92s (see text). Fig.3(b) at right shows the full-size PC board artwork. case (see photo). Smear some thermal grease on the mating surfaces before bolting REG1 to the heatsink. Case preparation The plastic case must have the five lugs at the front of the bottom and the two at the front of the lid removed, to 68  Silicon Chip allow the PC board to sit in position. This is easily done by drilling them out or using a sharp chisel and a small hammer. Next, use the front panel label as a template to mark out and drill the holes for the various items of hardware. The rec­tangular cutout for the The wiring between the front panel and the PC board is straightforward. Use light, flexible leads to allow the panel to fit close to the PC board without jamming or straining the wires but leave these wires long enough to be able to access the PC tracks if you fold down the front panel. The power leads from the back panel can be soldered in later. The anode leads for the two LEDs are wired to their respec­tive stakes on the PC board, while their cathodes are connect­ed together and wired back to the EARTH stake. The black test terminal also connects to the EARTH stake, while the red terminal goes to the wiper of switch S1. The other two The 7-segment LED displays and the rotary switch (S1) are mounted on the copper side of the PC board. Note that we modified the connections to S1’s wipers after this photo was taken (they now connect to PC stakes; see Fig.4). terminals on the switch go to the TEST and DISCH stakes on the board. The indexing lug on the rotary switch should be set to allow for four positions from the fully anticlockwise direction then a nut fitted to hold it in place (ie, three clicks, four positions). The toothed washer should be fitted behind the front panel but don’t attach the front panel just yet. Now plug the four displays into the 20-way pin strips, making sure that all the decimal points are at bottom right. This done, push the 40 pins through the PC board holes, then fit the front panel and secure it at one end with a second nut on the rotary switch. The other end of the front panel is secured to the board using an 18mm threaded spacer and two screws. Countersink the hole on the front panel so that the bolt head will not be visible when you fit the label. Fig.6 shows the dimensions of the window for the LED dis­plays. This can be made from either red or smoked Perspex and should be about 3mm thick. The 2 x 3mm rebate around the outside can be made using an engraving tool (ask your local engraver), a small router or even a flat file. Fit the window from the back and secure it with a couple of drops of super glue. This done, push the pin strips forwards until the displays touch the window, then tack solder the four corner pins. Check that the alignment is satisfactory before soldering the remaining 36 pins. Next, slide the front panel into the plastic case guides and check that the lid fits properly and does not foul the heat­sink. You can then fit the plugpack socket to the rear panel and connect it to the two PC stakes near the diodes. Testing To test the unit, first apply power and check that the power LED lights. If it doesn’t, you’ve probably got the LED wired the wrong way around. Next, use your multimeter to check for about 18V on the cathodes of D2 and D3 (the actual voltage measured will depend on the plugpack used). You should be able to measure the same vol­tage on one end of the 2.2kΩ resistor near REG2 and 5V less (ie, about 13V) at the other end. Pin 18 of IC1 should measure +5V. If all voltages are correct (within ±10%), connect a multimeter set to read a DC current of 50mA (probably Resistor Colour Codes  No.   1   2   1   1   1   1   1   1   1   1   1   2   1   1   1   2   7 Value 2.2MΩ 1MΩ 100kΩ 47kΩ 39kΩ 22kΩ 15kΩ 10kΩ 4.7kΩ 2.2kΩ 2kΩ 1kΩ 390Ω 220Ω 150Ω 100Ω 47Ω 4-Band Code (1%) red red green brown brown black green brown brown black yellow brown yellow violet orange brown orange white orange brown red red orange brown brown green orange brown brown black orange brown yellow violet red brown red red red brown red black red brown brown black red brown orange white brown brown red red brown brown brown green brown brown brown black brown brown yellow violet black brown 5-Band Code (1%) red red black yellow brown brown black black yellow brown brown black black orange brown yellow violet black red brown orange white black red brown red red black red brown brown green black red brown brown black black red brown yellow violet black brown brown red red black brown brown red black black brown brown brown black black brown brown orange white black black brown red red black black brown brown green black black brown brown black black black brown yellow violet black gold brown September 1999  69 SILICON CHIP + 3 TEST ELECTROLYTIC CAPACITANCE METER + x10 x1 x0.1 RANGE x100 -+ + POWER + + + + DISCHARGE DISCHARGING 17 Fig.5: actual size artwork for the front panel label. the 200mA range on most digital meters) across the capacitor terminals and read the value with the Range switch set to the x100 range. This should be around 44mA. Now, stepping anticlockwise, check that the other ranges measure close to 4.4mA, 440µA and 44µA. This is deter­mined by the actual output voltage of REG2 70  Silicon Chip 23 51 57 2 Fig.6: dimensions of the window for the LED displays. This can be made from either red or smoked Perspex and should be about 3mm thick. The 2 x 3mm rebate around the outside can be made using an engraving tool (ask your local engraver), a small router or even a flat file. and the exact value of the selected range resistor. Next, connect a 2200µF or 4700µF capacitor across the termi­nals, set S2 to the x10 range and set the toggle switch (S1) to TEST. The display should start counting up. Wait for five sec­onds, then flick the toggle switch to DISCHARGE. The Discharge LED should come on briefly, then extinguish, while the count should remain fixed on the LED displays. Calibration If you have an electrolytic capacitor with an accurately known value (say 10,000µF or more), connect it across the test terminals and check its value on the x10 range. Now adjust VR1 until the correct value is displayed. This will have to be done on a trial and error basis, with the capacitor re-tested after each adjustment. On our unit, VR1 had to be adjusted until it was almost against the clockwise stop. If you find you need more adjustment, reduce the 39kΩ resistor which goes to pin 7 of IC2 to 33kΩ. If you don’t have a known capacitor, then get several capacitors of the same nominal value (say 10,000µF or more) and test each one. You can then select a unit from the middle of the range and use this as the standard. Note that the overall accuracy is better on the x10 range. Using the meter Normally, a quick check is all that is needed to find a bad capacitor. For example, there will be times when the meter won’t stop counting. As illustrated by Table 1, this indicates that the capacitor has excessive leakage or an internal short. Note that when using the x100 range (44mA), you should let the meter warm up for a couple of minutes to allow the circuit to stabilise. That’s because Q9’s base-emitter junction voltage varies slightly as the transistor warms up. This doesn’t pose a major problem but it can decrease the accuracy until the circuit stabilises. Computer-grade electrolytic capacitors are designed to have a low equivalent series resistance (ESR) while memory backup capacitors have a fairly high ESR. When testing a capacitor that has a high ESR, use the formulas in Table 1 to find the correct capacitance and ESR. The formulas aren’t perfect but they’ll get you close enough. The meter can be allowed to roll over (count to 9999 and continue) if desired. That comes in handy when it’s necessary to test a larger capacitor on a lower range. If you suspect, for instance, that the capacitor being tested has high ESR, testing it on a lower range gives better accuracy because of the lower test current. It is also possible to test a capacitor larger than 1F on the x100 range using that method. Let’s look at a couple of examples: Example 1: while testing a 300,000µF computer-grade capaci­ tor on the x100 range, the meter’s readout displayed 3855. In that case, the actual measured capacitance is 3855 x 100 = 385,500µF. On the x10 range, the meter was allowed to roll over three times, producing a finished count of 9556. The x10-range capacitance would then be 39,556 x 10 = 395,560µF. Both readings are high compared with the capacitor’s specified value and both readings are within 10% of each other. The regulator heatsink must be offset as shown in this close-up photo, to clear the lid of the case on the x100 range. In such cases, you could just accept the reading obtained on the x10 range or use the formula in Table 1 to find the cor­rect capacitance and ESR. Here, the high-range reading is too low and the low-range reading is too high. Table 1 indicates that the capacitor has either leakage or series resistance and should be tested with an ohmmeter. Since we’re checking a memory-backup capacitor, it becomes obvious that the erroneous reading is probably due to series resistance. In such cases, we’d use the lowrange reading of 47,020µF. Using the formulas for ESR and capacitance: (1). ESR = (47,020 - 30,100)/(.011 x 47,020) = 32.7Ω. (2). C = ((11 x 47,020) - 30,100)/10 = 48,712µF. The test terminals, indicator LEDs and test switch are wired back to PC stakes on the board. Referring to Table 1, note that some leakage is indicated but the capacitor is otherwise OK and that the highrange reading is correct (the higher range reading is not the same as the high reading). This means that this capacitor is actually about 385,500µF which is about 29% higher than the manufacturer’s specifications – but well within tolerance. Example 2: while testing a memory backup capacitor (speci­fied as 0.047F) on the x100 range, the meter produced a readout of 30,100µF. On the x 10 range, the meter displayed 47,020µF. The two readings, of course, are not within 10% of each other. It is not difficult to recognise that the reading on the x10 range is closer to the correct value because the series resistance of memory backup capacitors can cause erroneous readings Example 3: while testing a 2200µF capacitor on the x10 range, the display produced a readout of 1610µF. On the x1 range, the reading was 1636µF. Both readings are low and within 10% of each other. Table 1 indicates that the capacitor has low ca­pacitance and to take the high-range reading (1610µF) as the cor­rect value. As it turned out, the capacitor was low by 27% and unusable! Finally, the accuracy of the meter can be increased using several approaches. For example, the CO (carry out) output (pin 14) of the counter/ display driver could be used to clock another counter driving another 7-segment display. This would allow you to see the rollover, instead of counting the number of rollovers. Anoth­er approach would be to divide the 105Hz clock frequency by 10 (when you know there will be roll over), to provide an SC additional x1000 range. September 1999  71