Fullscreen HTML5Med Res (??MB)HTML4

Electrolytic reformer an Got a bunch of old electrolytic capacitors you’d like to use . . . but don’t know if they are any good? Or do you need to re-form the electrolytics in an old valve amplifier or vintage radio set? This Electrolytic Capacitor Reformer and Tester will do the job for you, at any of 11 different standard voltages from 10V to 630V. I n addition, it provides the ability to apply the selected test voltage for any of seven periods ranging from 10 seconds to 60 minutes. Thus you can use it for ‘reforming’ electrolytic capacitors that have developed high leakage and high impedance due to years of inactivity. As well, it can be used to test the leakage of virtually all capacitors at or near their rated voltage. Of course, we have to state that not all old electrolytics can be restored – they can’t. Some will have very high leakage due to contamination of the can seal or breakdown of the electrolyte and some will have just dried out. In those cases, you cannot do anything to resurrect them but in many cases you will be able to restore and re-use Part 1: by JIM ROWE capacitors that have not been used for many years if not decades. Some very old caps (1960s vintage!) we had took several hours to come good while others, made in more recent years, were good within a few minutes. Most high voltage (ie, 250V and above) capacitors should be capable of being reformed to the extent that their leakage current drops to around 3mA or less. The Reformer circuit is designed so that no damage can occur if the capacitor connected to it is short circuit or has very high leakage, or is even connected back-to-front (ie, with reverse polarity). Furthermore, even if the capacitor leakage is very high, the output current is limited so that the maximum dissipation in the capacitor is no more than 2W. This means that some capacitors might get warm while they are being reformed but none will get so hot that they are in danger of swelling up and “letting the smoke out”. That’s a good thing because electrolytic capacitor smoke is particularly foul-smelling! And as any serviceman will tell you, the gunk (electrolyte) inside is particularly nasty if it escapes with the smoke. The Electrolytic Capacitor Reformer and Tester is housed not in a traditional instrument case or box but in a standard plastic storage organiser case WARNING: SHOCK HAZARD! Because the voltage source in this instrument can be set to provide quite high DC voltages (up to 630V) and can also supply significant current (tens of milliamps), it does represent a potential hazard in terms of electric shock. We have taken a great deal of care to ensure that this hazard is virtually zero if the instrument is used in the correct way – ie, with the lid closed and secured – even to the extent of quickly discharging any capacitor when the lid is opened. However, if the safety switching is bypassed, especially when the unit is set to one of the higher test voltages, it is capable of giving you a very nasty ‘bite’ should you become connected across the test clips or a charged high voltage capacitor. There are some situations where such a shock could potentially be lethal. Do NOT bypass the safety features included in this design. We don’t want to lose any SILICON CHIP readers to electrocution. 80  Silicon Chip siliconchip.com.au capacitor nd tester Most hobbyists would have collected many old electros over the years (maybe not as old as some of these!) – but are they any good, or can they be resurrected into life? which, together with a microswitch view the 2-line LCD which shows the of test voltages plus the inbuilt test interlock, provides a safe compartment capacitor voltage, its leakage current timer which allows the test voltage to for the capacitor when it has high volt- and the time elapsed. be applied for as long as 60 minutes. age applied. We also published an electrolytic Another compartment provides The design capacitor reformer designed by Rodhandy storage for the switchmode 12V This Electrolytic Capacitor Re- ney Champness in “Vintage Radio”, plugpack. former and Tester is based on the October 2006 issue. Opening the lid of the case means smaller and simpler unit described However, this new design not only that no voltage is applied to the in the December 2009 issue of SILICON offers higher voltages (the 2006 model capacitor – until the lid is closed – CHIP but with a much bigger selection only went to 400V), it is fully self but perhaps even contained, is a more more importantly, + elegant design and is RLY2 opening the lid very much safer to use. 16x2 LCD MODULE Q5 safely and quickly Commercial capacidischarges the cator leakage current pacitor so there meters/reformers are RA1 is no chance of available but they tend Q4 RA4 a nasty electric to be fairly expensive LED1 CAP UNDER TEST PIC16F88 shock – for you (well over $1000) and  MICRO or anyone else. we don’t believe any AMPLIFIER (IC3) +Vt AN2 A = 1.205 A charged 630V of them incorporate + – (IC2a) SELECTABLE capacitor with its a safety interlock to 100 DC VOLTAGE 1.770M AN5 leads exposed is avoid the possibility SOURCE RB7 RA4 RA7 (11 VOLTAGES, not something to of electric shock. 10V – 630V) be trifled with! With ours, you have RLY1 9.90k 6.8k With the transa choice of eleven dif(S1, IC1, Q1–Q3) S3 S4 S5 parent lid closed ferent standard test you can select the voltages: 10V, 16V, test voltage and 25V, 35V, 50V, 63V, the period of re- Fig.1: block diagram of the Electrolytic Capacitor Reformer and Tester. Not 100V, 250V, 400V, form/testing and shown here is the safety interlock microswitch and discharge resistors. 450V and 630V. These siliconchip.com.au August 2010  81 correspond with the rated voltages of most electrolytic capacitors which have been available for the last 30 years or so. If you have an “oddball” capacitor with a different working voltage, simply select the next voltage down. (In fact, in the vast majority of cases selecting the next voltage up won’t do the capacitor any harm either because most capacitors, especially electrolytics, can stand a short-term higher peak voltage than their working voltage, hence the labelling – eg, 400VW, 500VP). With any of these test voltages applied to a capacitor you can read its leakage current on the 2-line x 16-character backlit LCD screen, with two automatically selected current ranges: 0-200A or 0-20mA. As well, you can also read the voltage which appears across the capacitor at any time in the procedure. Importantly, for reforming capacitors you have the choice of ten test periods: 10 seconds, 30 seconds, 1 82  Silicon Chip minute, 3 minutes, 10 minutes, 30 minutes or 60 minutes. How it works Essentially the Reformer’s operation is quite straightforward, as you can see from the block diagram of Fig.1. This is broadly very similar to the abovementioned design in our December 2009 issue. There are only two functional circuit sections, one being a selectable DC voltage source (on the left) which generates one of 11 different preset test voltages when power is applied to the voltage source (actually a DC-DC converter) via relay RLY2, controlled by the PIC micro (IC3) via transistor Q4. This test voltage is applied to the positive terminal of the capacitor via a protective current limiting resistor and a microswitch, whose purpose we will look at shortly. The second functional circuit section is on the right in Fig.1 and combines a digital meter which is used to measure any direct current passed by the capacitor under test and the voltage appearing across the capacitor. There is also a digital timer which controls the DC test voltage source via Q4 and RLY2. The PIC micro (IC3) forms the ‘brains’ of this section. We use IC3 as a voltmeter to make the current measurement because any current passed by the capacitor flows down to ground via the 100Ω resistor, either alone or with the 9.90kΩ resistor in series. The resistor(s) therefore act as a current shunt and its voltage drop is directly proportional to the current flowing through the capacitor. The meter measures the voltage across the resistor(s) and is arranged to read directly in terms of current. We also use IC3 to measure the voltage across the capacitor for the duration of the leakage test or reforming period. That way, you can keep track of the leakage current and the voltage at any time. For a good capacitor, the voltage across it will rise while the leakage current steadily reduces. The reason for relay RLY1 and the siliconchip.com.au Inside the opened case, showing the main cut-out required. Inset top left is the interlock microswitch which cuts power and bleeds the charge on the capacitor when the lid is opened. And just in case you were wondering – yes, you do have to lay the capacitor down before closing the lid! Note this PC board is an early prototype – several changes have been made to the final version. 9.90kΩ resistor which it effectively switches in series with the 100Ω resistor is that this gives the digital current meter two ranges. This allows it to read leakage currents down to less than 100nA (0.1A), while also coping with charging and/or leakage currents of up to 20mA or thereabouts. Before the micro begins a test by turning on transistor Q4 and relay RLY2 to apply power to the test voltage source, it first turns on transistor Q5 and relay RLY1 to short out the 9.90kΩ resistor, giving the effective current shunt resistance a value of 100Ω, which gives a 0-20mA range for the capacitor’s charging phase. Only when (and if) the measured current level falls below 200A does it switch off Q5 and RLY1, increasing the total shunt resistance to 10kΩ and thus providing a 0-200A range for more accurate measurement of any residual leakage current. So that’s the basic arrangement. Pushbutton switches S3-S5 are used to select the test time period and also siliconchip.com.au to begin a test or end it prematurely. LED1 is used to indicate when RLY2 has applied power to the DC voltage source and therefore when the test voltage is present across the capacitor test terminals. The reason for the resistor in series with the output from the test voltage source is to limit the maximum current that can be drawn from the source in any circumstances. This prevents damage to either the voltage source or the digital metering sections in the event of the capacitor under test having an internal short circuit and also protects the 9.90kΩ shunt resistor and the digital voltmeter section from overload when a capacitor (especially one of high value) is initially charging up to one of the higher test voltages. In the full circuit you’ll find that this series resistance has a total value of 10.4kΩ, which was chosen to limit the maximum voltage which can ever appear at the input of the voltmeter’s input amplifier (IC2a) to just over 6V, even under short circuit conditions and with the highest test voltage of 630V. It is also used to limit the current when the instrument is being used for reforming electrolytics. Circuit description Now let’s have a look at the full circuit of Fig.2. The selectable DC voltage source is again on the left, based around IC1 – an MC34063 DC/ DC conversion controller IC. It used here in a step-up or ‘boost’ configuration in conjunction with driver transistors Q1 and Q2, switching transistor Q3, autotransformer T1 and fast switching diode D4. We vary the circuit’s DC output voltage by varying the ratio of the voltage divider in the converter’s feedback loop, connecting from the cathode of D4 back to IC1’s pin 5 (where the voltage is compared with an internal 1.25V reference). The four series-connected 75kΩ resistors, together with trimpot VR1, form the top arm of the feedback dividAugust 2010  83 12V DC INPUT FROM PLUGPACK POWER D5 1N4004 + REG1 7805 +11.4V K A IN S2 1000 F 25V – RLY2 1000 F 25V +5V OUT GND 220 F K D6 1N4004 A TEST VOLTS ON RLY2: 6V MINI DIL RELAY (JAYCAR SY-4058 OR SIMILAR) A +11.4V  LED1 K 47 D4 UF4007 A 0.27  1k 5W Vcc DrC Ct IC1 MC34063 SwE GND 4 1nF 1 C Q1 BC337 E 100 B 2 B 2.2k C Q3 IRF540N 470nF 630V S Q2 BC327 110  1% 100k 390k 75k 1% 100k 75k 1% 100k 47 F 450V VR1 50k (25T) 100k 1% ZD2 4.7V TPG 16  1% 390k 47 F 450V K TP3 SET VOLTS A 33  1% 220  1% 100k 75k 1% 390k +1.25V 560  1% 75k 1% D G E Cin5 390k 470nF 630V 10T 8 SwC 3 8.2k 5W +HV 80T 7 Ips 6 K T1 560  1% 30  1% 2.4k 1% 3.0k 1% 100 1% 2.0k 1% 4.7k 1% 150 1% 6.8k 1% 1k 1% 22k 1% 63V 50V 100V 250V 400V 450V 630V SC 2010 S1 35V 25V 16V 10V SET TEST VOLTS ELECTROLYTIC CAPACITOR RE-FORMER & LEAKAGE METER er, while the 100kΩ resistor from pin 5 to ground forms the fixed component of the lower arm. These give the voltage source its lowest output voltage of close to 10.5V, which is the converter’s output voltage when selector switch S1 is in the ‘10V’ position. When S1 is switched to any of the other positions additional resistors are connected in parallel with the lower arm of the feedback divider, to increase its division ratio and hence increase the converter’s output voltage. For example, when S1 is in the ‘16V’ position, all of the series-connected resistors in the string between the various positions of S1 are in parallel with the 100kΩ 84  Silicon Chip resistor, increasing the division ratio to increase the converter’s regulated output voltage to 16.25V. The same kind of change occurs in all of the other positions of S1, producing the various preset output voltages shown. Although the test voltages shown are nominal, if you use the specified 1% tolerance resistors for all of the divider resistors they should all be within ±4% of the nominal values, because the 1.25V reference inside the MC34063 is accurate to within 2%. IC1 operates only when the 11.4V supply rail is connected to it via relay RLY2, under the control of micro IC3. The converter circuit then operates and generates the desired test voltage across the two 470nF/630V metallised polyester reservoir capacitors, connected in series, with their voltage-sharing resistors in parallel. At the same time LED1 is illuminated, to warn you that the test voltage will be present at the test terminals. Note that the test voltage present at the top of the feedback divider is not fed directly to the positive test connector, but is first fed through a low-pass RC filter formed by the 8.2kΩ 5W resistor and the series-connected 47F/450V capacitors (which again have voltagesharing resistors in parallel). This filter is to smooth out any ripsiliconchip.com.au +5V 47 F 2.2k 100nF Q4 BC337 Q5 BC327 C E 2.2k B C B 2.2k 5W 12 10k NO MICRO SWITCH ON S6 CASE LID COM 1k 1W 10k S3 S5 TEST TERMINALS D2 + IC4 LM336Z 2.5 TPG ADJ – RB5 2 100 8 1 IC2a 1 RB3 AN2 RB2 RB1 560 K 10k K RB0 A = 1.205 11 4 10 6 56 CLKo EN K K A ZD1,ZD2 A K A B-L K 16 8 7 IC2: LM358 6 15 TP2 (2.0MHz) 6 – + ADJ IRF540N BC327, BC337 B E 7 7805 D GND IN G C IC2b 4 TPG LM336-2.5 LED K A R/W 5 9 Vss 5 3.0k 1N4004, UF4007 3 CONTRAST RS RLY1: 5V/10mA (JAYCAR SY-4030 OR SIMILAR) D1-D3: 1N4148 15 B-L A A A 2 2 Vdd 5 D1 7,8 22 D7 D6 D5 D4 D3 D2 D1 D0 GND 1 14 13 12 11 10 9 8 7 10nF LCD CONTRAST VR3 10k 16 x 2 LCD MODULE RB4 3 ZD1 6.2V 1W VR2 10k IC3 PIC16F88 DECR TIME 1k 6 A K – 1,14 +5V K D3 100nF RLY1 SET 2.49V REFERENCE A TP1 AN5 16 RA7 17 RA0 13 RB7 INCR TIME TEST + +2.49V RA4 10k S4 100nF 4 14 Vdd MCLR RA1 2 Vref+ 6.8k 680k NC 1M 3 270k 820k 1k 1W 18 4.7k E 2.4k 100nF 10k D S GND OUT Fig.2: similar to the block diagram, the circuit is divided into two distinct sections – the high voltage generation on the left side and the reforming/reading/metering section on the right, which itself is under the control of a PIC microcontroller. Don’t depart from this circuit diagram – a lot of effort has gone into making it safe! ple present in the output of the voltage source/converter. The filtered test voltage is then made available at the positive test terminal via a 2.2kΩ 5W series resistor, which together with the 8.2kΩ 5W series resistance of the filter forms the protective current limiting resistance shown in Fig.1. Charged electros can be lethal! Before the test voltage is fed to the capacitor’s positive test connector, it first has to pass through microswitch S6, which is attached to the case so that it switches when the case lid is opened. Normally, (ie with the lid closed) the test voltage is connected but when siliconchip.com.au the lid is opened, the test capacitor’s positive terminal is connected to its negative terminal via two 1kΩ, 1W resistors which will discharge even the largest high voltage capacitors normally encountered in less than a second. Two 1W resistors are used to obtain a sufficiently high voltage rating for the highest value test setting. Of course, very high value lowervoltage capacitors will take much longer to discharge (as much as a few seconds or so) but these are not considered as dangerous to life and limb. It is important for your safety (and more importantly, the safety of others) that the microswitch is not left out nor bypassed or worse, the circuit built into a case which does not have a hinged lid allowing this form of protection. The circuit is perfectly safe as described. Wiring external to the PC board (ie, the high voltage wiring) should be made with 250V AC-rated cable. The easiest place to get such cable is from a surplus flexible mains lead. In fact, you might be lucky enough to find that you have some with red and black insulated wires (which are needed for the test capacitor connections) and newer ones with brown and blue insulated wires (ideal for the connections between PC board and microswitch). We wouldn’t use the green or green/yellow wiring August 2010  85 Parts List –Electrolytic Reformer & Tester 1 Trojan TJW0510 38cm Storage Organiser (from Bunnings) 1 PC board, code 04108101, 210 x 120mm 1 Front panel label, 320 x 120mm, laminated 1 16x2 LCD module with backlighting (Jaycar QP-5516 or Altronics Z-7013) 1 Mini DIL reed relay, SPST with 5V coil 1 Mini DIL relay, SPDT with 6V coil 1 SPDT 250V 10A microswitch (Jaycar SM-1040 or equivalent) 2 19mm square TO-220 heatsinks 1 Ferrite pot core pair, 26mm OD with bobbin to suit 1 25mm long M3 Nylon screw with nut and flat washer 1 1m length of 0.8mm diameter enamelled copper wire 1 10m length of 0.25mm diameter enamelled copper wire 1 Single pole 12-position rotary switch (S1) 1 Instrument knob, 16mm with grub screw fixing 1 SPDT mini toggle switch, panel mtg (S2) 3 SP Momentary pushbutton switches, panel mtg (S3-5) 18 6mm long M3 machine screws, pan head 4 25mm long M3 tapped spacers 4 12mm long M3 tapped Nylon spacers (or two - see text) 3 Nylon flat washers (only for QP-5516 module - see text) 2 M3 nuts 1 7x2 length DIL socket strip, OR 16-way length SIL socket strip (see text) 1 7x2 length DIL pin strip, OR 16-way length SIL pin strip (see text) 1 18-pin IC socket 2 8-pin IC sockets 10 PC board terminal pins, 1mm diameter 2 100mm long Nylon cable ties Semiconductors 1 MC34063 DC/DC converter controller (IC1) 1 LM358 dual op amp (IC2) 1 PIC16F88 microcontroller (IC3, programmed with 0410810A firmware) 1 LM336Z 2.5V reference (IC4) 1 7805 +5V regulator (REG1) 2 BC337 NPN transistor (Q1,Q4) 2 BC327 PNP transistor (Q2,Q5) 1 IRF540N 100V/33A MOSFET (Q3) 1 6.2V zener diode (ZD1) 1 4.7V zener diode (ZD2) 1 5mm red LED (LED1) 3 1N4148 100mA diode (D1,D2,D3) 1 UF4007 ultrafast 1000V/1A diode (D4) 2 1N4004 400V/1A diode (D5,D6) Capacitors 2 1000F 25V RB electrolytic 1 220F 16V RB electrolytic 1 47F 16V RB electrolytic 2 47F 450V RB electrolytic 2 470nF 630V metallised polyester 2 100nF MKT metallised polester 2 100nF multilayer monolithic ceramic 1 10nF MKT metallised polyester 1 1nF disc ceramic Resistors (0.25W 1% metal film unless specified) 1 1MΩ 1 820kΩ 1 680kΩ 4 5 100kΩ 4 75kΩ 1 22kΩ 1 2 6.8kΩ 2 4.7kΩ 2 3.0kΩ 1 3 2.2kΩ 1 2.0kΩ 2 1kΩ 1W 3 1 220Ω 1 150Ω 1 110Ω 2 1 47Ω 1 33Ω 1 30Ω 1 1 0.27Ω 5W 1 50kΩ 25T vertical trimpot (VR1) 2 10kΩ mini horizontal trimpot (VR2,VR3) 86  Silicon Chip 390kΩ 8.2kΩ 5W 2.2kΩ 5W 1kΩ 100Ω 22Ω 1 5 2 3 1 1 270kΩ 10kΩ 2.4kΩ 560Ω 56Ω 16Ω for ANY purpose except earth wiring. Some readers may query the use of 250V-rated cable when the highest voltage check is clearly well above this figure – 630V to be precise. The justification is that Australian/New Zealand standard AS/NZ3017 calls for mains power wiring to be tested at 1000V DC so it follows that the insulation of 250V cable must be able to handle this, at least in the short term. Voltage & current metering Now let us look at the digital metering and control section, which is virtually all of the circuitry below and to the right of the negative test terminal. The 100Ω resistor and paralleled 1MΩ and 10kΩ resistors connected between the negative test terminal and ground correspond to the current shunts shown in Fig.1, with the contacts of reed relay RLY1 used to change the effective shunt resistance for the meter’s two ranges. For the 20mA ‘charging phase’ range RLY1 is energised via Q5 and connects a short circuit across the parallel 1MΩ/10kΩ combination, making the effective shunt resistance 100. But for the more sensitive 200A range RLY1 is turned off, opening its contacts and connecting the parallel 1MΩ/10kΩ resistors in series with the 100Ω resistor to produce an effective shunt resistance of 10kΩ. As you can see the voltage drop across the shunt resistance (as a result of any current passed by the capacitor under test) is passed to the non-inverting input of IC2a, one half of an LM358 dual op amp. And IC2a is configured as a DC amplifier with a voltage gain of 1.205 times, feeding the AN2 analog input of IC3, the PIC16F88 microcontroller which forms the ‘heart’ of the metering/control section. IC3 takes its measurements of the amplified current shunt voltage from IC2a by comparing this voltage with a reference voltage of 2.490V fed into pin 2 of IC3. The reference voltage is derived from the regulated +5V supply line via voltage reference IC4, an LM336Z device which is provided with a voltage trim circuit using D2, D3 and VR2. These are used to set its voltage drop to exactly 2.490V, where it displays a near-zero temperature coefficient. In fact IC3 takes a sequence of 10 measurements at a time and calculates the average of the 10 readings to reduce ‘jitter’ caused by noise transients. siliconchip.com.au This early prototype board has had several component and design changes to that shown in the circuit diagram on p84-85. The final version, along with the component overlay, will be shown next month in the constructional article. It then does mathematical scaling to arrive at the equivalent current readings, which it displays on the 16x2 LCD module. IC3 also monitors the voltage across the capacitor via a voltage divider feeding its AN5 input, pin 12. Timer function As mentioned earlier, pushbutton switches S3-S5 are used to select the test time period to be used and also to begin testing a capacitor. Switch S4 is used to increase the test period time, while S5 is used to decrease it. Then when the user has set S1 for the correct test voltage and has selected the test time period using S4 and S5, testing is begun by pressing S3. IC3 then turns on Q5 and RLY1 to set the metering circuit for the 10mA range, after which it turns on Q4 and RLY2 to feed power to the test voltage converter (and LED1). It also starts a software timer to control how long the test voltage is to be applied. While the test is being carried out, the metering section takes voltage and current readings and displays these on the LCD module, changing down to the 0-200A range automatically if siliconchip.com.au the measurements drop below 0.2mA. Then when the selected test time period ends or the user presses S3 again to end the test prematurely, IC3 switches off the test voltage source. The voltage and current measurements continue however, so you can monitor the current decay as the test voltage drops to zero. Zener diode ZD1 is included in the metering circuit to protect the pin 3 input of IC2a from damage due to accidental application of a negative or high positive voltage to the negative test terminal (from a previously charged capacitor, for example). On the other hand diode D1 is included to protect transistor Q5 from damage due to any back EMF ‘spike’ from the coil of RLY1 when it is de-energised. Trimpot VR3 allows the contrast of the LCD module to be adjusted for optimum visibility. The 22Ω resistor connecting from the +5V supply rail to pin 15 of the LCD module is to provide current for the module’s LED back-lighting. IC1 and the selectable DC voltage source operates directly from the 12V DC supply line (via polarity protection diode D5 and of course power switch S2) while the rest of the circuit operates from a regulated 5V rail which is derived from the battery via REG1, a standard 7805 3-terminal regulator. That’s basically it. The only other point which should perhaps be mentioned is that the PIC16F88 micro (IC3) operates here from its internal RC clock, at a frequency very close to 8MHz. A clock signal of one quarter this frequency (ie, 2MHz) is made available at pin 15 of IC3 and is brought out to test point TP2, to allow you to check that IC3 is operating correctly. Construction Now that we have the design and operation under our belts, we’re ready to move onto the construction. Unfortunately, though, space has beaten us this month, so the complete constructional details, including the mounting of the project within the special case, will be presented next month. In the meantime, the parts list is shown opposite so you can start collecting the bits required. Firmware for the PIC micro will also be on the SILICON CHIP website (siliconchip.com. SC au) next month. August 2010  87