Fullscreen HTML5Med Res (??MB)HTML4

Build this handy test instrument High-Voltage Insulation Tester This high-voltage insulation tester can measure resist­ance from 1-2200 gigaohms. It is battery powered and dis­plays the readout on a 10-step LED bargraph display. By JOHN CLARKE In all cases, when ever mains-operated equipment has been built or repaired, it is wise to test the insulation resistance between active and neutral to earth. This will verify that there is no leakage path to earth which could lead to a serious break­down later on or pose a hazard to the user if the earth connec­tion fails. Of course, a multimeter set to the high ohms range can often detect insulation problems but this is not always a valid test. That’s because a multimeter only produces a very low value test voltage (around 1.5V) and many types of insulation breakdown occur at much higher voltages. Another problem with a normal multimeter is that it will only show overrange for “good” insulation measurements rather than the actual value of the resistance. This is because insula­ tion resistance measurements usually result in readings of thou­sands of megohms (ie, gigaohms – GΩ) rather than the nominal 20MΩ maximum value for a multimeter. The Insulation Tester described here is a self-contained meter which will measure very high values of leakage Fig.1: block diagram of the Insulation Tester. The stepped-up high-voltage is applied to the test terminals via a safety resistor and the resulting voltage across the detector resistance then measured. 30  Silicon Chip resistance for a number of test voltages. It will also test capacitors for leakage. A 10-LED bargraph display is used to indicate the leak­age resistance. A test voltage switch selects between five possi­ ble values, while a 3-position range switch selects either x1, x10 or x100 scale readings. Block diagram Fig.1 shows the block diagram of the Insulation Tester. It is based on a high voltage supply, produced by stepping up from a 9V battery using a converter. This converter can produce either 100V, 250V, 500V, 600V or 1000V DC. Note that, because of the high voltages involved, a safety resistor is included in series with the output. This limits the output current to a minuscule level to (a) protect the circuit when the probes are short circuit; and (b) prevent the user from receiving a nasty electric shock. In operation, the leakage of the insulation under test causes a current to flow between the test terminals. This current is then monitored by the detector resistance between the negative test terminal to ground. The higher the leakage current, the higher the voltage across the detector resistance. This voltage is measured using a special voltmeter circuit which is calibrated to show the resistance on a LED bargraph readout. This is no ordinary meter since it cannot divert any significant current away from the detector resistance or false readings will occur. And the currents involved are extremely minute. A simple calculation will tell us exactly how small the currents flow- Feature s • LED b argraph display • Five test volt ages fr 1000V om 100 • Measu res from 1GΩ to 2200G Ω (2.2TΩ (1000MΩ) ) • Battery operated • Overr ange indicatio n the voltage across the detector resistor without drawing any more than a few picoamps (pA). Circuit details The prototype Insulation Tester was built into a standard plastic case. Be sure to use good-quality test leads, as cheaper types will show significant leakage at high test voltages. ing between the test terminals are. Assuming a 1000V test voltage and a 2000MΩ (2GΩ) resistance between the test terminals, the current flow will be just 1000/(2 x 109) = 500nA. The same resistance at a test voltage of 100V will allow only 50nA to flow. At 2200GΩ (the upper measurement limit of the Insulation Tester), the current flow is a minuscule 45pA (45 x 10-12) when 100V is applied. As a consequence, we need to measure Fig.2 shows the full circuit of the Insulation Tester. It uses six ICs, a transformer, Mosfet Q1 and a number of minor components. The step-up converter uses the two windings of transformer T1 to produce up to 1000VDC. When Mosfet transistor (Q1) is switched on, it charges the primary winding via the 9V supply. When Q1 is switched off, the charge is transferred to the second­ary and delivered to a .0033µF 3kV capacitor via series diodes D1-D3. These three diodes are rated at 500V each and so together provide more than the required 1000V breakdown. Following the .0033µF capacitor, the stepped-up voltage is filtered using a 4.7MΩ resistor and a 470pF capacitor. It is then fed to the positive test terminal via a second 4.7MΩ resis­tor. Note that these two 4.7MΩ resistors provide the current limiting function referred to earlier. Q1 is driven by an oscillator formed by 7555 timer IC2. This operates by successively charging and discharging a .0039µF timing capacitor (on pins 2 & 6) via a 6.8kΩ resistor connected to the output (pin 3). Let’s take a closer look at how this works. When power is first applied, the capacitor is discharged and the pin 3 output is high. The timing capacitor then charges to the threshold voltage at pin 6, at which point pin 3 switches low and the capacitor discharges to the lower threshold voltage at pin 2. Pin 3 then switches high again and so this process is repeated indefi­nitely while ever power is applied. The voltage at the output of the May 1996  31 32  Silicon Chip Fig.2: the circuit uses a step-up converter based on IC1a, IC1b, IC2 and Q1 to produce test voltages ranging from 100-1000V. PARTS LIST 1 PC board, code 04303961, 86 x 133mm 1 adhesive label, 90 x 151mm 1 plastic case with metal lid, 158 x 95 x 52mm 1 SPDT toggle switch (S1) 1 2-pole 6-position rotary PC board mounting switch (S2) 1 2-pole 3-position slider switch plus screws (S3) 1 red banana panel mount socket 1 black banana panel mount socket 1 test lead set (see text) 1 9V battery 1 battery holder and mounting screws 1 EFD20 transformer assembly (Philips 2 x 4312 020 4108 1 cores, 1 x 4322 021 3522 1 former, 2 x 4322 021 3515 1 clips) (T1) 1 150mm length of red hookup wire 1 150mm length of black hookup wire 1 150mm length of yellow hookup wire 1 150mm length of green hookup wire 1 400mm length of mains-rated wire 1 7-metre length of 0.25mm ENCW 1 80mm length of 0.8mm tinned copper wire 1 20mm knob 4 small stick-on rubber feet 13 PC stakes 1 100kΩ horizontal trimpot (VR1) 3 1N4936 fast recovery diodes (D1-D3) Semiconductors 1 LM358 dual op amp (IC1) 1 7555, TLC555, LMC555CN CMOS timer (IC2) 1 LM10CLN op amp and reference (IC3) 2 CA3140E Mosfet input op amps (IC4,IC5) 1 LM3915 log bargraph driver (IC6) 1 IRF820, BUZ74 or BUK455500A 500V N-channel Mosfet (Q1) 1 BC557 PNP transistor (Q2) 1 10-LED bargraph (LED1-LED10) 1 3mm red LED (LED11) Resistors (0.25W 1%) 1 10MΩ 1 36kΩ 1 8.2MΩ 1 22kΩ 1 4.7MΩ 1 20kΩ 4 4.7MΩ Philips VR37 1 1.2MΩ 1 11kΩ 1 820kΩ 3 10kΩ 1 470kΩ 1 9.1kΩ 1 390kΩ 1 8.2kΩ 1 180kΩ 1 6.8kΩ 2 120kΩ 1 1.8kΩ 3 100kΩ 1 1.2kΩ 2 82kΩ 1 1kΩ 1 56kΩ 1 100Ω 1 47kΩ 1 82Ω 1 43kΩ converter is controlled by monitoring the voltage across a resistor selected by S2b and feeding this to an error amplifier. In greater detail, S2b se­lects one of five range-setting resistors. This, in conjunction with two associated 4.7MΩ resistors, forms a voltage divider across the converter output. The voltage divider output is applied to error amplifier IC1a via a 10kΩ resistor. This stage is cascaded with IC1b for high gain. IC1b’s output, in turn, drives the threshold pin (pin 5) of IC2. If the output voltage goes too high, IC1b pulls pin 5 of IC2 slightly lower so that the pulse width duty cycle to Q1 is reduced. This in turn lowers the output voltage. Conversely, if the output voltage is too low, IC1b pulls pin 5 of IC2 higher. This then increases the duty cycle of the drive to Q1 and so the output voltage also increases. Basically, IC1a compares the voltage divider output with a fixed reference voltage applied to its pin 3. This refer- ence voltage is provided by IC3a and IC3b. IC3a is part of an LM10 dual op amp which includes a 200mV fixed reference at its non-inverting input (pin 3). It amplifies this reference by a factor of 10 to provide 2V at its pin 1 output. IC3b is connected as a unity gain buffer and provides a low impedance output for the 2V reference. Note that the reference voltage is taken from the inverting input at pin 2, while the output at pin 6 drives pin 2 via a 100Ω resistor. This resistor isolates IC3b’s output from the associated 100µF decoupling capacitor. Capacitors 4 100µF 16VW PC electrolytic 1 0.33µF MKT polyester 2 0.18µF MKT polyester 1 0.1µF MKT polyester 1 .0082µF MKT polyester 1 .0039µF MKT polyester 1 .0033µF 3kV ceramic 1 470pF 3kV ceramic IC4, a CA3140E FET-input op amp, functions as a buffer stage and is used to monitor the voltage across the detector resistor. This op amp offers a very high input impedance of 1TΩ (1000GΩ) and a nominal 2pA input current at the 9V supply. Howev­er, this input impedance and current is only valid if there is no leakage on the PC board. To prevent board leakage we have added a guard track around the input which is at the same voltage as pin 3. This effectively prevents current flow from the negative test terminal to other parts of the circuit. Specifications Test voltages ................................................100, 250, 500, 600 & 1000V Test voltage accuracy ...................................<5% Charging impedance ....................................9.4MΩ Current drain 50mA ......................................<at>1000V out May 1996  33 the test terminals are shorted, even at the 1000V setting. Switch S3 selects one of three possible resistance values for the separate ranges. Position 1 selects a 128.2kΩ resistance (120kΩ + 8.2kΩ), position 2 selects 1.282MΩ and position 3 se­ lects 12.82kΩ. These are unusual values but are necessary to correspond to a 1.28V full scale reading for the LED bargraph driver (IC6). Because of the high impedance at the negative test termi­nal, the input is prone to hum pickup and so it is filtered using a 0.18µF capacitor. Note that the earthy side of this capacitor is connected to the output of IC5 rather than to ground or to the 2V rail. This arrangement ensures that there is no DC voltage across the capacitor, thus giving the filter a fast response time. Conversely, if DC voltage had been allowed to appear across the capacitor, the circuit would have taken a considerable time to settle each time a measurement was taken. Buffer stage IC5 (another CA3140) monitors IC4’s pin 2 voltage via a 10MΩ resistor and a 0.33µF capacitor. The output from IC5 at pin 6 is thus a replica of the signal on pin 3 of IC4. It is connected to the earthy side of the 0.18µF filter capacitor, as mentioned above. Note that IC5 has been given a slow response by connecting a .0082µF compensation cap­ acitor between pins 1 and 8. IC4’s output is applied (via a 1kΩ resistor) to the pin 5 signal input of IC6. This is a log­ arithmic LED bargraph display driver which switches on LEDs 1-10 in the dot mode. Each step in the bargraph is 3dB (1.41) apart, giving a total 30dB range. Note that the lower threshold (RLO – pin 4) of IC6 sits at the +2V reference level provided by IC3b. This means that the upper threshold (RHI – pin 6) sits at 3.28V, since this pin sits 1.28V above RLO as set by an internal regulator. This 1.28V difference between RLO and RHI sets the maximum display sensitivi­ty. The 1.2kΩ resistor on pin Fig.3: install the parts on the PC board exactly as shown on this wiring diagram. Check that the LED bargraph display is correctly oriented and be sure to use Philips VR37 resistors where specified. Trimpot VR1 (between pins 1 & 5) is used to adjust the offset voltage at the output (pin 6) of IC4, while S2a sets the gain. This varies from x10 in the 1000V position up to x100 for the 100V setting. These gain adjustments 34  Silicon Chip are necessary to compen­sate for the voltage change that occurs across the detector resistance each time the test voltage is changed. The 100kΩ input resistor at pin 3 of IC4 protects the input from damage if Bend Q1 over as shown in this photograph, so that it doesn’t foul the front panel. The LED bargraph is installed so that its top surface is 19mm above the PC board. 7 sets the LED brightness. Q2 and LED11 provide the over­ range indication. If any of the LEDs is on, Q2 is biased on due to the current flowing through the 82Ω resistor. As a result, LED11 is off since Q2 effectively shorts it out. Conversely, if all the LEDs are out (which equates to a very high resistance), Q2 is biased off and so LED11 now lights to indicate an overrange. Power for the circuit is derived from a 9V battery via switch S1. There are several 100µF capacitors across the supply and these are used to decouple the 9V rail. Construction Most of the circuitry for the Insulation Tester is mounted on a PC board Fig.4: the primary of the transformer is wound first & covered with several layers of insulating tape before the secondary is installed. coded 04303961 and measuring 86 x 133mm. Fig.3 shows the parts layout on the PC board. Begin the assembly by installing PC stakes at the external wiring points (11 in all). These are located at the (+) and (-) battery wiring points, the wiring points for S3 (1-4), the three wiring terminals for switch S1, and at the (+) and (-) terminal points. Once the PC stakes are in, install the resistors, diodes and ICs. Don’t just rely on the resistor colour codes – check each resistor using a digital multi­meter, as some colours can be difficult to read. Take care to ensure that the semiconductors are correctly oriented. The capacitors can go in next, followed by the transistors and the trimpot (VR1). Note that Q1 must be mounted at full lead length so that it can be bent horizontally over the adjacent .0039µF capacitor. This is necessary to allow clearance for the lid of the case, when it is later installed. LEDs 1-10 (the bargraph) and LED11 can now be installed. Be sure to install the bargraph with its anode (A) adjacent to the 82Ω resistor. It should be mounted so that the top surface of the display is 19mm above the board, May 1996  35 The completed PC board mounts on the back of the lid and is secured using the nuts for switches S1 and S2. assembled PC board. This is fitted with a self-adhesive front-panel label measuring 90 x 151mm. Begin the final assembly by affixing the front panel label to the lid, then drill out and file the holes for the LED dis­play, LED11, switches S1, S2 & S3, and the two terminals in the end of the case. Holes will also have to be drilled in the base of the case for the 9V battery holder. This done, the front panel can be test fitted to the PC board. Check that everything lines up correctly and make any adjustments as necessary. You may need to adjust the height of the LED bargraph or LED11, for example. When everything is correct, set switch S2 fully anticlock­wise and move its locking tab (found under the star washer) to position 5. This ensures that S2 functions as a 5-position switch only. The external wiring can now be installed. Use light-duty hookup wire for the connections to S3 and the battery holder and mains-rated cable for the connections to the test terminals. Important: the leads to the test terminals must be kept well apart, as any leakage between them at the high test voltages used will affect readings. Testing so that it will later fit into a matching slot cut into the lid of the case. The top of LED11 should be 20mm above the board surface. Switch S1 is soldered directly to its PC stakes but with its pins touching the top of the PC board. You may need to cut the PC stakes to length to do this. S2 is installed directly on the PC board after first cutting the shaft to a length suitable for the knob. Transformer winding Transformer T1 is wound with 0.25mm enamelled copper wire – see Fig.4. The primary is wound first, as follows: (1) remove the insulation from one end of the wire using a hot soldering iron and terminate this end 36  Silicon Chip on pin 7; (2) wind on 20 turns sideby-side in the direction shown and terminate the end on pin 3; (4) wrap a layer of insulating tape around this winding. The secondary is wound on in similar fashion, starting at pin 4. Note that you will need to wind on the 140 turns in several layers. Use a layer of insulating tape between each layer and terminate the free end on pin 5. The transformer is now assembled by sliding the cores into each side and then securing them with the clips. This done, insert the transformer into the PC board, making sure that it is oriented correctly, and solder the pins. A standard plastic case measuring 158 x 95 x 52mm is used to house the To test the unit, apply power and check that, initially, one of the LEDs in the bargraph display lights. Assuming that the test terminals are open circuit, the bargraph reading should then slowly increase until the over­ range LED comes on. If this doesn’t happen, check that the LEDs are oriented correctly. Now check the circuit voltages with a multimeter. There should be about 9V between pins 4 & 8 of IC1; between pins 1 & 8 of IC2; between pins 7 & 4 of IC3, IC4 and IC5; and between pins 2 & 3 of IC6. There should also be a reading of 2V at TP2. If everything checks out so far, select the 1000V (or high­er) range on your multimeter and connect the positive meter lead to the cathode (striped end) of D3. Now check for the correct test voltages, as selected by S2. Note that if the output voltage is measured directly at the test terminals, the meter will show only about half the correct value because it loads the 9.4MΩ output impedance. Next, set your multimeter to read DCmV and connect it between TP1 <1 2 4 8 16 OVER RANGE + 1.4 2.8 5.6 11 22 GΩ RANGE + x1 x100 x10 ON 250V 500V 100V 600V 1000V + TEST VOLTAGE Figs.5 & 6: here are the full size artworks for the PC board and the front panel. Check your board carefully against the above pattern before mounting any of the parts, as any problems will be more difficult to locate later on. and TP2. This done, set the range switch to the x1 position and slowly adjust VR1 until you obtain a 0mV (or close to it as possible) reading. Note: nothing should be plugged into the test terminals during this procedure. Once all the adjustments have been completed, fit the front panel to the board assembly and secure it by fitting the nuts to switches S1 and S2. The unit can then be installed in the case and the knob fitted to S2 to complete the assembly. Test leads It is important to note that maximum resistance readings cannot be obtained from this instrument if the test leads touch each other or are twisted together, or if a standard test lead set is used. For measurements up to and beyond 220GΩ, we recommend high quality INSULATION TESTER test leads such as those from the Fluke range. DSE Cat. Q1913 test leads (or an equivalent type) are also capable of meaningful results above 220GΩ, provided rubber gloves are worn and the leads are not touching a common surface. Alternatively, you may be able to improve on a standard test lead set by WARNING! Take care with fully charged capacitors since they can provide a nasty electric shock. Always discharge the capaci­ tor after testing it by switching off the Insulation Tester with the probes connected. A 1µF capacitor will take about 10 seconds to discharge using this technique, while larger values will take proportionally longer. insulating the probes with heatshrink tubing. In most cases the protective shroud on the test lead banana plugs will have to be cut away to allow them to be inserted into the banana sockets. You can now check the unit by connecting the test leads across the terminals of an unwired switch. The leakage is then determined by first selecting the x1 range and then switching to the next range if necessary. If the display indicates 1GΩ on the x1 range, then the switch under test is either faulty or its contacts are closed. Note that the unit will display a reading of 1GΩ even if the actual resistance is much lower than this. Finally, when checking capacitors for leakage, be sure to select the correct test voltage. It is then necessary to wait until the capacitor fully charges before SC taking the reading. May 1996  37