Fullscreen HTML5Med Res (??MB)HTML4

By JIM ROWE and NICHOLAS VINEN Isolating High Voltage Probe for Oscilloscopes Here’s a low-cost project which will allow you to use your oscilloscope to observe and measure AC mains and other high voltage waveforms safely. It has three switchable input voltage ranges, wide bandwidth and high voltage isolation between input and output. O bserving and measuring waveforms on the AC mains and in other high voltage circuitry is quite dangerous using a standard oscilloscope or with the usual passive probes. And by “dangerous” we mean not only risking a possibly lethal electric shock to yourself, but also risking serious damage to your scope. The danger arises mainly because the “earthy” side of all scope inputs is connected to the scope’s internal frame, which is normally itself earthed via the mains cable. And it needs to be earthed in this way, both for correct operation and for the safety of the operator – you. (An unearthed or “floating” scope is an ac26  Silicon Chip cident/disaster waiting to happen, so never be tempted!) So the earthy side of all scope inputs is connected back to mains earth, which clearly poses a problem when you want to make measurements in circuits where everything is operating at a high or significant voltage with respect to earth. After all, where do you attach the “earth clip” of the scope probe? For example, in a circuit connected directly to the 230VAC mains, you can’t connect the earth clip to the Active line because this will at least blow one or more fuses and may even start a fire which destroys either the scope or various components in the circuit you want to make measurements in. On the other hand you can’t clip it to the Neutral line either, because this is often itself floating at a significant voltage with respect to earth. Another problem arises because the input attenuator on each channel of most scopes can only be switched to a maximum setting of 5V/division, which corresponds to 50V/division when a 10:1 divider probe is being used. Because there are usually only 10 vertical divisions on the display, this means that only waveforms of up to 500V p-p (peak-to-peak) can be displayed in their entirety. Since the peak-to-peak amplitude of a 230VAC mains waveform is around 650V, this means that it simply can’t be siliconchip.com.au The differential probe connects to the circuit being tested using a pair of standard multimeter probes, alligator clip leads or similar. The output signal is optically isolated and connects to the oscilloscope (or other test instrument) via a BNC lead. Three different attenuation factors are available; 10:1, 100:1 or 500:1, to suit the voltages being measured. The higher attenuation settings offer the best bandwidth, up to 1MHz. displayed or measured properly. it is not possible to achieve meaningful $385 and they rapidly move up into Things are even worse when it comes measurements. the four-digit range. to making measurements in circuits Even if the scope does offer a difWe estimate that you should be able connected to the 3-phase 400VAC ferential mode, the resulting waveform to build this new design for less than mains (415VAC with 240VAC mains). may not be a true portrayal because the $100. It’s true that 100:1 passive probes scope’s common mode rejection may are available and these can be used to not be adequate when measuring high The new probe extend a scope’s upper voltage limit to voltage circuits. Unlike other scope probes this one a nominal 500V/division or 5kV p-p. The best way of solving all of these is not meant to be held in the hand but But this type of probe does nothing to problems is to use a special probe with sits on the bench – with its insulated solve the main problem: where do you full high voltage isolation built in, like input leads running to the circuit under connect the probe’s earth clip? the one we’re describing in this article. test and its output connected to one With most modern scopes having at By the way, we know that this type input channel of the scope via a BNCleast two input channels, there is usu- of probe is available commercially. But to-BNC cable. ally only one way around this problem. the cheapest we could find was about It’s housed in a small ABS instruThat’s to use two ment box measur100:1 divider probes, ing 150mm long, one for each input 80mm wide and channel, and re- An isolating high voltage probe for oscilloscopes, providing three voltage division ranges. 30mm high. ÷500 (optionally, ÷200), ÷100, ÷10 move the earth lead Division ranges: All of the probe’s 2.0M|| ~10pF and clip from both Input resistance: circuitry, including probes. the two 9V alkaline Linearity: ±0.05% Then the two Bandwidth (see Fig.3): batteries it uses for 10:1 range: DC to 500kHz (±0.5dB) channels are used in power, is housed 100:1 range: DC to 1MHz (±1dB) differential mode, to inside the box. 500:1 range: DC to 900kHz (+0.2,-1dB) display and measure The input leads Residual noise: typically 1.4mV RMS, 2.5mV peak-to-peak the voltage differplug into insulated Input-output isolation resistance: >10G (500V) ence between the “banana” sockets two tips. But unless Maximum working isolation voltage: 1.4kV peak (1kV RMS) at one end of the 2.1kV peak (60 seconds) the scope provides a Isolation test voltage: box, while the BNC 8kV peak (10 seconds) differential (subtrac- Maximum transient I/O voltage: output connector 2 x 9V alkaline batteries tion) mode (Ch1-Ch2 Power supply: emerges from the 6.0mA from battery 1, 1.0mA from battery 2 or Ch2-Ch1) display, Typical operating current drain: other end. Specifications siliconchip.com.au January 2015  27 On the top of the output photodiode. Vcc1 Vcc2 box are the two main The close matchcontrols: a small ing of the two phoLINEAR ANALOG OPTOCOUPLER rocker switch to turn todiodes means that the probe’s power on when the LED is V IC1 l and off and a rotary passing a current IF switch used to select and emitting radiaI one of three volttion to both photoI I IC2 age division ranges: diodes, the current V ÷500, ÷100 and ÷10. IPD1 passed by the R2 R1 The important feedback photopoint to grasp is diode will have a OUTPUT CIRCUIT INPUT CIRCUIT that inside the box, value very close GROUND GROUND there’s a high voltage to that of the curFig.1: the simplified probe circuit. Op amp IC1 drives an LED in the opto“galvanic isolation coupler with feedback from one of the photodiodes. IC2 generates the output rent IPD2 passed by barrier” between the signal from an identical, isolated photodiode. Note that I the isolated outPD1 ≈ IPD2. input and output put photodiode. circuitry. By passing current This allows the input leads to be optocoupler), the other is located back IPD1 through resistor R1 to produce connected to circuits operating at many on the same side as the LED itself. a voltage proportional to the LED hundreds of volts above (or below) This allows the second photodi- current IF, we can use the resulting earth, despite the fact that the probe’s ode to be used to provide linearising voltage to provide input amplifier IC1 output is directly connected to the feedback, as a “proxy” for the isolated with negative feedback. This linearises earthed input of a scope – and without causing any distress or damage. 10pF In fact the isolation barrier inside + 1.5kV CON1 the probe is able to withstand a peak K 62k 620k 620k 560k D1 “working” voltage of 1414V, or 2100V 100nF 1N5711 for up to one minute (60 second), or 62k 500V 500V 500V A Q1 as high as 8000V peak for transients 0.5W 0.5W 0.5W INPUT 8 BC549 3 ÷10 RANGE 10pF of less than 10 seconds in duration. 100pF 56k B 1 S1a IC1a 150V ÷100 2 And if you’re curious about the isola500V tion resistance between the inputs and INPUT 4.7pF 330W SOCKETS 220pF 1nF ÷500 16k the output, this is more than 10G(10 K (÷200) (1nF) (10k) IC1: LM6132BIN Gigaohms or 10,000M). D2 F PD1 FEEDBACK PIN PHOTODIODE I S O L AT I O N B A R R I E R AlGaAs LED IN ISOLATED PIN PHOTODIODE PD2 OUT 28  Silicon Chip E 1N5711 How it works The probe achieves this impressive performance because of a very special component: a high linearity analog optocoupler. Understanding what this is and how it works is the key to understanding how the probe works as a whole, as we’ll see shortly. For the present, though, refer to Fig.1 which shows a basic linear analog isolation amplifier based on one of these devices. The linear analog optocoupler is like a conventional digital optocoupler except that it has two PIN photodiodes sensing the infrared (IR) radiation emitted by the high performance AlGaAs LED. The two photodiodes are very closely matched in terms of their optical sensitivity and linearity. The only difference between these ”identical twin” photodiodes is that while one of them is located on the far side of the device’s internal voltaic isolation barrier (like the output photodiode or transistor in a conventional C 2.0k (1nF) 4.7nF (1nF) (10k) 2.0k (link) CON2 A USE VALUES IN BLUE FOR 200:1 MAXIMUM DIVISION RATIO OMIT EXTRA CAPACITOR FOR 500:1 INPUT AMPLIFIER/BUFFER 200k IC1, IC2 BC549 B E C = INPUT SIDE GROUND 4 8 1 D1-D4 A K ON/OFF S2a MAXIMUM INPUT VOLTAGES (DC + AC, CON1 TO CON2) FOR THE THREE INPUT RANGES RANGE MAXIMUM VOLTS ÷10 80Vp-p (28V RMS) ÷100 800Vp-p (280V RMS) ÷200 ±800V peak (560V RMS) ÷500 ±1414V peak (1000V RMS)* *SET BY THE WORKING ISOLATION VOLTAGE RATINGS OF OPTO1 & S2 SC Ó2015 10k INPUT HALF-SUPPLY BUFFER 5 6 9V BATTERY1 100mF 16V D3 1N4004 6 4 100nF 7 IC1b 10k 150W 100mF 16V ISOLATING HIGH VOLTAGE PROBE FOR SCOPES siliconchip.com.au the operation of the input circuitry in converting input voltage VIN into LED current IF and hence the IR radiation passing over the isolation barrier. Since the output photodiode’s current IPD2 is virtually the same as IPD1, we are then able to use resistor R2 to convert this current back into a voltage VOUT which is also directly proportional to VIN. (IC2 is then used to buffer VOUT, to ensure that any load connected to the output does not upset this linearity.) In fact the resulting linear relationship between VOUT and VIN turns out to be very close to the ratio of R2 to R1, multiplied by the optocoupler’s “transfer gain” K3 (where K3 = IPD2/ IPD1). So +3 500:1 10:1 -1 0 -2 -3 90 -4 -5 180 -6 -7 270 -8 -9 50 100 200 500 1k (ISOLATION BARRIER) 100nF 56k FEEDBACK PIN PHOTODIODE 6 4 l 2 1 IR LED TO SCOPE INPUT 5 8 3 ISOLATED PIN PHOTODIODE 1nF (ISOLATION BARRIER) IC2: TLV2372IP V1 LED1 BLUE OUTPUT BUFFER A l V2 V2+ OUTPUT HALF-SUPPLY BUFFER (ISOLATION BARRIER) 10k 5 OFFSET VR2 ADJUST 2k 9V BATTERY2 V1 6 10k 100mF 16V IC3b IC 2b 4 100nF 100mF 16V D4 1N4004 100mF 16V 7 150W V2 Fig.2: the complete probe circuit. The voltage being monitored is attenuated by a resistor/capacitor ladder and the selected tap connects to input pin 3 of IC1 via rotary switch S1. IC1b and IC2b provide half-supply rails to allow signals with bidirectional voltage swings to be probed. siliconchip.com.au 360 50k 100k 200k 500k 1M A = VOUT / VIN = (R2/R1) It also turns out that we can compensate for any small deviation of the optocoupler’s K3 away from unity, simply by “tweaking” the value of R2. So the overall gain of the isolation amplifier can be adjusted to be exactly unity, or whatever other figure we want it to be. So we achieve linear analog voltage gain while at the same time passing over a high voltage isolation barrier. Performance = OUTPUT SIDE GROUND K S2b V1+ 100mF 16V VR1 50k GAIN CALIBRATE CON3 100W 1 IC2a 2 180k 5k 10k 20k linear analog optocouplers have a transfer gain K3 of very close to unity (1.0); within a few percent. So the overall gain of the basic linear isolation amplifier of Fig.1 simplifies down to: V2+ OPTO1 IC2 HCNR201 HCNR201 LINEAR OPTOISOLATOR 2k Frequency (Hz) Fig.3: frequency response of the probe for each attenuation setting. The response is flattest at 500:1 but there is slightly more bandwidth at 100:1. The output/ input signal phase shift for each setting is shown dashed, using the right y-axis. Because of the close matching between their twin photodiodes, most 3 0 -10 10 20 VOUT/VIN = K3.(R2/R1) V1+ 100:1 +1 Phase Shift (Degrees) Output/Input Relative Amplitude (db) +2 We tested our prototype by measuring signals under a number of different circumstances. The ‘litmus test’ was connecting the probe across the motor of a drill plugged into our 230V/10A Speed Controller For Universal Motors (February-March 2014). The result is shown in Scope1. This is gratifying as it gives a clear picture of the voltage across the load, despite the fact that it’s floating at mains potential and with the fast rise/fall times displayed correctly. In fact, this result is almost identical to what we get with a commercial differential probe. With its ~1MHz bandwidth, our probe can be used to view signals with a higher switching frequency than this. For example, it could be used to view January 2015  29 the probe to unity. At the probe’s front-end circuitry, the 200k resistor connected between pin 2 of IC1a and the input circuit’s negative rail is the equivalent of feedback resistor R1 in Fig.1. As you can see the anode of OPTO1’s feedback photodiode (pin 4) also connects to the 200kresistor, as in Fig.1. Note that the value of the 330 current-limiting resistor is important since its ratio with the 200kresistor sets the current gain of the optocoupler and this affects the open-loop bandwidth of the surrounding circuit (ie, including IC1a). Increasing this resistor value reduces output overshoot but also reduces overall bandwidth. The 4.7pF capacitor also has an effect on bandwidth (in combination with the 330resistor) and is required for the circuit to be stable, due to the phase shift inherent in the DC feedback path via the opto-coupler. Scope1: the voltage across a drill powered by our 230V/10A Speed Controller for Universal Motors, showing a rectified mains waveform chopped at about 1kHz. The spikes are generated by the circuit; they are not measurement artefacts. a floating Mosfet gate drive. We did try it out connected across the output of our Induction Motor Speed Controller (April/May 2012) which has a much higher switching frequency, 36kHz. While we were able to get a reasonable picture of the output waveform (Scope3 shows it “zoomed out”), the bandwidth of our probe is a little too low to show the very short pulses as a square wave. The voltage rise and fall times are simply too fast. The output photodiode of OPTO1 is connected to the non-inverting input (pin 3) of output amplifier IC2a, in exactly the same way as in Fig.1. Trimpot VR1 with its series 180kresistor takes the place of R2 in Fig.1, with VR1 allowing the exact value of R2 to be adjusted to set the overall gain of Input voltage divider The non-inverting input of IC1a (pin 3) is connected to input connectors CON1 and CON2 via a switched voltage divider, to provide the probe’s three division ranges. The switching is done by S1a, one pole of a 4-pole, 3-position rotary switch (the other poles are unused). The input divider is arranged so that it provides a fixed input resistance of The full probe circuit Now refer to the full circuit of Fig.2. The specific linear analog optocoupler device we’re using is the HCNR201, made by US firm Avago Technologies. This has very impressive features: • • • • • • • Low non-linearity: <0.01% Transfer gain: 1.00 ±5% Wide bandwidth: >1MHz Isolation: UL 5000V RMS for one minute Maximum working voltage: 1414V peak I/O test voltage: 2121V peak for 60s I/O transient over-voltage: 8000V for 10s The IR LED of optocoupler OPTO1 is driven by op amp IC1a via transistor Q1. The transistor is used as an emitter follower to provide the required current drive for the optocoupler’s LED, since IC1 is a low power device with low current drive capability. 30  Silicon Chip Scope2: a 1kHz scope compensation square wave as measured using the differential probe on its 10:1 setting. There are brief overshoot spikes at each edge but otherwise the shape is square with no ringing or distortion. siliconchip.com.au + + + + + OPTO1 + + siliconchip.com.au HCNR201 + 5711 /500 IC2 TLE2022 5711 4004 10k 10k + + C 2014 /100 62k 150W 560k /10 (500V 0.5W) 2Mon all three ranges. 9V BATTERY A series of capacitors (FOR CIRCUITRY ON have been connected in (500V 0.5W) OUTPUT SIDE OF 620k parallel with the divider ISOLATION BARRIER) resistors. These are re620k VR1 50k S1 100mF 100mF 100mF 100mF ADJUST GAIN RANGE quired for a number of +IN – 10pF 500V 180k reasons. BATTERY 2 OUTPUT TO 10pF OUTPUT 150W SCOPE 1.5kV D3 Firstly, they swamp the 100nF 100nF D1 100pF 100mF input capacitance of IC1a S2 150V CON3 (exacerbated by the capacIC1 OFF/ON LM6132 62k itance of D1 & D2), which 100nF 56k would otherwise form a 100W 4.7pF 16k D2 4.7nF –IN 100nF 100mF 1nF 1nF low-pass RC filter with the 200k A K 10k 220pF 2.0k 2.0k NOTE: NOTE:AAPIECE PIECE Q1 resistive divider network, 330W LED1 BC549 OF OF0.8mm 0.8mmTHICK THICK 10k VR2 2k seriously limiting the 9V BATTERY PRESSBOARD PRESSBOARD ADJUST OFFSET SHEET SHEET100 100xx23mm 23mm 56k available bandwidth. (FOR CIRCUITRY ON (CUT (CUT&&BENT BENTAS ASIN INFIG.8) FIG.7)ISIS 4004 INPUT SIDE OF They also keep the AC USED USEDTO TOPROVIDE PROVIDEEXTRA EXTRA D4 STRAIN ISOLATION BARRIER) – impedance “seen” by IC1a ISOLATION ISOLATIONBETWEEN BETWEENINPUT INPUT BATTERY 1 RELIEF AND ANDOUTPUT OUTPUTCIRCUITRY CIRCUITRY low, minimising noise and RF/hum pick-up. An extra 10pF capacitor placed across the top 620kresistor in the divider provides some extra compensation to cancel out the input capacitance of IC1a. Regarding the voltage ratings of these components, 90% of the voltage applied across inputs CON1 & CON2 appears across the top three resistors and parallel capacitor. Given the 1414V peak rating of the device, the resistors must therefore be able to handle at least 500V and the 10pF capacitor, 1.5kV. Similarly, the Fig.4 (top): the component overlay, which matches the near-same-size photo of the early proto100pF capacitor sees 9% type PCB (above). Note that the PCB is double-sided – make sure you solder the components to of the total voltage and the correct side! S2 is not yet soldered in place in the photo but is shown in situ above. thus must be rated for at So each section operates from its least 150V. the opto-coupler just below 1MHz (ie, Diodes D1 and D2 provide over- its roll-off point). This gives a flatter own 9V alkaline battery, with the input section running from battery 1 and the voltage protection for IC1a, ensuring frequency response (Fig.3). that input pin 3 cannot swing higher Note that we’ve also shown some output section from battery 2. We are using op amps 1C1b and IC2b than 0.4V above the input circuit’s alternative divider component values positive supply rail (V1+) or lower in the circuit. If used, these change the as buffers to give each supply its own than 0.4V below its negative rail (V1-). ÷500 range to ÷200. This results in a half-supply “reference ground”. The This prevents IC1 from damage should better signal-to-noise ratio but with buffers are very similar, in each case you accidentally connect the probe a more limited input voltage range using a resistive divider to establish a battery “centre tap”, with the ICs coninputs to high voltages when switch before saturation (see table in Fig.2). S1 is switched to the low voltage (÷10) Note that the resulting 800V peak nected as voltage followers to provide range. rating is sufficient for working with the necessary current capability. (The 150 resistors and 100µF caThe 100resistor at IC2a’s output even 3-phase mains. pacitors are to ensure that the voltage isolates this buffer from any cable followers remain stable.) capacitance or input capacitance of Power supply In the case of the input circuitry, the scope. Importantly, the input and output We’ve also added a 1nF capacitor circuits of the probe must be operated the purpose of IC1b is to establish a to form an RC low-pass filter here, to from separate power supplies, since “reference ground” voltage level for compensate for a peak in the frequency they are on opposite sides of the isola- the negative input connector CON2, so that when there is no input to the response of the circuit surrounding tion barrier. January 2015  31 if a battery happens to be connected backwards while S2 is on (easy enough to do, at least briefly), the diode will limit the voltage applied to IC1 or IC2 to no more than -1V, protecting it from damage. LED1 is fitted to make it harder to forget to turn the unit off when you’ve finished using it. As it’s a high-brightness blue LED, it only requires 100µA to operate, so it doesn’t add much to the battery drain during operation. Building the probe Scope3: the voltage across two outputs of the Induction Motor Speed Controller with an incandescent lamp as a load. The scope performs a sort of averaging when zoomed out like this, revealing the PWM-modulated sinewave shape. probe the non-inverting input of IC1a is biased midway between the V1+ and V1– rails. This allows the input circuit to operate the IR LED inside OPTO1 at close to “half brightness”, while also allowing IC1a to cope with the maximum possible AC voltage swing. On the output side, IC2b is again there to provide a half-supply reference ground, for the output connector CON3. And by making the exact reference voltage variable using trimpot VR2, we allow cancelling of any output offset voltage that might be caused by differences between the photodiodes inside OPTO1 at the quiescent current level. 9 Although the two supplies are on opposite sides of the probe’s isolation barrier, we switch them on and off in tandem using S2a and S2b, the two poles of a 250VAC-rated rocker switch. Typical mains-rated switches of this type are rated to withstand 1000V RMS, which just happens to be exactly what OPTO1 is able to withstand. To be safe, we’ve added some extra insulation between the leads connecting to the switch (as we’ll explain soon). Diodes D3 and D4 are connected to the switch such that the are reversebiased normally and thus do not affect circuit performance at all. But (SIDE VIEW) TIN THESE ENDS ONLY As mentioned earlier, all of the components and circuitry of the probe are built into a small ABS instrument case measuring 150 x 80 x 30mm. In fact everything except the two 9V batteries, on/off switch S2 and input jacks CON1 and CON2 is mounted on a single PCB measuring 122 x 70mm and coded 04108141. The board has cutouts on each side to provide spaces for the two 9V batteries, as you can see from the overlay diagram of Fig.4. On/off switch S2 mounts on the top of the case on the centre line and about 1/3 of the distance up from the output end, with short insulated and splayed leads connecting its lugs to the matching pads on the PCB. The two insulated input jacks CON1 and CON2 mount in the input end panel of the case with their connection lugs wired to the matching pads on that end of the PCB. Output BNC connector CON3 is mounted directly onto the PCB at the centre of the output end, with trimpots VR1 and VR2 spaced equally on either side. The trimpots are then easily adjusted using a small screwdriver or alignment tool, through matching holes in that end of the case. (END VIEW) (END VIEW) WHITE DOT MAKE SOLDER JOINTS SMALL AND SMOOTH HEATSHRINK SLEEVES 11.5 1 CUT 4 x 50mm LONG PIECES OF HOOKUP WIRE, STRIPPING INSULATION 4mm FROM ONE END & 37mm FROM THE OTHER END & LEAVING 9mm OF INSULATION ON EACH WIRE. TIN THE SHORT BARED ENDS OF ALL FOUR WIRES 2 IDENTIFY THE SWITCH LUGS TO WHICH THE WIRES WILL BE SOLDERED, ON BOTH SIDES OF THE SWITCH 3 SOLDER THE SHORT END OF EACH WIRE TO A SWITCH LUG, MAKING EACH JOINT SMALL & SMOOTH. THEN SPLAY EACH PAIR OF LEADS OUTWARDS TO SPACE THEM 11.5mm APART Fig.5: follow these steps in soldering leads to, then securing, S2 to the PCB. 32  Silicon Chip 4 CUT 4 x 11mm LONG PIECES OF 3mm DIAMETER HEATSHRINK TUBING AND SLIP OVER EACH WIRE & SWITCH LUG. THEN SHRINK THEM IN TIGHTLY USING A HOT AIR GUN OR THE SHANK OF A SOLDERING IRON. siliconchip.com.au To wire up the probe PCB, fit the components in the usual order: first the resistors (including VR1 & VR2), followed by the four diodes, the smaller capacitors and the six 100F electrolytics – taking care to fit the diodes and electrolytics with the correct polarity. Take care not to get the two types of diode mixed up. Next, mount transistor Q1, followed by the range switch S1, after cutting its spindle at a distance of 12mm from the end of the threaded ferrule. Then fit the switch to the PCB, taking care to use the orientation shown in Fig.4. Next fit IC1 and IC2, again making sure you orientate each one as shown. The next component to be added to the PCB is the HCNR201 linear analog optocoupler (OPTO1). Although it comes in an 8-pin DIL package, it has wider pin spacing than usual: 0.4” (10.16mm) rather than 0.3” or 7.62mm. It’s fitted to the PCB with the “notch” end towards the top. After this fit BNC output connector CON3 at the right centre of the PCB, midway between trimpots VR1 and VR2, followed by the four PCB terminal pins used to make the connections between the two battery snap leads and the PCB. Two of these pins are soldered into the pads just below the cutout for Battery 1 at upper left, while the other two go just to the left of the cutout at lower right, for Battery 2. You can see these quite clearly in Fig.4. Mount LED1 with the bottom of its lens 20mm from the top of the PCB. This will be with virtually the full lead length. Finally, cut the two battery snap leads themselves to about 45-50mm long (measured from the snap) and strip back about 5mm of the insulation from the wire ends. Thread the wires through the stress relief holes provided on the PCB and solder them to the terminal pins, again as shown in Fig.4. Your probe PCB assembly should now be complete, and can be placed aside while you prepare the box. Preparing the box There are no holes to be drilled in the bottom half of the case. All of the holes are drilled and/or reamed in the top half and in the two removable end panels. But as there are only nine holes in all, this shouldn’t be a problem. The size and location of all of the holes are siliconchip.com.au Parts List – Isolating High Voltage Probe for Oscilloscopes 1 PCB, code 04108141, 70 x 122mm 1 ABS instrument box, 150 x 80 x 30mm [Jaycar HB-6034] 1 4-pole 3-position rotary switch, (S1) 1 knob to suit S1, <25mm diameter 1 DPDT, 250VAC-rated rocker switch, single hole mounting (S2) [Jaycar SK-0994] 2 banana sockets, fully insulated, 1 red, 1 black (CON1, CON2) 1 PCB-BNC socket (CON3) 1 6mm long untapped spacer 1 15mm long M3 tapped Nylon spacer 1 15mm long M3 Nylon machine screw (cut from a 25mm long screw) 1 6mm long M3 machine screw 2 16.5mm long untapped spacers (cut from 25mm long spacers) 2 25mm long 6G or 7G countersunk self tapping screws 4 3.5mm ID flat washers 2 9V alkaline batteries 2 battery snap leads to suit 4 PCB terminal pins 1 100 x 26mm piece of 0.8mm Pressboard or Presspahn/Elephantide sheet Semiconductors 1 LM6132AIN/BIN dual high speed op amp (IC1) [element14 order code 9493980] 1 TLE2022CPE4 dual low current op amp (IC2) [element14 order code 1234686] 1 HCNR201-050E high speed linear optocoupler (OPTO1) [Digi-Key 516-2379-5-ND] 1 BC549 NPN transistor (Q1) 1 3mm blue LED (LED1) 2 1N5711 Schottky diodes (D1,D2) 2 1N4004 1A diodes (D3,D4) Capacitors Changes for 200:1 option: 6 100F 10V/16V PC electrolytic • Delete 220pF & 4.7nF ceramic 4 100nF multilayer monolithic ceramic capacitors 1 4.7nF 50V disc ceramic • Add three more 1nF ceramic capacitors 2 1nF 50V disc ceramic • Delete 16k& two 2kresistors 1 220pF 50V disc ceramic • Add two more 10kresistors 1 100pF 150V* disc ceramic 2 10pF 1.5kV* disc ceramic 1 4.7pF C0G/NP0 disc ceramic * 7.62mm lead spacing; 3kV types suitable Resistors (1% metal film 1/4W unless specified) 2 620k500V 1% 1/2W 1 560k500V 1% 1/2W (eg, Vishay HVR37) 1 200k 1 180k 2 62k 2 56k 1 16k 4 10k 2 2.0k 1 330 2 150 1 100 1 50kmulti-turn horizontal adjustable trimpot (VR1) 1 2kmulti-turn horizontal adjustable trimpot (VR2) shown in a drilling guide PDF which can be downloaded from siliconchip. com.au After drilling the smaller holes and reaming the larger holes to size, use a jeweller’s file or a sharp hobby knife to remove any burrs left around each hole on both the inside and the outside. To make a “dress” front panel for the probe you can make a photocopy of our artwork in Fig.8 (or download it from siliconchip.com.au) and then laminate it in a plastic sleeve for protection. After this it can be trimmed to size and attached to the top of the case using double-sided adhesive tape. Then cut holes in the dress panel for fitting the top PCB mounting screw, S2 and the control spindle for S1, using a sharp hobby knife and guided by the holes you have already cut and reamed in the case underneath. Making the isolation barrier Before you begin fitting everything into the case, you need to prepare the isolation barrier which will provide additional isolation between the input and output circuitry and their batteries. The barrier is cut from a 100 x 26mm January 2015  33 15mm LONG M3 NYLON SCREW (CUT FROM ONE 25mm LONG) EPOXY FILLET 6mm LONG UNTAPPED SPACER PRESSBOARD ISOLATION BARRIER S2 OFF/ON LED1 EPOXY FILLET 220p 15mm LONG M3 TAPPED NYLON SPACER RANGE Q1 BC548 S1 4.7nF 9V BATTERY BATTERY 2 IC2 6mm LONG M3 SCREW OUTPUT TO SCOPE + IC1 CON1 16.5mm LONG UNTAPPED SPACERS (CUT FROM 25mm LONG) IC3 25mm LONG 6G CSK HEAD SELF TAPPING SCREWS (BOTTOM OF BOX) CON3 4004 2x 3.5mm ID FLAT WASHERS ON EACH SCREW CUT OFF THESE SPACERS Fig.6: how it all fits into the case, as if looking through the side. Opposite is a photo of the completed unit. rectangle of 0.8mm thick pressboard sheet (similar to Presspahn Elephantide), using the upper diagram of Fig.7 as a guide, and then bent up as shown in the lower diagram. Preparing S2 The next step is to prepare on/off switch S2 by fitting it with the four well-insulated wires which will connect it to the PCB. As you can see from Fig 5.1 this needs four 50mm lengths of insulated wire, each with the insulation stripped by 4mm from one end but 37mm from the other end. (The long bared ends are to make assembly easier later.) We are using the two centre lugs and those at the ends opposite to the white dot on the red rocker actuator at the top of S2, as shown on the left in Fig 5.2. After soldering the short ends of the four wires to these switch lugs, each pair of wires is splayed away from the other pair as shown Fig 5.3, so that the pairs are spaced about 11.5mm apart. Then cut four 11mm-long lengths of 3mm diameter heatshrink tubing, and push each of these sleeves up one of the wires as far as it will go – that is, over the switch lug and the solder joint and until its top end is hard against the rear of the switch body (see Fig 5.4). After this use a hot air gun or the hot shank of your soldering iron to shrink each of the sleeves firmly into position around the wires and switch lugs. Then your “S2 switch assembly” should be complete, and ready to be fitted into place in the 18mm hole on the top of the case. This is done by unscrewing the large plastic nut, and then passing the switch and its splayed wires down into the box via the 18mm hole. Then screw 34  Silicon Chip the nut back on again inside the box, to hold it in position. But before you tighten the nut completely, make sure that the switch is positioned so that the white dot on its rocker actuator is positioned on the right, directly in line with the “ON” label of the dress front panel. Next, cut the two 25mm untapped spacers down to a length of 16.5mm, using a jeweller’s saw and smoothe off the cut ends using a small file. Then fasten them temporarily to the two mounting spacers moulded into the inside of the top of the case (at the output end), using the two 25mm long countersink-head self tapping screws with about five or six small flat washers under each screw head as packing, so the screws don’t enter the moulded spacers very far – just enough to hold the 16.5mm spacers in place. Then pass a 15mm long Nylon M3 screw (cut from a 25mm long screw) down through the central hole near the input end of the case front panel, slip the 6mm untapped spacer up over the end of the screw and fit an M3 nut – screwing it up to hold the 6mm spacer firmly against the underside of the front panel. You should now be almost ready to apply a fillet of epoxy cement around the top end of each of the three spacers, to hold them in place securely. But there’s one more thing to do first: fit the Pressboard isolation barrier into the top half of the case. Its 26mm-high “L section” should be over on the side ready to slip into the cutout for battery 2, with the 20mm-high section with its cutouts for S2 and OPTO1 passing “east-west” and aligned centrally between the contacts at the rear of S2. Once you’re happy that it’s in the correct position, it can be secured there using a few small dabs of epoxy adhesive between the barrier and the inside of the case top. Then while you have the epoxy cement mixed up, cement the spacers to the case top as well. When the cement has had time to cure, you can unscrew both of the Resistor Colour Codes p p p p p p p p p p p p No. Value 2 620k 1 560k 1 200k 1 180k 2 62k 2 56k 1 16k 4 10k 2 2.0k 1 330 2 150 1 100 4-Band Code(1%) blue red yellow brown green blue yellow brown red black yellow brown brown grey yellow brown blue red orange brown green blue orange brown brown blue orange brown brown black orange brown red black red brown orange orange brown brown brown green brown brown brown black brown brown 5-Band Code (1%) blue red black orange brown green blue black orange brown red black black orange brown brown grey black orange brown blue red black red brown green blue black red brown brown blue black red brown brown black black red brown red black black brown brown orange orange black black brown brown green black black brown brown black black black brown siliconchip.com.au Take note of the order of assembly in the text, especially the Presspahn isolation barrier (arrowed) which wraps around the lower battery and sits across the middle of the PCB, as indicated by the red dotted line. This is all necessary to ensure good isolation between the battery and PCB and between the two poles of the power switch. The next step is to attach the 15mm long M3 tapped spacer to the PCB (at top centre), using a 6mm long M3 screw passing up from underneath. It’s a good idea to tighten this screw firmly (but not TOO firmly) using a screwdriver, with the spacer held by a small spanner or nut driver. After this, mount the two input connectors CON1 and CON2 into the input end panel of the case, with the red one on the right as viewed from behind the panel. Tighten their nuts to secure them in place, and then solder a short length of tinned copper wire to the rear lug of each connector. Capacitor Codes Value μF value 100nF 0.1μF 4.7nF NA 1.0nF NA 220pF NA 100pF NA 10pF   NA 4.7pF   NA siliconchip.com.au IEC code 100n 4n7 1n 220p 100p 10p 4.7p EIA code 104 472 102 221 101 10 4p7 threaded ferrule of rotary switch S1 passes up through its matching hole in the top of the case. When the assembly can’t be pushed in any further, you should be able to secure it all together by screwing the two self-tapping screws back into the matching holes of the mounting spacers moulded into that end of the case top, and also by passing the 15mm long Nylon screw down through the matching hole in the centre of the input end of the case top, so it passes down through the 6mm untapped spacer and can then be screwed into the top of the 15mm long M3 tapped spacer. If you found this description somewhat confusing, try looking at Fig.6. This shows what you’ll be working towards. When the PCB assembly is secured 28 12 18 20 4.5 12 11 12 17.5 (FOLD UP BY 90°) Final assembly Then, with the centre axis of the two connectors positioned about 6mm above the top end of the PCB, solder each wire to its matching pad on the PCB. These pads are provided with a centre hole, so you can pass each wire down through the hole before soldering. Next, fit the output end panel of the case over the shank of CON3, after removing its nut. Then screw the nut back on again, to complete the PCBand-end panels assembly. By now you should be ready to fit this completed board assembly up into the top half of the case, by introducing it so that each of the two end panels slips into the matching slots in the ends of the case half, the four wires from S2 pass down through their matching holes in the PCB and the shaft and (FOLD DOWN BY 90°) 25mm long self-tappers and remove all but two of the washers on each, ready to secure the PCB shortly. At the same time you can unscrew the 15mm M3 screw and its nut holding the 6mm spacer in place, and you’ll be ready for final assembly. 30.5 26 17 100 MATERIAL: 0.8mm THICK PRESSBOARD/PRESSPAHN ELEPHANTIDE SHEET ALL DIMENSIONS IN MILLIMETRES Fig.7: here’s how to cut and fold the sheet of insulation material. It forms a physical barrier between the input and output sides. January 2015  35 in place as shown in Fig.6, you’ll be able to fit switch S1’s spindle with its control knob. Of course you’ll also need to solder the wires from S2 to their pads on the PCB, after which you can cut off their excess lengths. All that remains now is to attach each 9V battery to its snap connector, and then lower it into its waiting “slot” at the side of the PCB. The final assembly step is to fit the bottom of the case and fasten it in place with the four 20mm long countersink head M3 screws supplied with it. However just before you do this, you’ll need to cut off the two PCB mounting spacers moulded into the bottom of the case at the output end. This is because if left in situ, they’ll interfere with the heads of the mounting screws on the underside of the PCB. It’s not hard to cut off these spacers with a pair of sharp side cutters. After these “minor trimming” jobs, you should find that the bottom of the case will mesh nicely with the PCB-andtop assembly, allowing you to fit the four screws holding it all together. MAXIMUM INPUT VOLTAGES FOR THE THREE INPUT RANGES Set-up & calibration /500 1414Vp-p (500V RMS) /100 800Vp-p (280V RMS) /10 80Vp-p (28V RMS) There isn’t much involved in setting up and calibrating the probe. The first step is to connect a DMM (set to read DC volts, on its 2V range) to the probe’s output connector CON3 using a cable ending in a BNC plug. Now turn range switch S1 to the “/500” position, and also plug two input leads into CON1 and CON2. Connect their far ends together to make sure the probe definitely has “zero input”. Next turn on the probe’s power switch S2, and you’ll probably see the DMM reading move to a value slightly above or below 0V. The idea now is to adjust trimpot VR2 (Offset Adjust) in one direction or the other with a small screwdriver or alignment tool, to bring the reading as close as possible to 0V. This is the initial setting for VR2. However, it may have to be readjusted by a small amount after you have performed the second step – calibration. To calibrate the probe, the simplest approach is as follows. First connect its output (at CON3) to an input of your scope or DSO, using a reasonably short BNC-to-BNC cable. You can adjust the scope’s input sensitivity to, say, 1V per division and if it has a switch or option for setting its calibration to allow for a probe’s division ratio, set this to the 10:1 position. (This should change the effective input sensitivity to 10V/division.) Next turn the probe’s range switch S1 to the /10 position (fully clockwise) and connect the probe’s input leads to a source of moderately low voltage AC. This can be from an audio generator set to provide a sinewave at about 1kHz with an output level of say 10V RMS (= 28.8Vp-p) or a square wave or function generator set to provide a square wave of again 1kHz at about 20 - 25Vp-p. Or if you don’t have access to either kind of generator, you could use a step-down transformer with a known (ie, measured) secondary voltage of around 12-15V RMS (= 34 – 42.4Vp-p). When you now turn on the probe’s on/off switch (S2), you should see the waveform from your signal source on the scope’s display. Its frequency and amplitude should also be displayed if your scope has this facility built in, as most do nowadays. 36  Silicon Chip – INPUTS + DIVISION FACTOR /100 /500 /10 ON OFF POWER ISOLATING HIGH SILICON VOLTAGE PROBE CHIP FOR OSCILLOSCOPES OFFSET ADJUST OUTPUT TO SCOPE GAIN CALIBRATE Fig 8: same size front panel artwork – photocopy this (or download it from siliconchip.com.au) and glue it to your box before inserting S2. Now the odds are that while the frequency reading will be correct (either 1kHz or 50Hz as the case may be), the amplitude reading will probably be a little higher or lower than the known level of the signal being fed into the probe. So what’s needed now is to adjust the probe’s “Gain Calibrate” trimpot VR1 in one direction or the other using a small screwdriver or alignment tool, to bring the reading as close as possible to the correct value. After doing this calibration step, it’s a good idea to go back and repeat the first “Offset Adjust” step – especially if you had to turn VR1 quite a few turns to achieve calibration. This is done quite easily, simply by removing the probe’s input leads from your source of AC and connecting them together. Then after turning the range switch to “/500”, you can reconnect the probe’s output to your DMM and check what reading you get. If it has moved slightly away from the “0V” mark, it’s simply a matter of adjusting trimpot VR2 to bring it back again. Then your probe will be set up, calibrated and ready for use. SC siliconchip.com.au