Fullscreen HTML5Med Res (??MB)HTML4

It has often been described as about the third-mostuseful piece of test gear in a TV service tech’s arsenal. It’s easy (and cheap!) to build, easy to use and you will wonder how you got along without one . . . Design by Bob Parker Shorted Turns Tester www.siliconchip.com.au February 2004  63 OK, so you’re already asking: if it’s number three, what are one and two? Few would argue that the multimeter (or more likely a DMM these days) and a ’scope well and truly take the first two spots. But if you’re into repairing TV sets and/or video monitors, a shorted turns tester in your tool box or on the bench can save you hours of wasted time – not to mention a lot of expense. However, we’re getting a bit ahead of ourselves. What does it do? Ummm – isn’t that blindingly obvious? Shorted turns tester? Tests for shorted turns? Yes, it does just that – but unless you ARE a TV or monitor technician, you’re probably still none the wiser. Let’s go back a few steps. In all traditional (ie, CRT-equipped) TV sets and video monitors there is a horizontal output stage (also called the line output stage). You could regard this as the “business end” of the TV set/monitor. It’s job is to supply appropriate signals and the extra high tension (around 20-30,000V) the picture tube needs to make it operate. Operating at high voltages, frequencies and power levels, the horizontal output stage is one of the most-stressed sections of the circuit and is responsible for more than a fair share of faults in TV sets and monitors. Unfortunately, faults in the horizontal output stage are often difficult to find – and many a technician has replaced the principal (and most expensive) component, the line output transformer, only to find the fault is somewhere else. Perhaps the fault is in the highspeed rectifier diodes connected to the transformer’s secondaries. Maybe the horizontal output transistors have failed due to the stresses they are under. Or it could be an insulation breakdown in the deflection yoke on the back of the tube. But the fault that most technicians dread is a shorted winding within the line output transformer. Without the right test gear, the easiest way to test a line output transformer is by substituting a known good one. But we have already mentioned the fact that they are expensive – and, unfortunately, they are commonly NOT interchangeable from one brand to another. Another minor dilemma for the This project was first described in Electronics Australia in August, 1998 and has proved to be a very popular and enduring design with thousands sold around the world. It is re-presented here, with cosmetic changes only, for the benefit of 21st century readers! Note that existing stocks of the Dick Smith Electronics kit (Cat K-7205; $49.80) will include the old panel and instructions until the next run of kits. technician is that this section of the set can bite – badly. Most repairers are slightly less than enthusiastic about digging around the horizontal output stage while it is powered up. They’d much rather find a way to test a less angry set! All things considered, a technician needs to be fairly confident that the line output transformer IS faulty before going to the trouble of obtaining a good’un then substituting it (which usually means a bit of set disassembly). How do you test it? Most test equipment, including the one described here, is based on the fact that nearly all serious faults in horizontal output stage will greatly increase the losses in the primary circuit. The components in the primary circuit form a reasonably low loss resonant circuit (also called high “Q”), especially at low voltage levels. Shorted turns or components in the output stage will lower that “Q”. Find a way to check low Q and you have a handy piece of test equipment. Ring testing When you apply a fast pulse to the primary of the line output transformer (LOPT), the total inductance and capacitance will produce a decaying oscillation in the secondary, which may have a dozen or more cycles before it dies away to a low value. This is known as “ringing” Incidentally, it is called that because it is very similar to the effect you get when you strike a bell. You get a note that gradually dies away. If the circuit has shorted turns These two ’scope shots demonstrate not only the principle of operation of the Shorted Turns Tester (and also a ringing oscillation!) but also its effectiveness. The first shot is that of a known good line output transformer; the second is the same transformer with a dead short across one of the secondary windings (eg, a crook rectifier diode). In the first shot, all LEDs were lit; in the second only four. 64  Silicon Chip www.siliconchip.com.au or other faulty components in the secondary, the oscillations die away very much faster. Continuing the bell analogy for a moment, if you place your hands around the bell to stop it resonating, the bell sounds for a much shorter time. This principle is the basis of our Shorted Turns Tester. A fast pulse is applied to the primary of the transformer and the number of “rings” (or oscillations) are counted. If all is well, the circuit lights up a number of LEDs. If all is not so well, less LEDs light. If there is catastrophic failure (for example, a collector/emitter short in the horizontal output transistor(s) or a capacitor short) there will probably be no ringing at all, with no LEDs lighting. We’ll look at this in more detail shortly. Before we move on to the circuit description, it is worth noting that this Shorted Turns Tester works at low voltage and is designed to check the line output stage “in situ” – very much more convenient than having to remove the transformer or other components. The circuit There are three sections to the Shorted Turns Tester circuit (Fig.1): the oscillator, which produces the low frequency but fast-rising pulse; the comparator, which compares the amplitude of the oscillations produced by the transformer; and the LED bar-graph driver and display. The low frequency pulse generator: IC1b, one half of a LM393 dual comparator, is set up as a low frequency oscillator, whose output (pin 7) is normally pulled up to essentially the positive supply rail by the two 1kΩ resistors. The output switches down to 0V for about 2ms every 100ms, with the timing set by the feedback components between the inverting input (pin 6) and the output. It is during these low-going 2ms pulses that each ring test occurs. When IC1 pin 7 goes low, Q1 is driven into saturation and its collector voltage rises almost to the +6V supply. This makes two things happen. First, the 100pF capacitor, between Q1’s collector and the reset pins of IC2, sends a positive pulse of about 5us duration to those resets, which drives all the outputs of the four-bit www.siliconchip.com.au February 2004  65 switches cleanly between its low and high voltage levels. The result of all this is that an inverted and squared-up version of the ringing waveform appears at the output of IC1a, until the ringing amplitude has decayed down to about 15% of its initial value. This pulse train is connected straight to the clock inputs of the two shift registers in IC2. 3. The LED bargraph display: IC2 consists of a pair of identical four-bit serial-in/parallel-out shift registers, connected to form a single eight-bit unit, with each output driving one LED in the ‘bargraph’ display via the 1kΩ resistors. The serial data input of the first stage (pin 15) is permanently connected to the positive supply, or logic 1. Fig.2: follow the PC board overlay above and the photo at right and you should have no problems in assembling the project. It should take less than an hour to do. Remember to leave the LEDs until last, as explained in the text. shift registers to a low state. This switches off all the LEDs, in readiness for a new ring test. At the same time, D2 is forward biased, resulting in a brief 650mV pulse across the diode. This is coupled via the 47nF capacitor to the test leads and the LOPT primary winding. As previously explained, this causes (hopefully!) the LOPT circuit to ‘ring’, a bit below its natural resonant frequency due to the presence of C3 (which functions as the resonating capacitor when testing an LOPT on its own). 2. The ring amplitude comparator: The ringing waveform is coupled by a 10nF capacitor to the inverting input of comparator IC1b, itself DC biased to about +490mV by the voltage divider across the supply (4.7kΩ, 33kΩ and 150kΩ resistors). At the same time, D3 is constantly forward-biased and its entire voltage drop of about 600mV is applied to IC1a’s non-inverting input as a reference voltage, via a 10kΩ resistor. The 1MΩ resistor between the non-inverting input and the output of IC1a produces a small amount of positive feedback, ensuring that its output Partially assembled Shorted Turns Tester shows the battery holder in place in the case bottom and the PC board ready to mount on its threaded spacers with the LEDs poking through the front panel. In the DSE kit these holes are pre-punched, saving you a lot of time and trouble (rectangular holes are a cow to drill . . .) 66  Silicon Chip One measurement For the first 5us after the commencement of a new 2ms measuring pulse, both shift registers are reset to zero on all outputs, as described And here it is fully assembled, ready to close up and use. You might like to put some foam rubber between the PC board and batteries, just in case. www.siliconchip.com.au best use. Their responses are shown below, giving a good idea of the usefulness (and the limitations) of this tester. Putting it together Fig.3: the wiring is pretty simple because almost everything mounts on the PC board. If you get a DPST (or even a DPDT) power switch in your kit (as ours was), simply use the centre pin and one of the outside pins. earlier. At the same time the initial positive pulse applied to the LOPT drives IC1a’s output, connected to both shift registers’ clock inputs, to a low (logic 0) level - unless the test leads are shorted. If the LOPT primary circuit is OK, it will ring during the next several hundred microseconds. For each ring above about 15% of its initial value, it will cause a high-going pulse to be applied to the shift register clock inputs, resulting in the logic 1 on IC2 pin 15 being moved one shift register stage further along. It doesn’t matter if the LOPT rings more than eight times – all LEDs will still remain illuminated. So the overall result is that one LED illuminates for each LOPT ring cycle above 15% of the initial level, and this condition remains until the start of the next 2ms measuring pulse. Usage & limitations In order to assess the usefulness of this design, we gave several prototype Shorted Turns Testers to technician friends to evaluate for many months, then asked for their comments and thoughts on how to put the tester to Before soldering anything to the PC board, hold it up to a bright light and examine the copper side carefully for fine track breaks and, especially, whiskers or bridges - particularly where tracks pass close to component solder pads. Referring to the board overlay in Fig.2, begin installing the components, starting with the low-height components – the resistors and diodes - working your way up to the tall ones including the four PCB pins for `GND’, `HOT’ and `+6V’ terminal connections. Leave the LEDs off the board for now. Take care with the orientation of the polarised components, including the IC sockets. With everything but the LEDs installed on the PCB, once again illuminate it from the top, then check for and correct any solder bridges or other problems. Now turn your attention to the front panel, mounting the banana sockets and the power switch in their respective holes. Attach the tapped spacers to the corners of the board using plain 3mm screws and solder long component lead offcuts to the `GND’, `HOT Collector’ and `+’ solder pads, followed by the battery snap’s black wire to the `-’ pad. Next, without soldering them, poke the leads of all the LEDs through their respective holes in the board. Make sure the coloured LEDs are in their correct places, and that all the (long) anode and (short) cathode leads are correctly oriented as shown in Fig.??. Using black countersunk 3mm screws, Fig.4: this drawing should give you a pretty good idea of how it all goes together. Only the battery holder mounts in the case itself – everything else “hangs” off the front panel. www.siliconchip.com.au February 2004  67 Comments from the field: the Shorted Turns Tester under test! Our sincere thanks to Larry Sabo, Michael Caplan and Wayne Scicluna for their assistance in completing this project. We couldn’t have done it without you! Larry Sabo is an experienced monitor technician in Ottawa, Canada: One of the first things I do to check out a monitor is connect the tester between the HOT collector and ground. If no or only a few LEDs light, I check the HOT, damper diodes and tuning caps for shorts using a DMM. If these are OK, I check for an open fusible resistor in the circuit feeding B+ to the LOPT, and for shorts/ leakage in diodes on the LOPT secondaries. I also check the bypass capacitor on the DC supply to the LOPT primary for excessive ESR. If these check OK, I ring the horizontal yoke with its connector unplugged. It will normally ring seven times on its own. If the yoke rings OK, I unsolder all but the LOPT primary winding and ground pins, and ring the primary. If the primary still rings low with everything else disconnected, the LOPT is probably defective. Most LOPTs on their own will ring 8+ times, but some ring only four or five, even when they are perfectly normal. So it is prudent to confirm the diagnosis by ringing an identical known-good LOPT, if at all possible. Sometimes an LOPT is defective, but still rings normally with the tester, eg, due to leakage or arcing that only occurs at full operating voltage. The problem will sometimes be manifest by heavy loading of the B+ supply, spurious ringing and/or reduced voltages on the HOT collector, or excessively high EHT resulting in HV shut-down. Because this tester uses impulses of only 650mV to minimise the forward biasing of semiconductors, such defects will not be reflected in the ring count. In these circumstances, I check for measurable leakage resistance between the EHT cap and the other LOPT pins. It should be unmeasurable, otherwise the LOPT is defective. If I have gone through the above tests and have 68  Silicon Chip these symptoms and a normal ring count on the tester, the diagnosis can usually be confirmed only by substituting a known-good identical LOPT, or by testing with a chopper similar to the one described in Sam Goldwasser’s Electronics Repair FAQ, located on the Internet at http://www.repairfaq.org/ sam/flytest.htm. Something else I do when testing a LOPT is to supply it with a reduced B+ to enable scoping the HOT and measuring EHT (in situations where the monitor goes into HV shutdown). To reduce the B+, I use two light bulbs in series, one end to B+ supply, centre-tap to LOPT B+ connection, other end to ground. One bulb is 60 watts, the other is 100, so I can reverse the end leads and increase or decrease the B+ value used in testing. At the outset, when I have power supply cycling but have confirmed there are no shorts from HOT-C to ground, I substitute a dummy load (60W bulb) for the LOPT where the B+ enters, to see if the power supply works with the LOPT out of the equation. Overall, the LOPT tester can identify about 80% of LOPT failures. When trying to solve a puzzle, if someone offers information that is right 80% of the time, it’s a lot better than having to guess 100% of the time, especially if the ante is the price of a LOPT and wasted, valuable time. Michael Caplan does general electronic servicing in Ottawa, and added the following useful points in relation to TVs: It’s pretty straightforward to use, with the usual precautions of ensuring that the under-test unit power is off and any caps are discharged. When testing an LOPT in circuit, it might be necessary to disconnect some of the LOPT terminals, and/or yoke plugs that could load it down and upset the readings. The tester will often not detect bad HV diodes in integrated split-diode LOPT units, nor shorts/arcing that is voltage dependent - but then no other passive tester does either. I have found it useful for checking TV deflection yokes, both horizontal and vertical. A good yoke lights at least five and typically the full eight LEDs. However, many yokes have built-in parallel or series damping resistors, and these must be temporarily disconnected. Otherwise the reading will be low, even though the winding itself is fine. The tester can be used for checking high-Q transformers such as those used in SMPS’s. However, my experience has shown that it will not provide more than a two or three LED indication for good TV horizontal drive transformers. It can be used for these, however – to indicate shorts (no LEDs lit). On the other hand the ESR Meter (Dick Smith catalog number K-7204) can do much the same with these low resistance transformers. Wayne Scicluna services TVs in Sydney, and is the technician who talked me into developing the tester in the first place. Here are his hints: If you’ve already checked for the more obvious leaky and shorted semiconductors and capacitors etc., and are still getting a low reading on the tester, there are some other traps to avoid. You need to get a good connection with the test leads, because contact resistance can cause a low reading. The same applies to defective solder joints in the horizontal output stage, especially on the LOPT itself and HOT. In fact connecting the tester with clip leads, flexing the board and wiggling components is a good way to show up bad solder joints in this area. Body conductivity can also cause a lower than normal reading if you’re touching the test leads and your skin is damp. Low readings can also be caused by having the test leads reversed, i.e., connecting ‘HOT Collector’ to chassis, and by faults in an external voltage tripler. www.siliconchip.com.au Parts List – Shorted Turns Tester 1 PC Board, code ZA1137 (51 x 76mm) 1 plastic case, 130 x 68 x 41mm (DSE H-2853); 1 front panel to suit (prepunched and screened) 4 PC pins 1 red 4mm banana socket 1 black 4mm banana socket 1 set red/black test leads with 4mm banana plugs 1 4x AAA flat battery holder 1 battery snap 1 SPST power switch, push on/off 1 8-pin DIP IC socket 1 16-pin DIP IC socket 4 M3 tapped spacers, 15mm; 4 M3 x 6mm screws (zinc plated) 4 x countersunk M3 x 6mm screws (black) 4 x countersunk No4 x 6mm screws (black) double-sided adhesive tape Semiconductors 1 LM393 dual comparator (IC1) 1 4015 / MC14015 / CD4015 dual 4-bit shift register (IC2) 1 BC328 / 2N5819 PNP silicon transistor (Q1) 3 1N914 / 1N4148 silicon diode (D1-3) 3 Rectangular red LEDs (LED 1-3) 2 Rectangular yellow LEDs (LED 4,5) 3 Rectangular green LEDs (LED 6-8) Capacitors 1 100µF 16/6VW RB electrolytic 4 47nF MKT polyester (code 473 or 47n) 1 10nF MKT polyester (code 103 or 10n) 1 100pF disc ceramic (code 101 or 100p) Resistors (All 5% 0.25W carbon or better) 1 2.2MΩ 4 1MΩ 1 150kΩ 2 47kΩ 1 33kΩ 1 10kΩ 3 4.7kΩ 11 1kΩ 1 270Ω TEST COIL: 1 Balun core (DSE Cat. R-5440) 2 metre length 0.25mm enamelled copper wire attach the front panel to the board assembly and place the whole thing face-down on a soft flat surface. Manoeuvre all of the LEDs into their cutouts in the front panel, and push each LED down slightly to ensure its face is level with the front of the panel. In the unlikely event that a LED won’t fit, use a small file or similar to remove the excess powder coating inside the hole. Now solder all the LEDs into place, then connect the test lead sockets and the closest terminal of the power switch to their respective wires from the board, and finally the red battery snap wire to the free switch contact (refer to Fig.3, the wiring diagram). Snip off the battery holder’s PCB mounting pins, then install four ‘AAA’ cells into it. Connect the battery snap to the terminals, and switch the unit on. If everything’s OK then the bottom red (‘1’) LED will illuminate and shorting the test leads will cause it to go off. An effective way to test the unit is to connect the test leads to the primary winding of a known good LOPT out of circuit, which should bring all eight LEDs on. Then thread a loop of solder around the ferrite core of the LOPT (simulating a single shorted turn), and the LED count should drop to 1-3 as the loop is closed. If everything’s OK, use double-sided adhesive tape to stick the battery holder into the bottom of the case, with the cells aligned in a “north-south’ direction for easiest access. All that remains to be done now is to screw the front panel into place and try out your tester on some LOPTs and their associated circuitry. Winding a Test Coil In order for constructors to test the unit once assembled we have provided details and parts to construct a simple transformer coil which enables the circuit to light all ‘8’ LEDs. Your Dick Smith Electronics kit should include a Balun core (R 5440) and about two metres of 30B&S (0.25mm) enamelled copper wire. Construction is very simple. Using the balun core provided, wind around 45 turns (tightly wound) through the two centre holes. Once completed trim the wires to approximately 50mm and clean the enamel from each end so that a positive connection can be made. Now test the coil in the Shorted Turns Tester. It should display all eight LEDs. Feeding through an additional winding and shorting the ends (remember to remove the enamel!) will reduce the “rings” to either one or two LEDs, giving a good indication that the unit is working correctly. SC Resistor Colour Codes o o o o o o o o o No. 1 11 3 1 1 2 1 4 1 Value 270Ω 1KΩ 4.7kΩ 10kΩ 33kΩ 47kΩ 150kΩ 1MΩ 2.2MΩ www.siliconchip.com.au 4 Band (5%) red violet brown gold brown black red gold yellow violet red gold brown black orange gold orange orange orange gold yellow violet orange gold brown green yellow gold brown black green gold red red green gold 4 Band (1%) red violet brown brown brown black red brown yellow violet red brown brown black orange brown orange orange orange brown yellow violet orange brown brown green yellow brown brown black green brown red red green brown 5 Band (1%) red violet black black brown brown black black brown brown yellow violet black brown brown brown black black red brown orange orange black red brown yellow violet black red brown brown green black orange brown brown black black yellow brown red red black yellow brown February 2004  69