Oscilloscopes Episode 5 pptx

68 Oscilloscopes repetitive waveforms right up to and beyond the chopping frquency, though there is little point in so doing. The choice of trigger source is very important when working with a dual trace oscilloscupe. As mentioned earlier when discussing the waveforrns encountered in a decade divider stage, if the frcqucncics being displayed on the Y1 and Y2 t.races are different bt.~~. relat.ed, one should trigger from the lower frc- qurncy, whcthcr it be displaycd on the Y 1 or the Y2 (.race. Dual trace scopes usually have a 'mixed trigger' facility; this Incans that when used in the alternate mode with internal triggering, the sweep will be triggered from the Y1 channel when displaying thc Y1 trace, but on the next sweep will display the Y2 signal triggered from the Y2 channel. Consequently both traces will be perfectly synchronized with their respective displayed signals and the traces will appear to have a fixed stable relationship. In fact, the signals displayed on the two channels could have totally unrelated frequencies, as would he apparent if triggering from Y1 were selected, in which case the Y2 trace would not be synchronized, and vice versa. In the mixed triggering mode in fact, the oscilloscope is simply equivalent to two entirely separate single channel scopes, each internally triggered Ircm its owi signal. Nevertheless, mixed ~riggering can bc very uselul lor keeping an eye on two unrelated wavetornis simultanc~ot~sl~, prrividcd this fact is borne firmly in mind. Care is nccded even when ihc two Irc~quencics arc ha rm( )ri ica 1 I y re la 1 cd or id cn t ica I. M i x cd t riggcri ng w i I1 show the 0" reference oiitpLit and the 90" qtidralurc outpu.t of a quadratiirc oscillator as bcirig iii phase, whereas triggering from the reference input will show the correcL 90" phase diflerence between the two sine wavcs. The moral is to use mixed triggering only whcn it is specifically required. and to regard the selection of the appropriate triggering arrangements as an essential part of setting LI~ a dual trace scope. Many dual trace oscilloscopes provide the option of displaying a single trace which represcants the sum of the voltages applied to the Y1 and Y2 inputs. In addition, it is possible to invert one of the traces, say Y2, so that posi~ivc-going inputs deflect the trace downwards and negative inputs upwards. It is thus possible to using oscilloscopes 69 display Y1 - Y2, i.e. the difference between the two input signals, instead of the sum. This will result in no deflection of the trace if the same signal is applied to both Y inputs - provided they are set to the same volts/div setting (and both variable controls, if provided, are at the calibrated position). Thus the oscilloscope will only respond to the difference between the two inputs, just what is wanted for examining two wire signals that are balanced about ground. This property of ignoring or rejecting identical signal compo- nents at the two inputs is called 'common mode rejection' or 'input balance'. The unwanted 'push-push' or common mode component that is rejected is referred to as 'common mode noise', 'longitudinal noise' or 'noise to ground', whilst the push- pull signal is called the 'transverse', 'metallic' or 'normal mode' signal. Two-wire balanced transmission systems are widely used, e.g. for transducer signals in factory process control systems, as twisted pairs in multi-pair telephone cables and for the two-wire overhead subscriber's loop connecting the domestic telephone to the nearest telegraph pole. The Y1 - Y2 mode will typically provide a 26dB CMRR (common mode rejection ratio), meaning that the sensitivity to undesired common mode signals, e.g. 50 Hz mains hum, is only one-twentieth of the sensitivity to the wanted transverse signal. This is only a modest degree of input balance compared with special scopes and other instruments specifically designed for working on balanced systems. However, balanced systems are generally used only up to a few hundred kilohertz at most, and instruments specifically designed for such use are correspond- ingly limited in bandwidth. Note that if 10:1 passive divider probes are in use, the 20:1 CMRR may be degraded, owing to within-tolerance differences in the exact division ratios of the two probes. With or without probes, the CMRR can be optimized by connecting both inputs to the same signal source and adjusting one or other Y channel variable gain control to trim down the gain of one channel to exactly match that of the other. With care, up to 100:1 CMRR (40 dB balance) or more can be obtained for signals up to a few hundred kilohertz, but this will not usually be maintained over the full bandwidth of the scope. To maintain this 70 Oscilloscopes increased CMRR, readjustment will also be necessary if the two Y input volts/div switches are set to another (common) setting. When using an oscilloscope's Y1 -Y2 mode for balanced measurements, beware of a potentially severe limitation. If the unwanted common mode signal (e.g. mains hum) is much larger than the desired signal, it can overload the Y input amplifiers, resulting in a distorted and inaccurate display. This problem can be avoided by using a purpose-designed differential probe. In the Tektronix P6046 Differential Probe and Amplifier Unit, the differential signal processing takes place in the probe itself, the amplifier producing a single-ended (unbalanced) 50~ ouput suitable for connection to any oscilloscope's Y input channel. The P6046 provides 10 000:1 CMRR at 50 kHz and no less than 1000:1 even at 50 MHz, while common mode signals up to +5 V peak to peak (+50V with the clip-on xl0 attenuator) can be handled without overload, even when examining millivolt level signals. In power engineering it is often necessary to examine small signals in the presence of very large common mode voltages, for example when checking that a silicon controlled rectifier's gate to cathode voltage excursion is within permitted ratings, in a motor control or inverter circuit. The Tektronix A6902B Voltage Isolator uses a combination of transformer- and opto-coupling to provide up to +3000 V (d.c. + peak a.c.) isolation from ground for each of two input channels. Designed for use with any two-channel oscilloscope, the A6902B permits simultaneous observation of signals at two different points in the same circuit, or signals in two different circuits without respect to common lead voltages. The two channels can also be combined to function as an input to a differential amplifier, for floating differential measurements. Use of Lissajous figures It might seem that nowadays the use of Lissajous figures for comparing frequencies is 'straight out of the Ark'- why not simply use a frequency counter? But in fact there are several cases where the use of a Lissajous figure can provide more information, and provide it faster. Suppose, for example, one had a precision 1 MHz frequency standard consisting of an oscillator controlled by an ovened Using oscilloscopes 71 (i) frequency ratio 3:1 (Y:X) (ii) frequency ratio 3:2 (a) (iii)-frequency ratio 3:2 (but with different phasing) (b) Figure 5.6 (a) Lissajous figures (courtesy AEG Telefunken). (b) Frequency measurement with Lissajous patterns requires a known frequency sine wave on one channel, usually the X channel. If the unknown frequency has the exact ratio to the known frequency as shown above, then (depending on the phasing) the trace will be like one of those shown. Other ratios, e.g. 2:3, 3:4, etc., will give stable, though more complicated, patterns. In principle, any rational number (i.e. m:n where m and n are integers) will give a stable pattern (courtesy Tektronix UK Ltd) 72 Oscilloscopes crystal. One could check its frequency with a digital frequency meter if the latter's internal reference were accurate enough, or could be independently checked. In the UK (and over much of Europe), one could check by counting the carrier frequency of the BBC's Droitwich transmitter, whose carrier is maintained to an accuracy of one part in 10 ~. In fact, 'off-air frequency standards' are available commercially; these receive the Droit- wich transmission, strip off the amplitude modulation and supply a 1 MHz output locked to the carrier. However, even a 10 second gate time will only allow a 1 MHz frequency to be checked to an accuracy of +1 count in 107, which makes checking the frequency meter and adjusting the 1 MHz crystal oscillator a tedious business. Even then, the accuracy achieved will fall far short of that available from the Droitwich carrier. Suppose now that a Droitwich-derived 1 MHz sine wave and the crystal oscillator under test are displayed as a Lissajous figure; the effect of adjusting the crystal oscillator can be observed immediately and continuously. A frequency difference of as little as one-hundredth of a hertz can be noticed in an observation time of a second or so, as the figure slowly drifts through the line-ellipse-circle repertoire of patterns. A counter would still have an uncertainty of plus or minus one-hundredth of a hertz or more, even after an observation time of 100 seconds. The Lissajous figure can also provide information about the stability and spectral purity of an oscillator. For example, if two independent conventional r.f. signal generators are both set to 100 kHz the resulting Lissajous display should be stable, giving a clean line and a round circle as the inevitable small frequency difference causes the figure to cycle slowly through its series of patterns. If now a Wien bridge type of RC oscillator is substi- tuted for one of the signal generators, the poorer frequency and phase stability of this type of oscillator will be immediately apparent. The circle, instead of being perfeclly round, may show minor dents and the figure will wobble, rather like a jelly being carried on a plate. This is evidence of very low-frequency noise FM sidebands, which it would be difficult to resolve with even the most sophisticated spectrum analyser. Using oscilloscopes 73 Z axis input A useful feature of many oscilloscopes is a 'Z axis' input. In Cartesian coordinates the Z axis is the third dimension at right angles to the X and Y axes, and therefore the same as the direction of the electron beam when the spot is at the centre of the screen. With no connection made to the Z axis input, the oscilloscope works normally with the trace brightness controlled by the intensity control, also affected by the timebase speed and sweep repetition rate as explained earlier. Applying a varying voltage to the Z axis input alters the brightness of the trace in sympathy. Some oscilloscopes have d.c. coupling of the Z axis input, but a.c. coupling is much cheaper and therefore more common, whilst positive-going voltages result in a decrease of brightness if, as is commonly the case, the Z axis input is coupled to the cathode of the c.r.t. The facility is useful in a variety of ways, one interesting example being the display of 'eye diagrams'. These are a way of examining the degradation due to imperfections of the modems and noise accompanying the signal at the receiver, in a digital phase-modulation communications link- Figure 5.7. The receiver for such a system will have a clock timing recovery circuit; displaying the i.f. (intermediate frequency) waveform at the receiver with the scope triggered from this will not produce a coherent or useful picture. Bandwidth is a scarce and hence expensive commodity, and the sudden changes of phase shown in Figure 5.8(b) imply the presence of wide signal sidebands. The modulated carrier at the transmitter is therefore first processed to produce a smoothly changing phase (by filtering and limiting, or other means) before being transmitted- Figure 5.8(c). This illustrates 'BPSK' (binary phase shift keying) where there are just two possible transmitted phases. 'QPSK' (quadrature phase shift keying) systems have four possible phases at each clock or data stable time, permitting the transmission of two bi.ts of information per clock cycle or 'symbol'. To display an eye diagram, the recovered clock or symbol timing is used to generate a narrow pulse occurring at the clock edge or data-stable time. This is applied to the Z axis input to 74 Oscilloscopes bright-up the oscilloscope trace. The timebase runs repetitively, triggered from the receiver's carrier recovery circuit, or possibly in a bench test set-up, derived from the transmitter carrier as shown in Figure 5.7. As the trace is invisible except during the bright-up pulse, i.e. at the sampling instant of the receiving test C lOck : r ncOder [ TRANSMITTER baseband t i lterl phase equaIiser modulator PA ( generator l) I E ST GEAR I" I , .~ I_LL br~ht -up pulse generator ~~176176 ~-,! delay / I ~H'e'l "iext Z mad. osc it loscope t -" ext trig q I I I I monostable/ I /J n put ' l I I J __ lc,oc l 9 - 1 extrocl - dec~d e r J" ~] d ern~ ,~J ox L RF''- -ulator data to error-rote test set RECEIVER Figure 5.7 (simplified) dummy Ioacl and pat h- I o~s simulator Block diagram of digital phase-nl()dulation radio data link on test Using oscilloscopes 75 (a) (b) (c) 0 I I I II I I I I I I l 1 I i 1 c lock I I = ! I I I 0 0 A , l i I I I I I , J data (NRZ) [ , I I I i 1 I t ! I ! I impractical digital phase mOdulatlon cJbrupt transitions I I I I I I ; I Ic Ic Ic ;c Ic I I f ' ! i:)ractlcal modulator output, smooth phase trans~tlons L Figure 5.8 (a) Clock and typical data stream of data link shown in Figure 5.7. (b), (c) Modulated r.f. output waveforms; note that both have the same phase at clock times 'C'. For clarity the r.f. is shown as exactly five times the clock frequency; in practice it would be many thousands of times and with no exact relation modem, the phase of the received signal will be (ideally) in one of two possible positions 180 ~ apart, as indicated in Figure 5.9 (a), or in one of four possible positions in the case of QPSK. The resultant picture is called an eye diagram. In Figure 5.9 the open eye, such as should be obtained with a well-set-up system, indicates little distortion; the nearly closed eye shows a system with excessive 'intersymbol interference' due to poor modem design. Figure 5.9(b) alternatively gives an impression of what one might see 'for real' over a digital radio link with a very low received signal strength, the poor signal to noise ratio resulting in a nearly closed eye, and in consequence a high 'BER' (bit error rate) in the received data. 76 Oscilloscopes With the DSP (digital signal processing) capability built into modern DSOs, it is possible to derive more information than ever from an eye diagram. Figure 5.10(a) shows (diagrammatically) a DSO acquiring points on a 'clean' eye diagram; with a poorer signal there would be more randomness to the point positions. Figure 5.10(b) shows how with a 'bit mapped' display with I6 bits per 'pixel', the instrument can, over a period, totalize the number of sampled points falling in each pixel. The resultant eye Figure 5.9 (a) Two-level digital phase-modulated signal showing well-set-up system with n~, intersymbol interference. (b) Poor system with bad intersymbol interference Using oscilloscopes 77 diagram can be displayed in colour, with, say, single or low count pixels shown in shades of blue, through the spectrum to red for the pixels with the highest counts. Additionally, the data can be further processed to show histograms illustrating the 'openness' of the eye in various ways, Figure 5.10(c). The oscilloscope in servicing Several of the facilities of a good scope have been discussed above in connection with specific applications. The rest of this chapter looks at other particular areas of use for a scope. First, TV servicing is considered briefly; for a more extensive treatment of the topic reference should be made to one of the many excellent books available dealing specifically with this subject. It is important to pay due regard to safety when working on any type of mains operated equipment. This is doubly true when working on TV sets, as some of them do not have the circuitry and chassis isolated from the mains. The circuitry of the ubiquitous 12 in black and white portable set is designed to run from 12 V d.c. in order to permit operation from a car battery when required. For mains operation a step-down transformer, rectifier and smoothing supply the required 12 V d.c. Thus only the transformer primary is at mains potential, the rest of the set being isolated. Larger mains-only colour TV sets may have a type of switchmode power supply providing full mains isolation, but this is by no means invariably so. To avoid drawing a d.c. component from the a.c. mains (which was quite normal in the days of valved TV sets), non-isolated sets use a fullwave rectifier: as a result the set's circuitry and chassis can be at approximately half the mains voltage. The only safe way to proceed when working on a TV chassis is to run it from a mains isolating transformer of a suitable rating. A 500 VA transformer should be more than adequate. The television set's chassis should be firmly earthed, as is the case of the oscilloscope. Even then, one must be very wary of the high voltages present in the line deflection and e.h.t, sections of the receiver. No one should work on a TV set without adequate knowledge and expertise. Even apart from the safety aspect, many faults will prove difficult or impossible to rectify without the full servicing [...]...78 Oscilloscopes I I0 ~ =0 Trig 0 _i-l_fi.I-Lftfi_fi.fi~LllJlJ-l~~ F-! FL! LI-L.I L! L ,., Clock to Trigger Random Data to Ch 1 (a) t ! I 256 pixeis I i F J // i t/ I: 51 2 plxels / / I 11 A sample fell on this (b) pixei 27 times Using oscilloscopes Figure 5. 10 Ltd) 79 Measurements on eye diagrams, see text (courtesy Tektronix UK... examining a 25MHz sine wave with a scope having a quoted 25 MHz bandwidth the trace will show only 71 per cent of the true amplitude of the signal Furthermore, the waveform being observed may be ralher severely distorted if not a sine wave, since the harmonics have frequencies of 50 MHz, 7 5 MHz, etc., and the oscilloscope's response at these frequencies will be very low indeed Even an ideal 25 MHz squarewave... devices connected to the component under test (courtesy Tektronix Inc.) 86 Oscilloscopes Figure 5. 14 With its 100Ms/s 8 bit samples and 100MHz bandwidth, the Metrix OX8100 provides sensitivities down to 2 mV/division on both input channels, and sweep speeds down to 50 ns/div (reproduced by courtesy of Chauvin Arnoux UK Ltd) Figure 5. 15 Economical, both on price and bench space, the 100MHz bandwidth TDS224... Figure 5. 12 Tile two channel HM20 05 real-time analogue oscilloscope features a 200MHz bandwidth, limebase speeds tt~ 2ns/divisit~n (with • magnifier on) and separate lrigger controls for lhe A and I3 timebases Cursor functions provide alphanumeric readout of w~ltage, time and frequency measurements; an RS232 interface and component tester (see Figure 5. 13) are built in (courtesy Hameg Lid) Using oscilloscopes. .. Tliiis when trying t o obscrvc a signal that is beyond the lull-screen handwidth of tlic oscilloscope, tlic arriount of c x ~ r i handwidth IO ~ had b y rcducing the displayed b c b Using oscilloscopes 83 Figure 5. 11 With a m a x i m u m sampling rate of 40 Ms/s, the IEC 1010-I Cat III safety-rated OX8032 features true differential inputs With input sensitivity ranges from 10 mV/div to 200 V/div, floating... channels With a nlaximum sensitivity of 2 mV/div., the TDS224 offers display options including s i n x l x interpolation, dot or vector dot-joining, persistence of 1, 2 or 5 seconds, or infinite, or OFF (courtesy Tektronix UK Lid) Using oscilloscopes 87 Some m o d e r n mains-operated scopes use a direct off-line switching p o w e r supply In addition to cost and weight reduction, this a r r a n g e m... of the scope for ct)mparism Unlike third harmonic (and other higher odd-ordcr) distorticin c ~ j r r ~ p o n e ~ ~ t s , and second other cvcn-ordcr harrnonic disrortion affect the positive and Using oscilloscopes 81 negative half-cycles of the w a v e f o r m differently, usually m a k i n g one flatter and the other more peaky Consequently, with care even 1 per cent of second-order h a r m o n i... been propped up at the top end w i t h n u m e r o u s bits of compensation and peaking c i r c u i t s - this enables the m a n u f a c t u r e r ' s sales d e p a r t m e n t to quote an impres- 84 Oscilloscopes sive figure for t h e i n s t r u m e n t ' s b a n d w i d t h In this case, at f r e q u e n c i e s b e y o n d t h e design m a x i m u m , t h e Y amplifier r e s p o n s e m a y fall... o n g sequence can easily give you p e r m a n e n t loss of colour! The most convenient type of scope for TV servicing has built-in line and frame sync separator circuits, e.g the Fluke model 3094 80 Oscilloscopes featured i.n Chapcer 3 These are handy when examining t.h.e operation of line and frame dellrt:iion circuits respectively, particularly when the set is receiving live programme material... Cursor functions provide alphanumeric readout of w~ltage, time and frequency measurements; an RS232 interface and component tester (see Figure 5. 13) are built in (courtesy Hameg Lid) Using oscilloscopes 85 standing, a n oscilloscope can p r o v i d e c o n s i d e r a b l e u s e f u l inform a t i o n , e v e n if o n l y of a q u a l i t a t i v e n a t u r e , w h e n handling signals well b e y o n . single-ended (unbalanced) 50 ~ ouput suitable for connection to any oscilloscope's Y input channel. The P6046 provides 10 000:1 CMRR at 50 kHz and no less than 1000:1 even at 50 MHz, while common. RECEIVER Figure 5. 7 (simplified) dummy Ioacl and pat h- I o~s simulator Block diagram of digital phase-nl()dulation radio data link on test Using oscilloscopes 75 (a) (b) (c) 0. pixeis J // i t/ I: i I F t ! I I / / 11 51 2 plxels A sample fell on this pixei 27 times (b) Using oscilloscopes 79 Figure 5. 10 Measurements on eye diagrams, see text (courtesy