Fullscreen HTML5Med Res (??MB)HTML4

Vintage Electronics The Telequipment D52 Dual-Beam Oscilloscope The D52 dual-beam 6MHz oscilloscope was quite the creation in the late 1960s. It was British made and a definite competitor with the American-made Tektronix scopes. Within Australia, both Tektronix and Telequipment Scopes were marketed and sold by Tektronix Australia Pty Ltd in NSW. By Dr Hugo Holden T he Telequipment scopes had very characteristic front panels and knobs. Some Telequipment apparatus with these knobs got used as props on the panels of the flying craft in Gerry and Sylvia Anderson’s brilliant puppet TV series, The Thunderbirds. As soon as I saw a Telequipment scope, I recognised the appearance as being what I had seen on some control panels in that TV show as a boy. Perhaps that was one thing that made me more interested in them. The original arrangement on these units used UHF sockets for the scope’s probes, as can be seen from the advertising photos. I changed them to BNC connectors on my scope to make them compatible with many more modern probes. BNC panel connectors that are made to be an insulated panel mount fit perfectly into the hole for the UHF sockets. My scope has the orange filter. This was an option that Telequipment offered when it was fitted with a dual-phosphor CRT. The CRT has a short blue and long yellow persistence phosphor. This is designated as P7 (or a GM suffix). If you were interested in short timeframe events, you would fit a blue plastic filter in front of the CRT (this blocks yellow). Alternatively, if you were interested in slower events, such as a cardiac ECG, you fit the orange filter which lets the yellow through, tinting it orange, while blocking the blue. Many D52 scopes simply had the usual green medium-persistence phosphor designated P31 (GH) with a green filter. A blue (P11 phosphor) CRT option was also available, but only in Photo 1: using the orange filter shows just the slower of the two phosphors, so very fast transients are removed. Photo 2: without the filter, the traces look white; the result of the yellow and blue light mixing. Photo 3: the phosphor looks blue when viewed through the side wall of the cathode ray tube. Australia's electronics magazine siliconchip.com.au 92 Silicon Chip the 12-pin version. There were two CRT variants for this scope, with either 12- or 14-pin bases. Therefore, for the D52, there were five possible CRTs it could use, according to the manual. Photo 1 shows the typical result with the orange filter. With the filter removed, the trace looks white (Photo 2), which is the blue and yellow mixing. The yellow phosphor was applied to the CRT glass first, then the blue after that. Looking at the inside of the CRT (which can be seen through the side wall of the tube), the internal appearance of the phosphor is vivid blue (Photo 3). The D52 is a valve-based scope; nearly all the circuitry in the timebase and vertical amplifier circuits uses valves, mainly the ECC88 dual triode, ECF80 triode-pentode and 6AL5/EB91 dual diode. However, this design has an interesting arrangement to support the ×10 gain function. 2N3702 silicon transistors, two per vertical amplifier channel, are creatively switched into the circuit to achieve it. Also, the power supply uses solid-state rectifiers and a single ACY22 germanium transistor to support a -12V supply. There are also numerous 1N914 silicon signal diodes in the circuit. The CRT’s EHT rectifiers were long stick multi-disc element selenium types; these parts gave trouble and required replacement. The dual-beam oscilloscope Most cathode ray tube (CRT) based dual-beam oscilloscopes actually use a single-beam CRT; the two (or more) beams are created electronically. They have a channel switching circuit that effectively creates a duplicate channel. The switching is either done on alternate horizontal traces or it is chopped between traces at a high frequency. The latter switches between two vertical amplifiers and two beam positioning controls to create the two traces. The typical scope, in two channel mode, has an ALT or a CHOP switch to select the method. In other words, all the heavy lifting to make the scope two or more channels is done by the scope’s electronics, not the CRT. The D52 is different. It has a real twin-beam CRT, but with one electron gun – see Fig.1. The gun is arranged with a beam splitter element, which splits one beam into two after it is emitted from the CRT’s cathode. There is siliconchip.com.au electronic circuitry. Normally, there would be a blanking amplifier for the task. The timebase It has speeds of 500, 200, 100, 50, 20, 10, 5, 2 & 1ms/cm and those numbers again at μs/cm. The horizontal amplifier’s user X gain control expands the trace to 10 screen diameters, and the shift control has enough range to allow any part of that expanded traced to be centred on the screen. This timebase was known for easy triggering. I have had no difficulty with it. Vertical amplifiers Fig.1: the Telequipment D52 uses a special cathode ray tube that splits the electron beam into two streams that are steered together horizontally (X1/X2) but differently in vertical directions (Y1’/Y2’ & Y1”/Y2”). also an adjustable magnet on the rear of the CRT socket that makes sure the split beams have equal intensities. The two separate beams go on to pass via different sets of Y deflection plates in the same tube. Only one set of X deflection plates is required to create the horizontal trace for both beams. The CRT is quite the masterpiece of electron optics; it also sported post-deflection acceleration. This allowed the tube to have relatively high sensitivity of the deflection plates, but also a high EHT, which favours high beam brightness. The CRT also has an inter-plate shield (IPS) electrode to reduce the interactions of the two Y sets of plates. The service manual omitted advice on how to set the IPS voltage. It is usually set to the average deflection plate voltage, which is 207V in the D52. One other interesting feature of the CRT is that, to achieve retrace blanking, they incorporated an additional control element into the CRT, called a modulation plate. This is nothing to do with X modulation, which is introduced into the CRT’s grid in the usual way. It is to fully cut off the beam cleanly for horizontal retrace. It appears amazingly effective. This is another feature of the particular CRT that eliminated more Australia's electronics magazine The vertical amplifier circuit, Fig.2, shows the arrangement with the original 2N3702 PNP transistors that are used for the ×10 gain boost circuit. When I first got the scope, the transistors in both channels were damaged. At the time, I didn’t have the exact parts. Ultimately, I replaced them with 2N3906s, which are better for the task (as explained below). The cathode currents of the ECC88 cathode follower V2A & V2B drive the transistor’s emitters. Since it is a differential amplifier, the transistor’s inter-base resistance (RV36) controls the gain. The output voltage is developed across the 8.2kW collector load resistors. When ×10 gain is not wanted, the transistor’s collector and emitter terminals are simply shorted out by the switch, and the cathode follower behaves as a standard voltage buffer. The circuit is the same for both channels, although some components are shared. Frequency-compensation networks are generally required in oscilloscope amplifier circuitry, either at the emitters or bases, to keep the response flat. This is because a combination of resistance and capacitance rolls off the high-frequency response. The arrangements to solve the problem (typically used in Tektronix scopes) are shown in Fig.3. However, the Telequipment D52 scope did not have any frequency compensation networks associated with the transistors in the ×10 gain function. Thus, the scope’s bandwidth was significantly limited in the ×10 gain mode. The D52’s vertical amplifier performance is very good in ×1 gain mode. The vertical input sensitivity November 2025  93 Fig.2: the vertical amplifier circuitry of the scope uses four ECC88 dual triodes (V1-V4) and two 2N3702 PNP silicon transistors (TR1 & TR2) per channel. The two silicon diodes (MR21/MR22) are shared between the channels. The transistors are responsible for the extra gain required in ×10 mode. is 0.1V/cm or 10mV/cm in ×10 gain mode, which is good for a scope of this age. More modern CRT scopes of the 1970s and 1980s went to 5mV/ cm and eventually to 2mV/cm (eg, the Tektronix 2465B). The trigger circuits also sported filters to help the scope lock on to TV frame or line sync pulses. There is a general assumption that the bandwidth specification is for the -3dB point. In at least three cases I know of, that is not even close. The D52 was rated for DC to 6MHz on the 0.1V/cm setting and DC to 1MHz in the ×10 gain or 10mV/cm mode. However, they underestimated it. I tested it using a Tektronix SG503 levelled sinewave generator terminated into 50W at the scope’s input on the 0.1V/cm setting. The vertical amplifier’s frequency 94 Silicon Chip response was flat to over 6MHz, and only 3dB down at 7.9MHz. In ×10 mode, it was flat to 1MHz and 3dB down at around 1.6MHz with the original 2N3702 transistors. In ×1 mode, with the transistors shorted out, the output impedances of the cathode followers of V2A and V2B a few hundred ohms. However, with the transistors switched in, in ×10 mode, the collector load resistance becomes 8.2kW. This, in conjunction with the transistor’s output capacitance (about 12pF) and the additional capacitance of the wiring and V3’s input capacitance, rolls of the HF response to 1-2MHz. The best PNP silicon transistor replacement I could find was the 2N3906, which has an output capacitance of only 4.5pF. With these transistors installed, the frequency response Australia's electronics magazine in ×10 mode substantially improves to be 3dB down at 3.17MHz. It probably would be possible to improve this further by adding a frequency compensation network, but I decided that I would leave the scope original, aside perhaps from the better transistors I had installed. I also checked the attenuators in the D52; they are excellent and properly frequency compensated, so they do not alter the vertical amplifier bandwidth on any setting. Self-cracking resistors Valves V2 and V4 both have 100W resistors in series with their control grids. These are known as ‘stopper resistors’. They form a low-pass filter in conjunction with the valves’ input capacitance, which prevents (stops) very high frequency instability, siliconchip.com.au Photo 4: with the leads being within the bodies of these resistors, when the leads corroded and expanded, the bodies cracked. Photo 5: these VMI 1N6519 rectifiers are rated at 10kV & 500mA. They are quite rare. Fig.3: these two compensation networks can be applied to differential amplifiers to extend their high-frequency response. They compensate for the inherent roll-off due to Miller capacitances and non-zero source impedances. especially in the VHF and UHF region. Similar resistors are used in the timebase section. These particular 100W resistors were all made by the same factory to the same design, and it was a disaster waiting to happen. One would imagine the failure rate of a resistor in this application to be extremely low because the current and power dissipation are negligible. I pulled the D52 scope out from a period in storage and, on powering it, there were multiple failures in both the timebase and vertical amplifier stages. Initially, I thought it would have to be a power supply problem, but it was not. After several tests, I noticed that some of the valves had very low plate and cathode currents. The readings appeared to make no sense. Then I started to discover that several of the siliconchip.com.au 100W stopper resistors in series with the control grids had gone completely open circuit. The control grids were floating, accumulating a negative charge and cutting off the valves. I removed six of these resistors to study them. The construction of the resistor was a cylindrical ceramic rod coated in a carbon film. There was a hole in each end in the ceramic rod with a metallised coating where the wire leads were soldered in. This is in contrast to the method where metal end caps are used. Corrosion in the holes had caused the leads to expand, cracking the resistor bodies. One resistor was cracked totally in half and only barely holding together (Photo 4). Metals oxides tend to occupy more volume than the metals they’re based on, so if they are encased in a rigid structure, the pressure slowly builds up over time. For example, rust (iron oxide) crystals expand under the paint on painted steel surfaces, causing the paint to bubble. It is a superior idea for a ceramic bodied resistor to have pressed-on end caps, but I suppose the creators of these Australia's electronics magazine resistors did not consider what could happen to them over the next 50 years. EHT failures The 2.6kV EHT for the CRT’s final anode is derived from a 1060V tap on the main power transformer. It feeds two capacitors and two diodes in a typical twice-peak voltage doubler. The two rectifiers in the EHT circuit were a type of long selenium stick rectifier in a cardboard tube. These are made up of multiple small discs stacked in series to create a rectifier with a high reverse breakdown voltage. The method does result in a relatively high forward resistance and a high forward voltage drop, but the CRT’s final anode current is very low. For example, the tube’s beam current is limited to 500μA. However, these stick selenium rectifiers failed and developed significant reverse leakage, overloading the 1060V transformer output. I replaced them with some excellent EHT rectifiers made by VMI (Voltage Multipliers Inc). VMI makes high-quality high-voltage rectifiers for many industrial and military applications. Occasionally, some November 2025  95 Photo 6: I wrapped fibreglass tape around the new capacitors to make them the same size as the originals. Photos 7 & 8: the new rectifiers and capacitors in place; and the recapped power supply board (below). turn up on eBay, presumably parts left over from an assembly contract. I managed to land a pair of 1N6519 rectifiers and had them in my parts box for a rainy day (Photo 5). The original stick rectifiers were rated at 3.4kV and 5mA, while the 1N6519 rectifiers are rated at 10kV and 500mA. They have a relatively fast recovery, suited to high-frequency supplies. In this case, that feature is not required. The new EHT rectifiers resulted in an increase in the EHT output from 2.6kV to 2.9kV, ie, about +11%. The total CRT EHT voltage is higher too, because the CRT’s cathode circuit is configured to run at -960V. While the CRT’s maximum beam current is limited to 500μA by the circuitry, the individual electrons, being accelerated by a higher voltage gradient between the cathode and final anode, acquire more energy before they hit the screen phosphor. Thus, the beam brightness increased even without a significant increase in cathode current. Some people fit a series resistor when replacing selenium rectifiers with silicon types, to lower the resulting voltage to near what the selenium rectifier gave before. In this case, I decided it was not required, and the improved performance was helpful. I also discovered that both the capacitors in the EHT voltage doubler section were electrically leaky. This had possibly provoked the failures of the selenium rectifiers. The main 96 Silicon Chip output filter capacitor appeared to be a large oil-filled type, rated at 0.05μF (50nF) & 3.5kV. The other coupling capacitor to the first rectifier is a lot smaller, rated at 0.05μF (50nF) & 2kV. The replacement capacitor I used was created from two 0.1μF 3kV capacitors in series to halve their capacitance and double their voltage rating. Balancing resistors are not required for film caps of the same value to share Australia's electronics magazine charge, as they have practically zero leakage. Due to the fact that the new capacitors have a smaller diameter than the originals, I wrapped them in 0.2mm-thick fibreglass sheet and finished them off with Scotch 27 fibreglass tape. The capacitor in the righthand side of Photo 6 is the original 50nF 3.5kV part, which was 36mm in diameter and 80mm long. siliconchip.com.au An advert from page 48 of Electronics Australia magazine, April 1969, showing multiple different Telequipment oscilloscopes for sale. These scopes were sold by Tektronix distributors in Australia. siliconchip.com.au Australia's electronics magazine November 2025  97 Photo 7 shows the two VMI rectifiers fitted and the two new capacitors in the voltage doubler. The electrolytic capacitors Most of the other capacitors in the scope were in good order, although some of the electrolytics in the power supply had started to draw excessive current and heat up. I replaced the defective ones, shown in blue in Fig.4. I went to a considerable amount of trouble to decide if C412, a three-­ section 32μF 450V capacitor (highlighted in green) should be replaced. After removing it, extensive testing of its capacity, leakage at its full rated voltage and its ESR were all perfectly normal, so I re-fitted it. I also could not find anything wrong with the main 120μF voltage doubler capacitors (highlighted in red). Summary The Telequipment D52 is a very nice vintage oscilloscope. It does have limitations compared to more modern CRT scopes; its bandwidth is not particularly wide, although better than the 6MHz advertised. The D52’s power supply system is non-regulated (that probably would have given the engineers at Tektronix bad dreams), so line voltage variations can affect the trace. The internal physical construction is good. One plus is that its unique twin-beam CRT does not have any problems associated with CHOP and ALT modes that can sometimes affect traditional twin-beam scopes. If you find one of these scopes and want to restore it, I would replace the selenium EHT stick rectifiers and EHT filter capacitors off the bat (if it still has the original parts), because when they fail, it stresses the main power transformer. Likely at least one or two of the electrolytic capacitors will require replacing. Also, it pays to check all the 100W grid stopper resistors in case they suffer from the self-cracking disease. When the cracks start, the resistor initially goes high in value, then after a while, it suddenly goes completely open circuit. It is probably worth replacing the original 2N3702 transistors with 2N3906s to improve the high-­frequency performance in ×10 gain mode. The scope is a great workshop asset, especially when fitted with a dual-phosphor tube, making it particularly good at examining long-­duration events. One application I put it to was to record the output of Sputnik-1’s Manipulator circuit, which switches at 2.5Hz, with characteristic steps in the waveform that correspond to the time when neither relay in the manipulator is closed. You can see a video of the scope displaying this waveform at https://youtu.be/k15GSKK_UY0 SC Fig.4: the scope’s power supply circuitry with parts of interest highlighted in different colours: the EHT voltage doubler in orange, main voltage doubler capacitors in red, a special three-section 32μF 450V capacitor in green and the faulty electrolytic capacitors that needed to be replaced in blue. 98 Silicon Chip Australia's electronics magazine siliconchip.com.au