Fullscreen HTML5Med Res (??MB)HTML4

Vintage Radio Kriesler’s Kriesler’s 41-21 41-21 mantel/portable mantel/portable set set By Ian Batty Electrically, the set is also somewhat interesting. It uses a reflexed second intermediate frequency (IF) amplifier, with that transistor also acting as an audio preamp. The design is similar to the Philips MT4 that I described in September 2017 (siliconchip.com.au/ Article/10806). Like that set, the reflexed stage needs carefully-managed signal levels, so the 41-21 has a two-gang volume control potentiometer. More on that later. So despite its dial drive, it’s an Australian set worth an article. First appearances The Kriesler "Triplex" 41-21 is an all-transistor, battery-powered radio which uses reflexing. It was produced in the late 50s/early 60s and was sold with a plastic case that came in one of three colours (pink, brown or red). Engineers are a chummy lot. During my Air Force days, I encountered a variety of engineering types needed to keep an aircraft flying: mechanical engineers for the engines, airframes and controls, electrical engineers for the electrical systems and controls, electronics engineers for the radio, radar navigation and instrument systems, and commerce types for supplying all the parts needed. But when I took a look at the Kriesler 41-21 (manufactured from 1959 to 1961), I started wondering whether the 100 Silicon Chip mechanical engineering folks at Kreisler were ‘in dispute’ with the electronics engineer cohort. Surely no-one could have come up with the labyrinthine dial drive in this otherwise fine set unless they had some axe to grind. Yes, I get that it’s a way of accommodating a 130mm long dial with a 41mm diameter drum on the tuning gang, when an 82mm diameter drum would otherwise be needed. But I would have put in a 2:1 gear set to the drum and simplified the rest of the arrangement. Australia’s electronics magazine The curved, rippled front with its coloured inset and black case rear is a pleasing alternative to the “square black box” so often resorted to in the late 50s/early 60s. The “slide rule” dial is some 130mm long; plenty of space to list all the stations of the day. The side-mounted volume control is placed for easy adjustment. The separate on/off switch eases the load on the volume control; ie, it doesn’t need to be rotated every time you turn the set on or off, giving a longer trouble-free life. Circuit details The set’s circuit is shown in Fig.1. The main difference between the 4121 and the identically-cased 41-21A is the 21A’s use of a single-tuned third IF transformer. All transistors are Philips/Mullard “OC” series germanium PNPs, with a negative power supply (ie, positive ground). Ferrite rod L2 is tuned by the antenna section of the tuning gang, C3A. A low-impedance secondary matches to the base of the converter via capacitor C2, in parallel with 2.2kW resistor R1 (the bottom half of the converter’s bias divider). C2 is there to overcome the resistance of R1 at radio frequencies siliconchip.com.au and deliver full signal to the converter. The ferrite rod also provides a primary for connection to an external antenna and Earth. L1 (100µH) helps to match the capacitance of an electrically-short wire antenna to L2’s tuned secondary. Kriesler’s original circuit correctly describes L1 as a compensating coil. It isn’t there for interference suppression. L1 is wound using its parallel 3.9kW resistor as a former. This resistor dampens the L1/antenna resonance. Converter TR1, an OC44, uses selfexcitation and emitter injection with feedback from collector to emitter via oscillator coil L3 and 10nF capacitor C4. This allows signal injection directly to the base for testing. As usual, the base-emitter bias voltage is lower than you’d expect, as the converter needs to operate closer to Class-B than the normal Class-A used in the IF and first audio stages. Class-B operation allows the converter to go into cutoff over part of the local oscillator’s (LO) waveform, ensuring the non-linearity vital to converter action. You may wonder at TR1’s emitter and base voltages of 1.15V and 0.75V, with the emitter higher than the base. This makes it seem as though TR1’s base-emitter is reverse-biased and would never conduct. This would be the case for a Class-A amplifier, but TR1 works as an oscillator, and I measured a signal of around 6V peakto-peak at the emitter. TR1 repeatedly swings in and out of conduction due to the oscillator’s excursions, creating the non-linearity needed for conversion. I measured values of 0.75V and 0.94V respectively when the LO was stopped. For this set, LO operation can be confirmed by circuit measurement. It’s not a wholly reliable test though, and my preference is always to use the radiation test first. Tuning The 41-21’s tuning gang uses identical sections, so 480pF padder capacitor C6 restricts the LO’s frequency swing and ensures that it’s kept 455kHz above the incoming signal. The converter’s 455kHz signal is developed across the tuned, tapped primary of first IF transformer IFT1. Its tapped, tuned secondary feeds first IF amplifier TR2, an OC45. TR2’s significant collector-base capacitance demands neutralisation, siliconchip.com.au The Kriesler 41-21 with the case open shows the double-sided PCB and 5-inch Magnavox speaker. The in-built ferrite rod is hidden behind a cover at the top of the case. and this is done on the circuit board, with traces from the collector and base passing by each other. There’s no actual connection, but their proximity is engineered to provide 4pF of capacitance (C12). Neat. (See photo Fig.2). As usual with first IF amplifiers, TR2’s upper bias resistor, R5, is high in value at 150kW. This allows the AGC voltage developed by demodulator diode D2 to be fed back via 10kW resistor R7 to reduce TR2’s collector current with increasing signal strength, thus reducing its gain. Stage bypassing (C11, C14) is di- rectly back to the emitter rather than to ground, saving on emitter resistor R8’s customary bypass capacitor and giving improved bypassing. Extended AGC action 2.7kW dropping resistor R9 works in combination with R6 and D1 to provide extended AGC action. With no AGC applied, TR2’s collector voltage is around 6.1V. Although OA70 diode D1 and its series 3.9kW resistor R6 connect to the input of the first IF transformer, they have no effect with weak signals as D1 is reverse-biased. A close-up of the dial and the latch for the case. Australia’s electronics magazine March 2021  101 As signal strength increases and TR2’s DC collector voltage rises towards 6.8V, D1’s cathode becomes more negative, and it eventually comes into conduction. At this point, the signal at the first IF transformer’s input is partly shunted to AC ground, reducing the converter stage gain. This extends the AGC’s control range from the approximate 30dB increase in signal input achieved with AGC on the first IF amplifier alone, to as much as 60dB. TR2 feeds the tuned, tapped primary of second IF transformer IFT2. Its untuned, untapped secondary feeds the base of the second IF amplifier, TR3 (OC44). Why use the premium OC44 where you’d expect to find the lower-spec OC45? The answer is gain. The OC45’s hFE is 50~125, while the OC44 offers an improved range of 100~225. This should be advantageous to the audio function of this reflexed stage. Reflexing As mentioned earlier, the stage around TR3 is reflexed, amplifying both the 455kHz IF signal and the demodulated audio signals. The IF section follows common design practice. Like the first IF amplifier, this stage employs printed circuit tracks to provide neutralising capacitance (C16, 3pF). TR3 feeds the tuned, tapped primary of third IF transformer IFT3 and its tuned, tapped secondary feeds OA79 demodulator diode D2. The 41-21A set uses a single-tuned transformer (tuned, tapped primary, untuned, untapped secondary) for IFT3. D2’s output is applied, via R19, to the top of volume control R16’s first section. Confusingly, it’s labelled R16B. The DC component is fed, as the AGC voltage, via R7 to AGC filter capacitor C9 and then to first IF amplifier TR2’s base. R16B’s wiper feeds audio, via C20 and R13, to the base of the reflexed IF amplifier, TR3. Now, as an audio amplifier, TR3’s emitter needs to be bypassed for audio by 33µF capacitor C17 (in the original schematic this was 32µF). So why use a dual-gang volume pot? TR3 has a difficult job: it must amplify millivolt-level IF signals and much higher level audio signals without interaction. Recalling that valve reflexes were bedevilled by cross-modulation and 102 Silicon Chip Fig.1: The original Kriesler 41-21 schematic shows capacitors C21/23 in reverse polarity, and neutralising capacitor C16 should connect directly to the base of TR3, both have been fixed here. On some sets R12 is not fitted; if IF regeneration occurs, it's best to fit this R12 as shown. Similarly, an extra OA79 diode was fitted across the oscillator coil (L3), with its cathode to the collector. minimum volume problems, Kriesler’s designers have restricted the maximum possible IF signal (via the AGC system) and audio signal (by R16B) to ensure TR3’s correct operation at audio and IF signal frequencies. Audio stages Amplified audio is developed across 1.5kW collector load R15, and fed via C21 to the second section of the volume control, R16A. Audio stage gain is around 5.5 times, which might seem poor. But it’s in line with other similar circuits: the Bush TR82C’s first audio stage (TR4 on that circuit) delivers a gain of just 5.0 times. R16A’s moving contact feeds audio, via C23, to the base of audio driver TR4. This is an OC75, a higher-performing version of the OC70/71 types with a higher hFE (current gain) of 90~130 compared to 20~40 and 30~75 respectively. The manufacturer’s diagram for this set has the symbols for C21 and C23 mistakenly reversed. My redrawn diagram fixes this. TR4 drives the primary of phasesplitter transformer T1, with its secondary matching anti-phase signals into the low base impedances of the two output transistors, TR5 and TR6. TR4 gets audio feedback from the speaker via R32 (47kW), while R24 (560W) and C29 (22nF) apply top-cut. In common with transformer-coupled Australia’s electronics magazine stages, TR4 delivers a volt of signal into T1’s high-impedance primary for a stage gain of around 50. As T1 is a step-down transformer, the signal applied to the bases of TR5 and TR6 is considerably lower. TR5/6, both OC74s, operate in ClassB, with bias provided by the divider R25-27. R27, a CZ9A thermistor, acts to reduce the applied bias at higher temperatures, compensating for the natural fall in base-emitter voltage needed for a particular collector current as transistor junction temperature rises. 10W emitter resistors R29/R30 help equalise gains between TR5 and TR6, as well as providing some local negative feedback. The output transistor collectors drive output transformer T2, which matches them to the speaker. There’s another top-cut network across its primary, comprising 100nF capacitor C31 and 330W resistor R31. Disaster awaits The manufacturer’s diagram shows the output stage’s bias divider with a single adjustable resistor between the decoupled battery supply (at C28) and the output bases. What if you accidentally set this resistor to its minimum value? You’ll be attempting to apply many volts to the output bases. Expect them to draw massive collector current and possibly to suffer overheating and destruction. siliconchip.com.au I have a suggested modification below to solve this. Maybe the bloke who designed the dial drive also did this part of the circuit. Clean-up My sample was in good physical condition, with no cracks in the case. It just needed a bit of polish to bring it back to a reasonable condition. Mechanically, though, it had a broken/missing dial cord. Cue Lalo Shifrin music: “Your mission, should you choose to accept it...” In addition to the dial problem, I found it extremely noisy with the volume control wound up; less so at low/ zero volume. Contact cleaner on the volume control helped a bit, but I eventually traced the fault to capacitor C8. This 50nF green ceramic capacitor was acting like an erratic partial short circuit. Converter TR1’s collector voltage would crash down by as much as a volt, then recover, then drop by maybe half a volt, and so on. Leaky caps usually soak up a fairly constant amount of current; this was the first that I’ve seen like this. I thought of replacing it and all the others with greencaps to eliminate possible future recurrences. But that dial drive was lurking in the background, and I was wondering how I could make up those drive pulleys. About this time, I attended the HRSA RadioFest in Canberra. I was griping about this set when another member said he might have one among siliconchip.com.au the boxes of transistor sets he was getting rid of (he was ‘downsizing’). Bingo! It was the 41-21 version (double-tuned third IF), but otherwise identical, and with a functional dial drive. A simple cabinet swap gave me the set in this article: a good cabinet with a working dial mechanism. It was the classic case of “collect two, get one good”. The only bother was the original wire trimmer, which insisted on tuning to above 1700kHz. It’s easy to remove the tinned wire from the ceramic former but harder to add to it. I popped a Philips “beehive” into its place. That dial drive mechanism Kevin Chant’s website has the dial cord diagrams (www.kevinchant.com/ kriesler2.html). It has three assemblies: the securing loop (top), the pulley cord driving the gang’s drum, and the station scale cord (bottom). I’ll leave you to download it and try to work out how to fix it if your set has a broken dial cord. The one I started on had nothing but the dial drum, pointer and driveshaft remaining. You also need two floating pulleys to complete the job. How good is it? It’s good without being great. The reflexed audio stage helps it produce 50mW output for 150µV/m at 600kHz or 120µV/m at 1400kHz, with noise figures of 11dB and 12dB respectively. Fig.2: TR2 needs neutralisation, which is done on the circuit board via parallel traces from the collector and base of TR2. This provides approximately 4pF of capacitance. Australia’s electronics magazine March 2021  103 For the standard signal-to-noise ratio of 20dB, the required signal figures are 370µV/m at 600kHz and 225µV/m at 1400kHz. RF bandwidth is ±0.95kHz (-3dB) or ±23.7kHz (-60dB). AGC works reasonably well, with a 35dB signal increase giving a +6dB rise in output. Its audio response is 95~1200Hz from the antenna to speaker and 170~7000Hz from volume control to speaker, with a 2dB rise around 1kHz. Maximum output is around 130mW for 10% total harmonic distortion (THD). At 50mW, THD is 4.2%; at 10mW it’s 2.5%. At half battery, the maximum audio output is 25mW at clipping, and 20mW output gives 6% THD. 41-21 versions As noted above, the significant change from the 41-21 to the 41-21A was the substitution of single-tuned third IF transformer IFT3. There was one minor change: IFT3 retained a tuned, taped primary, but was fitted with an untuned, untapped secondary, simplifying the circuit and making alignment easier. The service manual also hinted at a 41-21B version which used a new dial drive mechanism, although no other information could be found on whether this set ended up being manufactured. Special handling Be very careful when adjusting the output stage bias. As noted above, the design contains a potentially catastrophic mistake: with only R25 in the “hot” end of the output stage’s bias divider, it’s possible to apply almost the full 9V to the bases of TR5/TR6. I have modified the review set with a 3.3kW series resistor. This allows plenty of adjustment without the danger of frying the output transistors. Conclusion I like the way this set looks, and it has good performance with just enough circuit quirks to make it interesting, without baffling us poor electronics engineers. While I would not buy one with the dial cord apparatus missing, “your mileage may vary”. Hopefully, the de104 Silicon Chip Even though they've used a double-sided PCB, the radio still has an ample amount of wiring, along with a number of unused holes. scription above will be of use if you do take the plunge. Thanks to Jim Greig of the HRSA for the loan of his set, and Charles McLurcan (also of the HRSA) for a set with the dial cord assembly intact. Not a member of the HRSA? Go to: http://hrsa1. com to see how we can help you with our exciting radio hobby. Australia’s electronics magazine Further Reading For the circuit and service notes, see Kevin Chant’s fine website: www. kevinchant.com The service notes contain the cording diagram with dimensions. This model in particular can be found at: www.kevinchant. com/uploads/7/1/0/8/7108231/41-21. pdf SC siliconchip.com.au