Fullscreen HTML5Med Res (??MB)HTML4

Vintage Radio Rebuilding the Kriesler 11-99 Never one to take the easy way out, when rebuilding this Kriesler 11-99, I decided to ‘upgrade’ it to 10-pin ‘decal’ TV-type valves. I then had a lot of fun resolving the stability problems that created! By Fred Lever T his Kriesler 11-99 is a smart-­ looking portable radio from 1968 with a modern-looking dial and single control knob. It has production number 3878 and ARTS&P number AA035355. The set is quite light at 3kg. The designs of the cabinet and circuit are efficient in material and component use, respectively. I purchased this example intending to refurbish it using new valves. As I had purchased a box of 10-pin decal TV-type valves and sockets for other projects, I decided to see if the original Kriesler 9-pin valve line up could be replaced with 10-pin TV types. If successful, I could use the same technique for future repairs. The original 11-99 used a series of 9-pin dual valves: a 6AN7 triode-­ hexode, a 6N8 dual-diode-pentode and a TV deflection valve, the 6GV8 triode-­pentode, for audio amplification. A 6V4 full-wave vacuum rectifier rounded out the lineup. That gave it the equivalent of a six-valve lineup – see the full circuit in Fig.1. I reckoned I could provide those functions using a 6V9 heptode-­triode to replace the 6AN7, a 6U9 pentode-­ triode in place of the 6N8, and a 6Y9 dual-pentode instead of the 6GV8. 94 Silicon Chip The 6V4 could be dispensed with and replaced with a silicon bridge rectifier. Being a next-generation series, the decal valves have similar or higher gains compared to the 9-pin types. So the performance of the set should be at least as good as the original. The 10-pin valves are designed to work at medium HT voltages of 200V, so the original HT power supply of around 150V looked adequate. Changing the valves I started by removing all the original valve sockets and small components, leaving large parts like the transformer, tuning gang and speaker in place. The 10-pin sockets bolted into the holes left by the 9-pin sockets. I then fitted a gland for the mains cord into the empty rectifier socket hole. Stage-by-stage, I rebuilt the set and fixed any problems at each stage before moving to the next. Fig.2 shows the changes that were required in red. To reduce the heat and increase efficiency, I deleted the valve rectifier and used a Jaycar ZR1320 silicon mini bridge rectifier. That gave about 150V DC at the first filter and 135V DC at the second filter with a draw of 20mA, the figure I wanted to limit the whole set to. A concern I had straight away was the exposed nature of both the mains and HT terminations and wiring. The Photos 1 & 2: the mains wiring was too exposed and a shock hazard. Australia's electronics magazine siliconchip.com.au Fig.1: the original Kriesler Circuit. Source: www.kevinchant.com/uploads/7/1/0/8/7108231/11-99.pdf safety level for servicing is very subpar in this type of set with a very shallow chassis. With the chassis out of the cabinet, making accidental finger contact with a live mains wire was way too easy – see Photos 1 & 2. I covered the whole of the power supply section and the volume control with insulated covers, seen in Photo 3. The audio section I wired up the audio section first, as shown in Photo 4. The 6Y9 is quite useful as an audio valve; there are a couple of watts possibly available, depending on the HT voltage and the suitability of the output transformer. It should reflect a load of around 10kW to the plate. The original output transformer had an open-circuit primary! That explained the blackened 6GV8 and tired 6V4. I fitted a Jaycar MM1900 line transformer to the speaker with the primary connected to the 0.5W tap. I was mindful of the high gain of the 6Y9 valve, but having used the type in two other projects, knew at least to start off with a grid stopper to combat self-oscillation. Photo 3: the mains wiring after fitting safety covers. siliconchip.com.au The valve is designed as a video amplifier, so it has a wide bandwidth. It was up to me to restrict the bandwidth using external components and limit the gain for stability. One step I took to reduce the total stage gain was to connect the preamplifier pentode section of the 6Y9 as a triode with grid bias. At first, I wired the 6Y9 output pentode with back bias. Upon turning the set on, I was greeted with screaming and popping noises laced with hum. To cut the story short, I simply had to ditch the back-bias and cathode Photo 4: I wired up the power supply and audio amplifier first. Australia's electronics magazine January 2026  95 self-bias the power section to be rid of the screaming and hum. Once cathode bias was applied, the crankiness went away, but the stage was drawing a lot of current. Sure enough, it was running as a sinewave oscillator at about 16kHz! Bridging the output control grid (pin 8) to ground did not stop the oscillation, so it was apparent the output pentode section was doing it all by itself, using the output transformer as a resonator. I stopped that by putting a 20nF 400V damping capacitor across the transformer’s primary. The two valves then amplified well, yielding a voltage gain of 2400 times from input grid to output plate. The overall gain of a 6GV8 is about 625 times, so the 6Y9 with 2400 times has plenty in hand to implement negative feedback if required. The power pentode drove about 4V into the speaker coil at clipping, indicating a couple of watts of output power. With that running well enough, I moved on to the detector and intermediate frequency (IF) stage. The IF section The 6U9 IF valve powered up in a grounded cathode configuration and immediately ran as a 470kHz sinewave oscillator, so I needed to bias it back up its gain slope to stabilise it and provide amplification. The simplest way of doing this was to self-bias the valve, get it under control, and let the automatic gain control (AGC) just do extra bias gain control in service. Even with that done, the stage was making hissing noises from the detector diode (the triode) and the 455kHz IFT coil resonance was limited. I had to manually cathode bias the valve back up its gain slope by increasing the bias to a silly value to get it stable. It became apparent that the base Scope 1: the initial IF system response looked pretty good. 96 Silicon Chip Photo 5: my original socket orientation had the wires crossing over, resulting in instability. Photo 6: rotating the socket 180° fixed the problem. wiring was involved in this, as moving the grid and plate leads around could ‘tune’ the instabilities in and out with squeals and popping noises! This was a lead dress issue, and all my own fault. I had mounted the socket without much thought to lead dress; the grid, diode load and plate leads were crossing over, making the valve unstable – see Photo 5. The yellow wire is the control grid drive to pin 3, the blue wire is the IF plate output from pin 7, and white is the diode load for pins 9 and 10. The leads were all too long, too close and crossing each other as the socket pin orientation was wrong. I fixed this by rotating the socket 180°, as seen in Photo 6. The yellow, blue and white wires are then short, direct and well away from each other, and the centre shielding ferrule can do its job. That was all it took to restore serenity. I left the IF stage with a low gain of about 40 times, with plenty of scope for reducing the cathode bias later to up the gain, and moved on to the front end. Scope 2: the initial oscillator grid (cyan) and plate (yellow) signals required some tweaking. Scope 3: the oscillator signals looked a lot better after adding a damping capacitor. Australia's electronics magazine Mixer and tuning coils I wired in the 6V9 as a classic grounded-cathode, grid-tuned biased-triode oscillator and a heptode tuned-grid mixer. Upon power-on, siliconchip.com.au Fig.2: my modified circuit. Besides swapping the values, most of the changes I had to make related to keeping the higherfrequency valves stable. there was no oscillator action. The tuned winding on the oscillator coil was open circuit! I stripped and rewound the coil, putting 100 turns on for the tuned winding and 30 turns for feedback. The oscillator then sprang into life, and the set received signals over the band with an indoor aerial, but there was obviously something not well as the IF output level was low, generating only -1V for the AGC signal. I performed a quick sweep check of the IF strip response to check the coils; that looked OK (Scope 1). Next, I checked the oscillator waveforms and amplitudes, looking for problems. The good news was that the triode tuned-grid amplitude was strong, at 30V peak-to-peak, and level from 1MHz to 2.5MHz (the cyan trace in Scope 2). The bad news was that the plate circuit was full of resonance (the yellow trace). I put a 220pF damping capacitor across the plate winding, and that got rid of a lot of the harmonics, or at least the higher ones (Scope 3). That siliconchip.com.au looked nicer, but it did not solve my IF problem. While scoping the 6V9 and 6U9 pins, I realised there was a large 100Hz component on the signal plates! That was not right. The HT ripple at the 6V9 and 6U9 supply point was about 0.4V, and this was getting into the signal streams and appearing at a level higher than the RF/IF signal! I fixed that in two ways. One, by separating the RF/IF stage’s HT from the audio HT with a low-pass filter comprising a 2.2kW series resistor and 10μF electrolytic capacitor to ground. Two, by providing a separate screen supply for each valve. Then the 6V9 and 6U9 were settled and stable, even when biased to a higher gain. Those two changes fixed all the stability problems the front end had. A rough check of the IF valve gain showed it was now around 60 times, depending on the AGC bias level, which now was regulating up to -6V depending on the received signal level. With the whole set working reasonably well, it was time to align the front end. Australia's electronics magazine The tuning range The dial tuning range is fixed by the reactance of the ferrite aerial rod winding and the gang. The only adjustment is the trimmer on the tuning gang. I checked the resonance range of the aerial coil (80 turns for tuning with a five-turn primary) and determined it varied from around 500kHz to 1700kHz. The stated tuning range of the 11-99 was 525kHz to 1635kHz. I took that as the figure and decided to tune the oscillator coil to give a range of 980kHz to 2080kHz (455kHz higher). Luckily, my 100-turn to 30-turn winding reached that by adjusting the one slug in the coil with the gang trimmer set at halfway. In theory, the specially shaped padderless gang should maintain 455kHz between the tuning coil and the oscillator coil. Even without an aerial wire attached, while tuning over the range, each local station came in at a good volume. Tuning over each station, I could scope a strong 455kHz resonance from the IF valve plate, and the detected audio January 2026  97 Scope 4: reducing the stability capacitor and some other tweaks made the audio sound much better. Scope 5: the oscillator tank signal. Scope 6: the mixer plate has the IF signal modulated by the audio waveform. pushed the volume control setting down below halfway. However, the speaker audio sounded muffled and unpleasant. There was around 30V peak-to-peak on the detector, so that was working well, clear of low-level diode knee distortion. The waveform on a sinewave-modulated test signal was clean, so the muffled sound was more likely due to a poor frequency response. With the front end working well, I moved back to the audio stage and investigated this further. I swept the audio response from the triode grid to the speaker coil over a range of 50Hz to 5kHz. The test showed the treble rolling off and too much bass for the tiny speaker. That rolled-off frequency response, or excessive bass, was the cause of the muffled sound. Applying a feedback loop from the voice coil to the 6Y9 triode’s cathode evened the response out a bit, while reducing the overall gain of the audio section. That was a good thing, as the volume control was working right near the start of its travel, especially with an aerial attached. I left the feedback loop and paid some more attention to the output transformer circuit. The 20nF value for the stability capacitor across the output transformer primary was the main culprit rolling off the response. The 6Y9 valve was always close to UHF instability, and would spill over at 16-30MHz if provoked, as well as oscillating at audio frequencies. That UHF instability was reduced by strapping a 220pF disc ceramic from the triode grid to ground at the valve socket end, and placing a 470pF mica capacitor across the grid feed point at the volume control end. The wire connecting the two points had to be shielded. The capacitors appear to be duplicates, but they do two different things. The disc is a shunt for UHF signals at the socket, while the mica capacitor acts as a roll-off for the IF filter at the other end of the shielded cable feed to the grid. This bypassing would be a bit unusual with normal superhet valves, but these decal valves are tiny internally and do not have the higher internal capacitances that would normally roll off the high-frequency response. I find you have to make circuits with decal valves the way IF strips are made in a TV. That is, bypass points with ceramic caps at the socket pins. Most of the resistors and capacitors in this Kriesler build are soldered right onto the valve sockets with short leads, to follow this practice. With the stage not bursting into oscillation above 1MHz, I then examined the effect of the plate roll-off capacitor on the frequency response. A few sweep tests indicated that the initial 20nF was overkill; reducing that to the final value of 6.8nF made the tonal balance much better. Scope 4 shows the final sweep response. All these changes perked up the mid-range and removed the muffled aspect of the sound from the speaker. The speaker is a small, 5-inch (127mm) type of average quality with not much baffling, so a sweep of its cone audio output would probably show most of the bass absent; the output would mostly contain mid-range frequencies. Scope 10: the audio signal at the AGC point is even cleaner. Scope 11: the audio preamp valve plate has picked up a bit of RF hash. Scope 12: the picked-up RF is gone by the time the signal reaches the power amplifier plate. Australia's electronics magazine siliconchip.com.au The audio response 98 Silicon Chip Final tweaking With those changes made, the set tone was more balanced and speech clear, so I then turned my attention back to the IF and mixer stages to get a bit more gain and better AGC. The most practical thing I could do was to carefully set the oscillator coil to shift the stations closer to the dial markings by using a deeper slug position. Scope 7: the grid of the IF valve has a much reduced IF signal component. Scope 8: the signal at the IF valve plate is significantly amplified. Scope 9: the detector diode output has a mostly clean audio signal. That had the effect of increasing the coupling of the primary and secondary coils, lifting the oscillator activity another 25%. I now had the set tuning local stations using the stick antenna alone, with the volume control between halfway and three-­ quarters, and around -3V on the AGC line. I think the oscillator coil could do with a rewind and closer coupling, but it is sufficient as-is. The tuning is very selective, mainly due to the IFT response being very sharp. I noted in the original 11-99 that they loaded the first IFT secondary with a damping resistor, presumably to broaden the skirts and widen the audio bandwidth. I tried that in my build, but there was no practical difference. I drove the antenna terminal with a 1000kHz signal amplitude modulated at 420Hz and scoped various points. In Scope 5, we see the tank resonance is not a pure sinewave. The harmonics in the primary are disturbing the tank resonance. I assume the triode is running into cutoff and saturation, but that is the best it can do. I could try re-winding the coil with fewer turns, but the effect is minor. In Scope 6, the mixer plate signal contains the 420Hz modulation and a mix of carrier and oscillator sinewaves. In Scope 7, at the IF valve grid, the bulk of the carrier station RF is rejected and a 455kHz ‘carrier’ bears the 420Hz audio signal. In Scope 8, at the IF plate, the 455kHz ‘carrier’ and modulation have been amplified by the IF valve to about 100V peak-to-peak and applied to the second IF transformer’s primary. That appears at the detector diode, on the secondary of IFT2, as shown in Scope 9. The positive swing of the IF signal has been shunted to ground, and a negative DC offset modulation signal exists. Most of the RF has been removed by the diode circuit’s filter capacitances. After the diode load resistor, the audio signal can be seen in Scope 10. Here, more RF is filtered out, leaving mostly audio-frequency components to the volume control. In Scope 11, the 6Y9 triode preamp plate has picked up a surprising amount of RF radiated hash, but by the time we get through the power valve, that has been lost and the voltage swing can be over 200V peakto-peak into the output transformer (Scope 12). Photo 7 shows the under-chassis component layout, while Photo 8 Photo 7: all the signal connections (via wires, resistors or capacitors) have been kept as short as possible to improve stability. Moving a wire or component by just a few millimetres can make all the difference! siliconchip.com.au Australia's electronics magazine January 2026  99 Photo 8: the top of the chassis has been left alone as much as possible, except for swapping over the valves. Shielding the valves didn’t seem to improve the stability, so I didn’t bother. shows the view from the top. The parts such as non-signal dropper resistors are mounted mainly on the centre tag strip, while signal-carrying parts are mounted directly on the valve base pins as much as possible. In Photo 7, the signal runs from the upper right-hand corner of the chassis down from the tuning coils, through the 6V9 mixer, IFT1 at lower right, then to the left across the chassis from IFT1 to the 6U9 IF valve, through IFT2, up to the volume control, back down to the 6Y9 output and out to the speaker. The tuning coils are unshielded, causing low-level whistles while tuning. I tried a shield can over the IF valve, and that helped there, but the problem is minor. Valve shield cans did not help with instability problems at any time, so I did not include any. Cosmetics The cabinet was originally all a bone colour, but it was warped and badly scuffed, with cracks and missing ventilation bars on the back. I just did a rough patch-up job with some epoxy resin to hide the worst of the damage. The front panel cleaned up nicely, so I left that and the dial in the original colours. The cabinet received a sanding and a couple of coats of “Go Go Blue” to give a ‘two-tone’ look, and it turned out fairly neat looking. Conclusion Photo 9: the chassis fits neatly into the restored cabinet. Photo 10: I used epoxy to fix the broken bars and fill in the cracks, then gave it a coat of paint. 100 Silicon Chip Australia's electronics magazine The set will play the local stations using the internal stick aerial. With an indoor wire attached, it will pick up all the available stations with good sensitivity and selectivity. The audio system is sufficient; the speaker level is loud, with good quality on speech and average quality on music. The AGC characteristic in still not perfect; a better-sorted system would hold the level constant no matter what aerial is connected. Still more work could be done on that. I have successfully demonstrated that the 6*9 decal series can be substituted for the original 9-pin type in a 1960s radio as long as a little care is taken to stabilise them. The article is a much-reduced version of a series posted in the special builds section of the “Vintage Radio” website hosted by Brad Leet: • https://vintage-radio.com.au/ docs/Kriesler-11-99-rebuild-111122. SC pdf siliconchip.com.au