Taylor model 67A Signal Generator

On the quarterly radio flea-market of the Dutch association for historical radios I bought this signal generator. It looked sufficiently old to contain some nice valves, but modern enough to still function. I wanted an RF generator to trim one or two radios. And for the fun of it. I did have plans to build a generator myself but hadn't yet found the time. To build a good generator the mechanical work, such as the frame, the tuning mechanism and the tuning scale, is at least as important as the schematic. And good mechanical work takes time.

By the way, if you want to build your own RF generator using transistors, there is a schematic and description on the Vintage Radio Repair and Restoration site.

Before I was going to try the generator, I wanted to check whether it was complete and safe. And, of course, I wanted to know how it worked. But I had no service documentation with it and could not find any on the Internet.

By following the wiring, I was able to reverse-engineer it and draw a schematic diagram. The generator is fairly simple. It has only two valves. The generator circuit itself is a classical Collpits oscillator around an EF91 penthode (actually I'm not sure about the valve. The lettering is partly illegible, only "EF.." is left, the "9" is obvious because of its 7-pin miniature base, but this leaves 8 penthodes in the "EF9x" range in my handbook. Looking at the pinout and electrode configuration I am fairly confident, but I may be wrong. Does anybody have a copy of the original manual for me?). There is a double tuning capacitor to produce the Collpits' tuned circuit with a tapped capacitor. The 6 frequency ranges are selected by switching the inductors. Amplitude modulation is achieved by varying the plate voltage to the generator. The modulation signal comes from a simple oscillator built up around an EC91 triode. This signal is available on an AF output that can be used to trigger a scope.

Apparently, the designer considered the valves as the most expensive part. So he used only two of them. On the other hand, the mechanical aspects have been done quite solidly. The generator is contained in a shielded subassembly. The attenuator switch works 20dB steps, each step is realised by a separate attanuator section that is shielded from the other sections to prevent crosstalk. To realise this, a 5-deck rotary switch has been used, whereas a cheap attenuator could have been realised with a single deck. The whole attenuator is housed in a shielded box. The tuning capacitor is driven by a gear mechanism to make frequency adjustment easy. On the photograph on the right you can see the inductors and trimming capacitors through the removable trimming hatch on the back of the oscillator case.

I had to fix some damage before I tried the generator. Two of the inductors seemed to have had a blow and one soldering joint had come loose. One resistor in the attenuator was broken, so only the 0dB position worked. That was easily fixed. And some screws were missing, but that was not a real problem.

The generator worked right away. On most of the ranges, it produces an output signal of 0.5-1 V_tt. Except on the highest range (30-240 MHz) where the signal is smaller. On the low end of this range, the generator even stopped. Understandable, because the tuning capacitor is around 500 pF, which is too large for these frequencies. Besides, 240 MHz is largely exaggerated. On my (100 MHz)scope it turned out to produce only 120 MHz.

Although many generators use an output amplifier to isolate the oscillator from the load, this one doesn't. This makes it rather sensitive to the load on the output. The output is connected to a coupling winding on each inductor. Using the generator on the 10-40 MHz range, connecting my frequency counter to the output, I observed a noticed a significant decrease in frequency when I added my scope to the output.

Measured on the reservoir capacitor, the supply voltage was only 80 V, produced from 100 V AC from the power transformer. For the generator itself, a meagre 50 V was left. The load on the power supply is only a few mA. This indicates a bad metal rectifier, so I braced it with a silicon rectifier and a 470 Ω resistor. Now I have 138 V on the reservoir capacitor and 90 V on the generator. After this, the generator worked much better. It even worked over the whole span of highest frequency range.

Modulation turned out to be poor. The oscillator already has some distortion, but the modulated signal looks even more distorted. That is not a surpise, given the simple modulation principle used. Modulation depth cannot be controlled. On most ranges it is only 20% or so. On the highest range, however, modulation is stronger and on the lower half of the range the generator is completely cut off during part of the negative half-cycle of the modulation signal. That behaviour is consistent with the observation that the generator worked marginally on a low supply voltage.

Well, now I have a usable RF generator for a number of trimming jobs. As a finishing touch, I replaced the 2-wire mains cord with a 3-lead cord to have a safety earth. I also thought of replacing the fixed output cable with a BNC connector, but didn't. It didn't seem worth the trouble. Besides, the output seems to be intended as 75 Ω, which does not go well with the standard 50 Ω measuring cables I have.