Today we can build a 20 to 20 kHz sweepable oscillator using a single IC that fits in a small BAND·AID® box. The well known, though mostly obsolete, ICL8038, Exar 2206, and other monolithic circuits made this a trivial task a few years ago and today we have powerful digital synthesis chips smaller than your little fingernail. Back in 1952 when the 1304-A was introduced things weren't so easy. Every active device meant adding another tube, and almost all sweep arrangements were mechanical.
Covering a wide range was difficult, thus the BFO arrangement using a fixed 190/210 kHz oscillator mixed with a variable oscillator. Setability required a big dial with excellent calibration. If you wanted a log tuning characteristic, that had to be built into the tuning capacitor geometry. Output voltage and power had to be higher than today's generally puny requirements.
The 1304-B tunes from 20 to 20 kHz in a single sweep of it's 6" dial, and has a "plus 20" switch that adds 20 kHz to the range, giving a sweep from 20.020 kHz to 40 kHz. The maximum output is 50 VRMS open circuit and 1 Watt into 600 ohms. Four ranges and a meter are provided. Distortion is less than .25% midband. Built like a battleship, the 1304-B weighs in at 39 pounds.
I bought my 1304-B at a barn sale. It was owned by a aging college professor who was no longer doing research. It was in excellent condition and I expected it to work perfectly after removal from the rack where it appeared to have spent most of its life. Not!
Oscillator alignment was way off and it took me awhile to find a manual. This is one instrument where the manual is essential. There are many adjustments and without detailed instructions it would be almost impossible to get it working correctly. The distortion level was visibly high on the scope and I didn't know if this was the nature of a BFO, or if something was wrong. The specifications suggested that it should be much cleaner, but nothing seemed amiss in the circuitry.
I determined that the basic oscillator was very clean, and the problem was related to the output section. That made it quite easy to troubleshoot, as it was now just an ordinary audio amp problem- nothing exotic or related to the BFO sections.
After much head scratching, I realized that there was a tiny grid voltage on the first tube of the output stage. It should have been zero. This changed with the setting of the level control, which shouldn't be the case if the coupling capacitor was doing it's job. DC was getting to the grid, biasing the amplifier into distortion. Replacing the coupling cap brought the distortion well below the GR specification. My guess is that the previous user thought this was normal performance, or it just didn't matter for his application.
Remember that it doesn't take much leakage to cause a problem. In fact, the capacitor tested good in every regard until a high voltage leakage test was performed. It would have worked fine in most applications except this one. It was one of the infamous black plastic paper caps, so I replaced it with a modern polyester of the same value.
The 1304-B was often paired with the 1521-B Graphic Level Recorder to make frequency response measurements. The two were coupled with sprockets and a chain, allowing the frequency sweep to synchronize with the chart paper. There were also motorized dial drives that would automatically sweep the 1304-B in similar applications.
You'd think the 1304-B is so outclassed by modern signal generators that it would sit on the shelf, but this isn't the case. The extremely accurate dial, high output voltage capability, and fully isolated output (transformer coupled) make the device an excellent companion to my many bridges. It also has the desirable property of maintaining its waveform regardless of load. Because of the built in attenuator and single range convenience I often use it as a signal source for testing various audio components. I consider the "B" version a bit more desirable than the "A" version, both aesthetically and technically, but as with all GR products, either one is very capable. The 1304-A and 1304-B were the replacements for the prior 913-A introduced in 1942 and upgraded to the 913-C in my 1948 catalog. Any of these units will do a fine job of keeping your lab warm in the winter, especially if combined with a big old tube oscilloscope!
C. Hoffman
last edit August 28, 2010