Featured VT: The 6BQ7

This is a small signal dual triode in the nine pin mini format. From the spec sheet:

The 6BQ7A is a miniature, medium-μ, twin triode primarily designed for use as a cascode radio frequency amplifier in very high frequency TV tuners. In this application, its performance is characterized by low noise figures and high gain.

From the piccie, it's obvious that this is indeed a high frequency triode. It has the additional advantage of a high gm, made possible by the comparatively enormous cathodes. These are much larger than those of other small signal types like the 12AU7, 6FQ7, 6SN7, 6C4. The 6BQ7 also has a μ-Factor that falls nicely between types like the 12AU7 or 6FQ7, and the 12AT7.

The spec sheet even includes a composite plate characteristic for cascode operation. What you don't find is any mention of any sort of audio applications. The RCA Receiving Tube Manual does make a concession to this sort of usage, including the 6BQ7 in its example designs for RC coupled voltage amps.

There is something squirrelly about this plate characteristic. Is that an undocumented variable-μ characteristic? This makes it difficult to find a good, low distortion, audio loadline, but it is not impossible. The key is to get the V_PK up, and keep the voltage swing well away from the bottom of those curves. This will require either an unusually high rail voltage with passive plate loading, or active plate loading, then good audio performance may be obtained. The 6BQ7 is a good deal more linear than its VHF cascode cousins, types like the 6BK7B, which is hopeless as an audio amp. Whether it's as linear as the more common audio small signal triodes is another matter entirely.

So what good is it? The one thing that brought this type into consideration was the design of an audio cascode subsystem. In solid state practice, transistor cascodes are not that unusual. In hollow state practice, I came up with very little information. Nearly all references to hollow state cascodes were about VHF amps. For audio, it was all for guitar amps, and the emphasis was on voltage gain, not sonic performance. Of course, you like distortion in a guitar amp, as this makes for "tone". Still, there is the solid state practice, since cascoded transistors operate more linearly than do singleton transistors. Was it just that hollow state audio cascodes were "weird", and therefore not used so much?

The usual audio suspects didn't work out so swell. Cascoded 6SN7s didn't have enough gain, and 12AT7s showed a g_m rolloff with decreasing current that likewise made them unsuitable for this design. The RCA Manual mentioned the 6BQ7 as a cascode.

This was a case of try it and hear. I designed and built an LTP phase splitter made from two cascoded 6BQ7s. Even though this costs you half your voltage gain, there was sufficient to eliminate an additional gain stage while maintaining sensitivity even when gNFB is included. As to performance, it was quite excellent, producing excellent phase-to-phase balance with an active tail load, and undetectable harmonic distortion. Since the LTP phase splitter is also a differential, it provides a ready gNFB summing node. As an audio cascode LTP splitter/differential amp, the performance is excellent. A singleton 6BQ7 can also serve quite nicely as a triode LTP, especially if you include active tail loading.

The fat cathodes glow nicely as well.

Cascode LTP Example

You do have to watch out when using the 6BQ7A. Pay attention to how the heaters are connected. There are two ways this was done: an internal series connection wherein the heaters are wound with one continuous filament, and those where each heater is individually connected to the heater support pins. The series heater versions tend to be microphonic, and they ring like bells. The parallel heater versions aren't microphonic. That goes with the territory: series heaters are cheaper, and less attention to detail was paid. I suppose they figured that ringing at audio frequencies was not going to be a problem at VHF. The series heater versions are also more susceptible to filament burn-out. That unguarded length of heater filament is the first to heat up. It can flare brightly for a second or two before the rest of the filament has a chance to heat and develop enough resistance to limit the current. Cheap tubes and cheaply made tubes aren't good tubes. Going to ham meets to acquire a stock is well worth it, as you can examine before buying.

6BQ7-oids:

There are two other types with different heater voltages:

4BQ7A: 4.2V / 0.6A
5BQ7A: 5.6V / 0.45A

These types were designed for use in TV sets that didn't have a PTX, and used a series heater string to light the VTs. The currents were quite "standard" for series heater operation. Types with odd heater voltages might be more available and/or less expensive. If using with 6.3V heater supplies, just add enough series resistance, preferably split equally between both legs to maintain balance, to drop the voltage to the rated heater voltage.

One last consideration is that the type includes an internal baffle shield between triode sections. This shield needs to remain negative to any cathode inside, or it could start to function as another plate. That could possibly throw off the bias, or lead to instabilities. It may be a problem with some LTP applications if the cathodes go negative.

The 6BQ7 isn't an easy VT to use in audio work, but when you need more gain than a 6FQ7 can provide, but not so much as a 12AX7 or 12AT7, then it definitely meets the design criterion. The cascode LTP splitter/gain stage is pretty much my universal, "go to" front end design. It provides the gain that you would get from a cascade of a 6SL7 LTP DC coupled to a 6SN7 differential stage, and with a greatly reduced input capacitance that would otherwise interact poorly with a high resistance volume control or long cable runs.