Featured VT: The 6J5

 

This type is a singleton triode with an octal base. As nine pin mini types proliferated during the early 1950s, few singleton triodes were made, and the few that you do find are almost always VHF amp types intended for 400MHz+, running as grounded grid amps. Most triodes were made in pairs, such as the 12A*7 series, or the 6FQ7, 6DJ8, etc. If you need a singleton triode in the seven or nine pin mini format, you will either have to go with a type like the 6C4 (a type that is not recommended as an audio triode, but rather a low power, Class C RF driver/final) or make a pseudotriode from a small signal pentode. (The 6AU6 works nicely for this, and it has a μ-factor that fits nicely between the medium-μ triodes like the 12AU7 and the high-μ types like the 12AT7 or 12AX7).

The 6J5 appeared with the Octal base, and metal envelope. There is also a glass Octal as well, but is much harder to find, and more costly to acquire. Regardless of packaging, the type is a small signal, medium-μ, singleton triode. It is also basically half of a 6SN7. As such, it can be used for some of the same things: small signal voltage amps, QRP finals, cathode followers, and cathodyne phase splitters, and oscillators. The characteristics are also quite linear, making this type a good one for audio amplification where moderate levels of gain are required.

These are nice, linear plate curves indeed. It is easy find loadlines that exhibit very little harmonic distortion. The type has the linearity for small signal amplification, and the plate dissipation that allows for use as a cathode follower for driving moderate currents into difficult loads. The latter does include the grids of PP finals driven from a cathodyne phase splitter.

There is a bit of audiophoolery regarding the type. Some dislike the metal envelope, and have made accusations of all sorts against it. Yes, the metal cans do cost you a few glowey bottle kewlness points, and the paint offers little protection in circuits that use negative rail feed to the cathode, instead of/in addition to the positive DC rail. In such cases, it would be safer to opt for the all-glass version. Otherwise, I don't know what difference it could make, and the auto-shielding of the metal envelope is an advantage for the typically low level signals of small signal work. Unless you can specify a specific need, like the extra insulation or something. Otherwise, it's an unnecessary audiophool premium.

That does bring up one extra consideration: the cathode must never become more negative than the envelope. If it does, then the metal can can work like a second plate, attracting stray electrons. This can lead to, at the very least, increased cathode current and excessive bias. Various parasitic instabilities are also a possibility. The safest thing to do with the metal can is to connect its pin to the cathode pin. It's never a problem when using the type in the most common manner: cathode bias resistor, the plate supplied from the positive, DC rail only.

The other bit of audio mythology is that the 6J5 needs to pull at least 10mA to sound good. This just isn't true, and you seldom need that much plate current unless you're designing some sort of QRP application with it. The 6J5 was designed for low power applications, and the 6SN7 was often used in B & W TV sets as a combination plate coupled multivibrator/power amp for vertical deflection duty. The 6SN7GTB is a "hardened" version made as TV screens grew larger, and the demands on the deflection systems greater.

Here you see the loadline for low current operation. The estimated H₂ is virtually non-existent, and in practice this design worked as promised. There was no measurable distortion in evidence, the Twin-T test showed nothing above the noise floor, and subtractive testing showed nothing more than a pure sine wave arising from phase shift. No distortion of any importance.

The main draw back to such operation is that the low plate current drives up the plate resistance. This compromises high frequency performance, with the high frequency -3.0db point coming in at 45KHz. It is trivially easy to get bandwidths a decade larger with any small signal transistor operating at similar voltage gains. Whether or not that makes a difference depends on the application. It definitely isn't how you'd want to operate a 6J5 in a wideband,"DC to Daylight" application.

For the design where the 6J5 was used as a "current trickler", it worked very well indeed. If you need just such a VT, or you have an especially low voltage application, then you don't need to exclude the 6J5 (or the other 6SN7-oids) from consideration.