What about the KT120?

We are occasionally asked about the KT120 tube. Most of the inquiries make an observation about some competitor adopting the tube and making claims for its virtues. In many cases these same manufacturers did much the same previously for the Russian made Tung-Sol 6550 and the Russian made Genalex KT88.

Generally speaking, the KT120 is an instance of Eastern European & Russian tube makers assuming that scaling up the internal structure of a tube makes everything better. However, this is not unambiguously true with vacuum tubes, so let’s get out the KT120 datasheet and look at the results, both pro and con (or, if you just want the bottom line, your KT88 based VAC amplifier should only be used with tubes that conform to the KT88 specification).

On the negative side:

1) Stray capacitances are increased by the larger structures, in this case by 50% to 70% for tetrode operation. The KT120 datasheet does not give the numbers for triode connection, so no conclusions can be drawn there, and that typically is more critical. Increased stray capacity reduces high frequency bandwidth. In amplifiers that have marginal stability this conceivably could lead to parasitic oscillations. It’s off the KT88 spec, and would have been rejected if it showed up in a production batch at the old British GEC/Genalex/M-O Valve group.

2) The heater current draw is up by as much as 22%. This puts a significant additional load on the power supply, and could lead to reduced heater voltage with attendant sonic degradation, or in some cases cause the power transformer to overheat and fail. Again, off spec for a KT88.

3) The data sheet shows transconductance about 9% high. Ordinarily not a significant concern, but it does relate to another problem in the characteristic curves that we’ll get to in a minute.

4) The KT120 data sheet is showing a surprisingly high 14% THD figure for typical operation. That’s much too high, and one might speculate about a typographical error, if not for the characteristic curves… we’ll get to that in a minute, too.

5) The maximum allowable grid resistances quoted for the KT120 are much lower than for the KT88, representing loads on the driver tubes that might be 1.12 x to 4.3x heavier than with a KT88. The typical reason for limiting the maximum grid resistance value is due to gas currents in the tube. So, this number might suggest that the manufacturer is concerned about gassy tubes.

6) Examining the plate family characteristic curves for the KT120, the first thing that emerges is the extremely sharp transition between the vertical and horizontal sections of the curves; the ‘knee’ is very sharp. The KT120 shows this behavior at much lower power levels than does a good KT88, and this may lead to more complex distortion with reactive loads (i.e., loudspeakers).

The second thing that stands out is the area of ambiguity that develops in very low output levels extends over a much larger range with the KT120 than with a good KT88. Further, this misbehavior continues to twice the power output level than it does with a good KT88.

Next, the spacing between the stepped grid potential curves is much less even than with a good KT88. Ideally, when you increase the signal on the grid by a unit of voltage, you want to see the same percentage increase in output current regardless of your starting point. In this regard, consider 5 volt grid increments, starting at -5 v (near max power), and then going down. With a good KT88, the intervals are 100%, 100%, 69%, and 56% as you head toward cutoff. With the KT120, it is 100%, 78%, 56%, and 39% … which is much less linear, and tends to confirm the surprisingly high THD figure quoted in its data tables.

Lastly, due to the behavior just discussed, the tube does not cut off cleanly at the expected point; it’s about twice the proper spec point for a KT88.

Now, positive points:

1) In triode mode, the max plate/screen voltage rating is up 50 volts over the KT88 to 650 volts. In the tetrode mode, the plate rating is up 50 volts to 850 volts, but the screen rating is unchanged at 600 volts (thus the effective rating for normal ultralinear techniques is still 600 volts, just as with a normal KT88).

2) The anode dissipation rating is up 43% over a standard KT88, so you could push more idle current through it, and potentially get more output power if you also raise the plate voltage (and thus make the amplifier incompatible with the KT88 or 6550). However, the screen dissipation rating is unchanged at 8 watts, so there is more to consider in practical design calculations than first meets the eye, and it might be difficult to exploit this.

3) The KT120 ratings showing an 8% increase in maximum cathode current, but for the tetrode mode only … this makes no sense at first blush, and so may be a typo. Either way, the 230 mA KT88 spec and triode KT120 spec is more than adequate.

The practical results:

1) Plugging a KT120 in place of a KT88 will not give you significantly more output power any more than switching one KT88 for another might. Keep in mind, though, that even with the same production run of tubes, one pair might yield 95 watts and another pair 118 watts; that’s just part of the normal manufacturing batch variation. Maximum power output generally corresponds to how ‘hot’ the tube wants to idle. The more negative the required grid bias voltage, the greater the possible drive signal swing before the onset of grid current, i.e., more watts out. However, hot tubes are not necessarily the best sounding, most stable, or most linear.

2) More power output on a consistent basis generally will require some redesign of the amplifier, which may in turn make it incompatible with KT88s or 6550s..

3) You should expect more distortion with the KT120, possibly particularly at moderately low listening levels and with highly reactive loudspeakers.

4) The situation is much as it was with the first generation EiRC KT90 (Mk I) back in 1990. You could redesign an amp to give more power, but the tube didn’t really conform to KT88 specs and had higher distortion.

5) If you use the KT120 in amplifiers designed for the KT88 or 6550, on rare occasions it may fry a power transformer.

6) You’re likely to give up a few kHz of high frequency response.

7) It might last longer in very hard service if you leave the settings as they would be for a KT88.

8) An amplifier developed to exploit the KT120 fully would not be compatible with the KT88. Since the KT120 is only made by one factory and has very little history, there might be difficulties with retubing such an amplifier in the future.


At the end of the day, it’s always fun for marketers to advertise a magical new ingredient, so it’s understandable that the KT120 has made quite a splash. Many manufacturers using it probably are hoping that it will be a more reliable power tube, and the comparison is clouded by the fact that many of the currently manufactured versions of the KT88 aren’t done as well as they should be, so this is not a knock against other manufacturers giving them a try. However, do not expect the KT120 to be a panacea.

Earlier I noted that there has been a progression with some other amp manufacturers from the Russian TungSol 6550, then to the Russian Genalex KT88, and now to the Russian KT120. At each change, the new tube is hailed briefly, and then thrown aside. At the introduction of each of these tubes we have tested them and found them to be wanting with respect to reliability and sound quality in comparison to our normal KT88; thus, we have not adopted them and do not recommend them.

VAC does not approve use the KT120 for VAC instruments. Our KT88 based amplifiers should only be used with tubes that conform to the KT88 specification.