Please click your language flag. Bitte Sprachflagge klicken.

The optimal use of the 6AF6G

Jacob Roschy Martin Renz Ernst Erb Bernhard Nagel Dietmar Rudolph Eilert Menke 
Please click the blue info button to read more about this page.
Forum » Valves / tubes - Semiconductors » VALVES/TUBES / SEMICONDUCTORS in RADIOMUSEUM » The optimal use of the 6AF6G
Jacob Roschy
Jacob Roschy
D  Articles: 1715
Schem.: 17
Pict.: 68
20.Oct.11 21:03

Count of Thanks: 16
Reply  |  You aren't logged in. (Guest)   1

In the late 1930s magic eye tubes were developed in Europe with two shadow angles, the EM4 and the EM34, or with four shadow angles, the EM11 and the early version of the 6AF7, which were controlled by integrated triode units with different plate sensitivities in order to handle a wide range of input voltages. The more sensitive shadow angle reaches full closing even at moderate input signals, while the less sensitive shadow angle closes only at strong signals.

However these two-range magic eye tubes have some disadvantages:

  • both shadow angles have a maximum opening of 90°, thus a big part of the fluorescent screen remains unused for the actual signal display, as it is lit without any change. This deficiency could be partially compensated by the tubes having 4 shadow angles.

  • The less sensitive section moves from the beginning in parallel with the more sensitive section, so that a substantial portion of the less sensitive section is wasted until the more sensitive section is at full scale. Only the remaining part of the display area is actually useful after the sensitive section is fully closed.

An ideal dual-range magic eye, in which the luminescent screen is completely used for the display, would have to have the following properties :

  • without input signal both sectors show no display, i.e. the fluorescent screen is completely dark with only two narrow light stripes as zero signal marks.
  • Weak signals are only displayed on the first sector. The second sector starts to display, after the first sector is fully closed.
  • Both sectors move from 0° to 180° so that the entire surface of the luminescent target is involved in the signal display.

Actually there is no tube available with these properties, but with some effort, one can come relatively close to these demands by using the 6AF6G eye tube.
The 6AF6 has two shadow angles similar to the European tubes EM4 and EM34, but it
contains no built-in amplifier units, which always makes a separate control tube necessary. Since this allows complete freedom of circuitry, the 6AF6 can be configured with display properties, which are not possible with any other magic eye tubes.

The shadow angle of the EM4 and the EM34 is 90° at maximum because,
even with zero input voltage, the positive voltage of the control rods is still relatively high with respect to the cathode. By using a 6AF6, this problem can be solved, if its cathode is biased at a higher voltage against ground (GND) up to +75 V with a voltage regulator tube 0A3 or 75C1.

With a proper double triode as control tube, such as a 6C8, it is now possible to pull the control rods to to a negative voltage against the cathode, so that both the shadow angles become
very wide and only a narrow strip is left lit with zero input signal.



A positive current bias is applied via a 100 MΩ resistor to the G II grid the of 6C8's second triode unit, while this G II grid is also connected to the control voltage U'in via a 5 MΩ trimpot. The G I grid of the first triode unit is directly connected to the control voltage U-in.

As long as less current flows through trimpot P2 than through resistor R3, the voltage at grid G II remains at 0 V while
the negative control voltage moves freely due to the grid diode clamping effect. The voltage at G II starts to become negative only at the point where the current through P2 equals the current through R3.

That means, the voltage at the second control grid G II will not become negative until the negative voltage at control grid G I has already exceeded a certain amount.

If trimpot P2 is adjusted so that the rise of the negative voltage on the second control grid just starts at the point, when the first section of the 6AF6G has reached full scale, a flying change of indication from shadow angle 1 to shadow angle 2 is obtained.

That way the entire range of control voltage is distributed successively to both indicating sections.

Resistors R5 and R6 form a voltage divider with R1 and R2 respectively, to prevent overlap of the luminous fields.

In the test circuit, the operating voltage + Ub was 272 V, the voltage Uk at the cathode of 6AF6 was 72 V, the luminous screen voltage Ul against cathode was 187 V. The voltage swing of the 6C8 plates went from 30 to 202 V (AI), and from 17 to 202 V (A II), by a control voltage swing from 0 to -15 V.

Double triodes with µ = 30 .. 40 and internal resistance of 20 to 30 kΩ, such as the 6C8 or the 12AY7, are particularly suitable as control tubes. Also some double triodes, which were developed for computer-flip-flop circuits are promising, e.g. 1684, 2033, 6211, E90CC, etc.

These tubes require a relatively low control voltage for full scale on both indicating sections and they also pull down the plate voltage low enough to get the luminous fields of the 6AF6 very narrow if control voltage is 0 V.

Low-µ tubes such as ECC82, 12AU7 or 6AH7, pulling down the plate voltage even further without any substantial benefit, but they require a rather high control voltage for full output.

If less control voltage is available, high-µ tubes such as ECC83, 12AX7 or 6SL7 may solve the problem. Since their internal resistant is also higher, the plate pull-up resistors R1 and R2 must be changed to 1 MΩ, the pull-down resistors R5 and R6 have to be removed. This way, a 6SL7 requires a control voltage of –7.5 V for full output of both 6AF6 sections, respectively -3.75 V for full output of the first section.


The image shows the course of the two sectors of the 6AF6, driven by a 6C8 as shown in the test circuit, while the control voltage U-in rises from 0 to -15 V.
On the 1
st and 2nd row from the top, only a change in the left-hand sector appears, whereas in the 3rd and 4th rows only a change in the right-hand sector appears, while the the left-hand sector remains at full-scale.

 Our fellow RM member Otmar Jung was so kind to make an animated image from the photo series, which is gratefully acknowledged to him.

However, it becomes obvious, that the transition from sector 1 to sector 2 is not optimally tuned, because at -7 V very little changes. The delayed start of sector 2 comes slightly too late, so that a portion of the voltage change is swallowed. This bug has already been detected and corrected, but the effort to make a new series of images would have been too much.

Best Regards, Jacob



This article was edited 20.Oct.11 21:23 by Jacob Roschy .

Michael Watterson
IRL  Articles: 996
Schem.: 650
Pict.: 2497
20.Oct.11 23:02

Count of Thanks: 6
Reply  |  You aren't logged in. (Guest)   2

I presume the 6AD6, 6AF6 and 1629 are similar to 6AF6G and only vary in voltage or package?

Very interesting article. I had mentioned the dual control of 6AF6 here but with no example of how to use it. I've linked the paragraph to your article now.

Jacob Roschy
Jacob Roschy
D  Articles: 1715
Schem.: 17
Pict.: 68
24.Oct.11 18:37

Count of Thanks: 6
Reply  |  You aren't logged in. (Guest)   3

Hello Michael,

the 6AD6 is indeed similar to the 6AF6(-G), but has a maximum target voltage of 150 V, while the 1629 is merely a 12.6 volt heater octal based version of the 6E5 with built-in driver triode and only one shadow angle, hence dissimilar to the 6AF6.

Thank you for setting a link to my article.

Best Regards, Jacob