John Broskie's Guide to Tube Circuit Analysis & Design

17 April 2006

Today’s blog entry was supposed to be on using the Aikido amplifier as a headphone amplifier, with special emphasis on making our new Aikido PCBs the foundation for such an undertaking (since we have them, we should put them to good use).

Well, this topic must be postponed until the next blog entry, as I have to perform some conceptual housekeeping before littering your mind with any more new interesting projects. (Besides, I am not done drawing the headphone amplifier schematics and I am waiting for some more feedback from a few readers who blazed ahead and our using the Aikido as headphone amplifiers.)

Getting R15 & R16 straight

The tipoff came from longtime reader, Paul, who told me that he measured more noise at the output of his Aikido after replacing his two 100k voltage-divider resistors with two resistors that followed the noise-voltage-divider formula,

Resistor ratio = 1/mu + ½

or (for those with my Aikido PCBs)

R16 = R15(mu + 2)/(mu - 2)

I have received enough e-mail from Paul to know that he can teach me plenty about carefully measuring an amplifier, so it was unlikely that he had incorrectly measured the noise. In fact, he had tested the resistor values with a VOM and he had tried other ratios and other 6SN7s, but always the two 100k resistors worked best. I was about to pull out my hair, when I asked if the bottom resistor was the larger. He replied that it wasn’t, but also noted that that wasn’t what I had posted here—and he included a screen capture of the Aikido schematic. He was right: I had posted the wrong order in several of the Aikido schematics. (I have since corrected the erroneous schematics.)

What a headache. First, there was the same error in the PCB user guide PDFs, now this. How did it happen? Easy. First, I used a different formula than the one I posted, which a reader had provided; my formula yields the same ratio, but does not lead to division-by-zero errors with a 6AS7 triode (I think like a programmer):

R15 = R16(mu – 2)/(mu + 2).

For example, a 6AS7’s mu of 2 renders R15 = 0, rather than R16 equals infinity. So I am accustomed to changing resistor R15’s (the top resistor's) value and leaving R16 (The Bottom resistor) fixed at 100k in my own design work and experiments. But I figured it was best to use the other formula on this website. Unfortunately, it proved harder to do than I expected. In other words, feel free to use either formula, but do so consistently.

Second, I cut and paste a great deal when drawing schematics, which means that an error can spread from one schematic to the next.

Yet even after I thought I had fixed the PDFs, I was told that nothing had changed; the PDFs still read the same. Madness. I used my FTP program to open the PDFs at the website and, indeed, they had been updated. However, when I downloaded them, they hadn’t been. What was going on? It turns out that you have to press the reload button on your web browser, otherwise it just keeps loading the old PDF from its cache on your hard drive, not the revised PDF off this website.

R15 and R16 Values for common triode types

The table below uses a fixed 100k R16 resistor value and show R15 in exact and the closest standard 1% resistor values.

Triode

mu

R15

R15 1%

R16

6AQ8

57

93,220

93.1k

100k

6AS7

2

0

0

100k

6BK7

43

91,111

90.9k

100k

6BL7

15

76,471

76.8k

100k

6BQ7

38

90,000

88.7k

100k

6BX7

10

66,667

66.5k

100k

6CG7

20

81,818

82.5k

100k

6DJ8

33

88,571

88.7k

100k

6FQ7

20

81,818

82.5k

100k

6GM8

14

75,000

75k

100k

6H30

15

76,471

76.8k

100k

6N1P

36

89,474

88.7k

100k

6N27P

14

75,000

75k

100k

6SL7

70

94,444

93.1k

100k

6SN7

20

81,818

82.5k

100k

12AT7

60

93,548

93.1k

100k

12AU7

17

78,947

78.7k

100k

12AV7

40

90,476

90.9k

100k

12AX7

100

96,078

96.3k

100k

12BH7

16.5

78,378

78.7k

100k

12BZ7

100

96,078

96.3k

100k

12DJ8

33

88,571

88.7

100k

12FQ7

20

81,818

82.5k

100k

12SL7

70

94,444

93.1k

100k

12SN7

20

81,818

82.5k

100k

12SX7

20

81,818

82.5k

100k

5687*

17

78,947

78.7k

100k

5691

70

94,444

93.1k

100k

5692

20

81,818

82.5k

100k

5751

70

94,444

93.1k

100k

5963

20

81,818

82.5k

100k

5965

41

90,698

90.9k

100k

6072

41

90,698

90.9k

100k

6080

2

0

0

100k

B65

20

81,818

82.5k

100k

E80CC

27

86,207

86,6k

100k

ECC33

20

81,818

82.5k

100k

ECC81

60

93,548

93.1k

100k

ECC82

17

78,947

78.7k

100k

ECC83

100

96,078

95.3k

100k

ECC85

57

93,220

93.1k

100k

ECC86

14

75,000

75k

100k

ECC88

33

88,571

88.7k

100k

*Tube cannot be used with the 9-pin Aikido PCB and is only listed for reference.

Coupling capacitor issues

First, how large in value do the output coupling capacitors have to be? Two variables come to play. The first is the load impedance to be driven and the second is the desired low-frequency cutoff. Then all that is needed is this simple formula:

Capacitor value = 159155 / Frequency / Resistance

The answer is in µF. For example, a cutoff of 100Hz and an input impedance of 47K would require a coupling capacitor of 0.0339µF, or after rounding to the closest standard value, 0.033µF. Another example: 5Hz and 100k requires 0.31831µF, or after rounding to the closest standard value, 0.33µF.

Can both outputs (from coupling capacitors C1 and C2) be used at once, with each one driving its own amplifier? Yes, indeed. Why would anyone want to do such a thing? I can think of several possibilities. For example, a bi-amped system could be easily assembled, using one of the coupling capacitors to limit the low-frequency energy to small satellite loudspeakers, while the other feeds the bass amplifier that powers the subwoofers. Thus, small and expensive coupling capacitors can be used for the high-frequency outputs and large—but cheap—coupling capacitors can be used for the low-frequency outputs.

Or one set of coupling capacitors could be used for a dedicated, solid-state, low-input-impedance headphone amplifier, while the other is used for a high-impedance tube power amplifier. I am sure that many of you will find other uses for the two coupling capacitors.

Here's a question I have gotten twice, which means there 200 hundred readers wondering the same thing: How do I wire up a rotary switch for switching between the two coupling capacitors? We need a four-pole, three-position switch and some hookup wire. All four coupling capacitors attach to the input contacts and the two channels of output can receive either coupling capacitors C1’s or C2’s or both capacitors’ outputs. The drawing below shows the knob on the faceplate and the rotary switch from behind. (The switch is shown on the "C1 + C2" position.)

Aikido amplifier background information:
blog 11 (Here is where it all began, back in 2004)
blog 13
blog 16
blog 30
blog 32
blog 51
blog 56
blog 57
blog 58
blog 59

Next time
Sorry for the quick detour. Next time, the promised headphone amplifier circuits.

//JRB