[[email protected]]

I have now made noise measurements on 28 6AU6 tubes, all connected with
plate and suppressor grounded and using the screen grid as anode. The
cathode bias resistor is fully bypassed. Initial tests showed that noise
was masked by hum from the heaters so for these tests 4xAA batteries were
used for the heater supply. Noise measurements were made with the grid
shorted to ground.The output noise in some tubes was so low it was
difficult to measure accurately. In general the peak to peak measurement of
the noise waveform on a scope connected to the output of a Ac
millivoltmeter (20KHz bandwidth) was divided by 8 to get the rms value. The
results we

1. Set of 5 Brimar EF94, and set of 4 Novosibirsk 6HN6 and 4 GE 6AU6W all
measured 50uV rms noise

2. An RCA 6AU6A, 3 of Sylvania 6AU6. 1 Tungsolo 6AU6 and 1 Philco 6AU6 all
measured 75uV rms noise.

3. 2 RCA 6AU6A measured 125uV rms noise.

4. GE 6AU6 measured 185uV rms noise.

5. 3 RCA 6AU6A, 1 Sylvania 6AU6, 1 Tungsol 6AU6 and 2 Motorola 6AU6 all
measured 250uV rms noise.

Conclusions. I measured the stage gain of one of the EF94 samples. The
result was 25. This implies the best tubes are capable of an EIN of 2uV;
not earth shattering but good enough for a mic pre with a 10:1 transformer
at the input with a 150 ohm source. Nearly half the tubes tested achieved
this noise level which is encouraging.

Ian

[Don Pearce]

On Mon, 09 Jul 2007 22:46:22 +0100, wrote:

I have now made noise measurements on 28 6AU6 tubes, all connected with
plate and suppressor grounded and using the screen grid as anode. The
cathode bias resistor is fully bypassed. Initial tests showed that noise
was masked by hum from the heaters so for these tests 4xAA batteries were
used for the heater supply. Noise measurements were made with the grid
shorted to ground.The output noise in some tubes was so low it was
difficult to measure accurately. In general the peak to peak measurement of
the noise waveform on a scope connected to the output of a Ac
millivoltmeter (20KHz bandwidth) was divided by 8 to get the rms value. The
results we

1. Set of 5 Brimar EF94, and set of 4 Novosibirsk 6HN6 and 4 GE 6AU6W all
measured 50uV rms noise

2. An RCA 6AU6A, 3 of Sylvania 6AU6. 1 Tungsolo 6AU6 and 1 Philco 6AU6 all
measured 75uV rms noise.

3. 2 RCA 6AU6A measured 125uV rms noise.

4. GE 6AU6 measured 185uV rms noise.

5. 3 RCA 6AU6A, 1 Sylvania 6AU6, 1 Tungsol 6AU6 and 2 Motorola 6AU6 all
measured 250uV rms noise.

Conclusions. I measured the stage gain of one of the EF94 samples. The
result was 25. This implies the best tubes are capable of an EIN of 2uV;
not earth shattering but good enough for a mic pre with a 10:1 transformer
at the input with a 150 ohm source. Nearly half the tubes tested achieved
this noise level which is encouraging.

Ian

Is 150 ohms the optimum source through a 10:1 transformer? Have you
measured the noise level with a 15k grid resistor rather than a short?

d

--
Pearce Consulting
http://www.pearce.uk.com

[Patrick Turner]

wrote:

I have now made noise measurements on 28 6AU6 tubes, all connected with
plate and suppressor grounded and using the screen grid as anode. The
cathode bias resistor is fully bypassed. Initial tests showed that noise
was masked by hum from the heaters so for these tests 4xAA batteries were
used for the heater supply. Noise measurements were made with the grid
shorted to ground.The output noise in some tubes was so low it was
difficult to measure accurately. In general the peak to peak measurement of
the noise waveform on a scope connected to the output of a Ac
millivoltmeter (20KHz bandwidth) was divided by 8 to get the rms value. The
results we

1. Set of 5 Brimar EF94, and set of 4 Novosibirsk 6HN6 and 4 GE 6AU6W all
measured 50uV rms noise

2. An RCA 6AU6A, 3 of Sylvania 6AU6. 1 Tungsolo 6AU6 and 1 Philco 6AU6 all
measured 75uV rms noise.

3. 2 RCA 6AU6A measured 125uV rms noise.

4. GE 6AU6 measured 185uV rms noise.

5. 3 RCA 6AU6A, 1 Sylvania 6AU6, 1 Tungsol 6AU6 and 2 Motorola 6AU6 all
measured 250uV rms noise.

Conclusions. I measured the stage gain of one of the EF94 samples. The
result was 25. This implies the best tubes are capable of an EIN of 2uV;
not earth shattering but good enough for a mic pre with a 10:1 transformer
at the input with a 150 ohm source. Nearly half the tubes tested achieved
this noise level which is encouraging.

So compared to 12AT7, 12AX7, 6DJ8, 6CG7, 12AU7 or most other real
triodes
the noise figures you have are about average.

2uV isn't bad for MM or MC with a step up tranny, or a microphone with
step up tranny,
but with MM without a tranny and if the rated output was 2mV,
then 2uV is only -60dB SNR, and barely good enough.

Weighting the SNR figure makes things seem better, so don't, because
weighting allows for higher LF noise to which the ears are not supposed
to be so sensitive,
but many ancient amps both SS and tube only could boast -55dB for phono,
unweighted.
OK for rock and roll maybe but not good enough for much classical where
there are quiet passages in the music, which leaves the venue noise,
tape noise, vinyl noise and amp noise to spoil what may have been a fine
recording.

With a single j-fets such as 2SK369/2SK147, noise figures of 20dB lower
and thus better are routine.

Patrick Turner.









Ian

[[email protected]]

Don Pearce wrote:

On Mon, 09 Jul 2007 22:46:22 +0100, wrote:

I have now made noise measurements on 28 6AU6 tubes, all connected with
plate and suppressor grounded and using the screen grid as anode. The
cathode bias resistor is fully bypassed. Initial tests showed that noise
was masked by hum from the heaters so for these tests 4xAA batteries were
used for the heater supply. Noise measurements were made with the grid
shorted to ground.The output noise in some tubes was so low it was
difficult to measure accurately. In general the peak to peak measurement
of the noise waveform on a scope connected to the output of a Ac
millivoltmeter (20KHz bandwidth) was divided by 8 to get the rms value.
The results we

1. Set of 5 Brimar EF94, and set of 4 Novosibirsk 6HN6 and 4 GE 6AU6W all
measured 50uV rms noise

2. An RCA 6AU6A, 3 of Sylvania 6AU6. 1 Tungsolo 6AU6 and 1 Philco 6AU6 all
measured 75uV rms noise.

3. 2 RCA 6AU6A measured 125uV rms noise.

4. GE 6AU6 measured 185uV rms noise.

5. 3 RCA 6AU6A, 1 Sylvania 6AU6, 1 Tungsol 6AU6 and 2 Motorola 6AU6 all
measured 250uV rms noise.

Conclusions. I measured the stage gain of one of the EF94 samples. The
result was 25. This implies the best tubes are capable of an EIN of 2uV;
not earth shattering but good enough for a mic pre with a 10:1 transformer
at the input with a 150 ohm source. Nearly half the tubes tested achieved
this noise level which is encouraging.

Ian

Is 150 ohms the optimum source through a 10:1 transformer?

It is the typical source resistance of a microphone.

Have you
measured the noise level with a 15k grid resistor rather than a short?


No, but the calculated equivalent input noise resistance is about 12K so
with 15K across the grid I would expect about a 3dB increase in output
noise.

Ian
d

[[email protected]]

Patrick Turner wrote:



wrote:

I have now made noise measurements on 28 6AU6 tubes, all connected with
plate and suppressor grounded and using the screen grid as anode. The
cathode bias resistor is fully bypassed. Initial tests showed that noise
was masked by hum from the heaters so for these tests 4xAA batteries were
used for the heater supply. Noise measurements were made with the grid
shorted to ground.The output noise in some tubes was so low it was
difficult to measure accurately. In general the peak to peak measurement
of the noise waveform on a scope connected to the output of a Ac
millivoltmeter (20KHz bandwidth) was divided by 8 to get the rms value.
The results we

1. Set of 5 Brimar EF94, and set of 4 Novosibirsk 6HN6 and 4 GE 6AU6W all
measured 50uV rms noise

2. An RCA 6AU6A, 3 of Sylvania 6AU6. 1 Tungsolo 6AU6 and 1 Philco 6AU6
all measured 75uV rms noise.

3. 2 RCA 6AU6A measured 125uV rms noise.

4. GE 6AU6 measured 185uV rms noise.

5. 3 RCA 6AU6A, 1 Sylvania 6AU6, 1 Tungsol 6AU6 and 2 Motorola 6AU6 all
measured 250uV rms noise.

Conclusions. I measured the stage gain of one of the EF94 samples. The
result was 25. This implies the best tubes are capable of an EIN of 2uV;
not earth shattering but good enough for a mic pre with a 10:1
transformer at the input with a 150 ohm source. Nearly half the tubes
tested achieved this noise level which is encouraging.

So compared to 12AT7, 12AX7, 6DJ8, 6CG7, 12AU7 or most other real
triodes
the noise figures you have are about average.


That is good news. I did not expect them to be better but it is good to know
they are not really worse and they are a hell of a lot cheaper than any of
the tubes you listed.

2uV isn't bad for MM or MC with a step up tranny, or a microphone with
step up tranny,
but with MM without a tranny and if the rated output was 2mV,
then 2uV is only -60dB SNR, and barely good enough.


Indeed, a tranny seems to be essential.

Ian