Over the time playing with tube amplifiers, I accumulated some number of various tube to allow me play “tube rolling” game. I experimented with two major types of low signal tubes: 6SL7 and 6SN7. I decided to try scientific approach and measure tube parameters, which may influence perceived sound of the amplifier.
Today’s notes are about 6SL7 type tube, which are usually used in first stage of amplifiers wherever high gain is needed. There were many tubes of that type made by various manufacturers and they are still made in Russia and China today. I have several different specimens of that tube to play with:
1. Two modern production Tung Sol 6SL7GTtubes, made in Russia
2. One modern 6N9P tube made in China
3. Tree 1950s vintage 1579 tubes made in Soviet Union
4. One WW2 vintage 7F7 tube made in USA by Motorola
Internal structure of these tubes can be classified by three categories:
- 1579 and 7F7 tubes internally look very similar to each other. It is likely that when Soviet engineers designed their tube, they copied it from American example.
- 6N9P is somewhat similar to 1579/7F7, but has a bit different plate support structure. But internal geometry is likely very close.
- 6SL7GT tubes do not look like their vintage brothers at all. It seems that they were designed independently from original tube.
In my tests I assumed that the “sound” of tube mostly depends on harmonics profile – amplitude of harmonic distortions in their relation to each other. That is why I decided to build simple circuit similar to one commonly used in audio amplifiers, and measure the individual distortion components at the output of that circuit. Circuit is a simple resistor based amplifier. Component parameters were selected from example in RCA Tube Handbook. They are close to what can be found in many tube pre- and power amplifiers.
I decided to measure distortion at two operation points. One is when input signal is 100mV. This is a common case for most audio devices when 6SL7 tube is used in first stage. Second operation point was when input signal is 2V peak to peak. The reason for that selection was that in circuit which I used, tube automatic bias is about 2V. Thus this is close to maximum possible input signal for the test circuit before any clipping may be observed. Surprisingly all vintage tubes from my set required LESS heater current (290mA) than modern production tubes (330mA). But there should be no issues unless your device has dozens of tubes.
First test was to check DC operation point. For all tubes plate voltage was 200V, bias 2V and idle current 0.9mA, with variations no more than 5% from tube to tube. It means that all tubes could be used in the same circuit without need to adjust other component values.
When I measured gain, I found that there were two groups: 6SL7GT tubes had lower gain than all other tube types. It means that with circuit that depends on feedback, there will be significant difference in response, due to lower gain of circuit before feedback is applied. Considering that 6SL7 tube is mostly used in first stage and gain is multiplied by following stages, it could become a serious problem.
Finally I ran series of tests where I measured distortion components. Results are presented in table and graphical forms. As you can see, variation of harmonic profile are bigger at low signal level than at high signal level. How that influence perceived “sound” of the tube is open for discussion. But nevertheless I could not find any “huge” difference and sound differences between tubes should not be easily heard. You need to be well trained to there these differences, and this is likely only possible wen you hear two amplifiers with different tubes side by side.
You are welcome to comment on my findings.
I also have different specimens of 6SN7 type tube and plan do a similar experiment with them. Stay tuned.
Today’s notes are about 6SL7 type tube, which are usually used in first stage of amplifiers wherever high gain is needed. There were many tubes of that type made by various manufacturers and they are still made in Russia and China today. I have several different specimens of that tube to play with:
1. Two modern production Tung Sol 6SL7GTtubes, made in Russia
2. One modern 6N9P tube made in China
3. Tree 1950s vintage 1579 tubes made in Soviet Union
4. One WW2 vintage 7F7 tube made in USA by Motorola
Internal structure of these tubes can be classified by three categories:
- 1579 and 7F7 tubes internally look very similar to each other. It is likely that when Soviet engineers designed their tube, they copied it from American example.
- 6N9P is somewhat similar to 1579/7F7, but has a bit different plate support structure. But internal geometry is likely very close.
- 6SL7GT tubes do not look like their vintage brothers at all. It seems that they were designed independently from original tube.
In my tests I assumed that the “sound” of tube mostly depends on harmonics profile – amplitude of harmonic distortions in their relation to each other. That is why I decided to build simple circuit similar to one commonly used in audio amplifiers, and measure the individual distortion components at the output of that circuit. Circuit is a simple resistor based amplifier. Component parameters were selected from example in RCA Tube Handbook. They are close to what can be found in many tube pre- and power amplifiers.
I decided to measure distortion at two operation points. One is when input signal is 100mV. This is a common case for most audio devices when 6SL7 tube is used in first stage. Second operation point was when input signal is 2V peak to peak. The reason for that selection was that in circuit which I used, tube automatic bias is about 2V. Thus this is close to maximum possible input signal for the test circuit before any clipping may be observed. Surprisingly all vintage tubes from my set required LESS heater current (290mA) than modern production tubes (330mA). But there should be no issues unless your device has dozens of tubes.
First test was to check DC operation point. For all tubes plate voltage was 200V, bias 2V and idle current 0.9mA, with variations no more than 5% from tube to tube. It means that all tubes could be used in the same circuit without need to adjust other component values.
When I measured gain, I found that there were two groups: 6SL7GT tubes had lower gain than all other tube types. It means that with circuit that depends on feedback, there will be significant difference in response, due to lower gain of circuit before feedback is applied. Considering that 6SL7 tube is mostly used in first stage and gain is multiplied by following stages, it could become a serious problem.
Finally I ran series of tests where I measured distortion components. Results are presented in table and graphical forms. As you can see, variation of harmonic profile are bigger at low signal level than at high signal level. How that influence perceived “sound” of the tube is open for discussion. But nevertheless I could not find any “huge” difference and sound differences between tubes should not be easily heard. You need to be well trained to there these differences, and this is likely only possible wen you hear two amplifiers with different tubes side by side.
You are welcome to comment on my findings.
I also have different specimens of 6SN7 type tube and plan do a similar experiment with them. Stay tuned.
Attachments
Last edited:
Black Glass Tubes
In the past I recall reading that old tubes with black glass coatings have lower distortion, it would be interesting if you could get a pair of those and measure them.
In the past I recall reading that old tubes with black glass coatings have lower distortion, it would be interesting if you could get a pair of those and measure them.
Member
Joined 2009
Paid Member
and Jan Jurco are now making new 6SL7 and 6SN7 valves too.
Attachments
Last edited: