I was just kidding about laser trimming a tube. I wouldn't know where to start with so many variables affecting the plate curves.
Like cathode area and surface roughness, heater power, knee voltage, grid alignment, gm, Ri, Mu, screen current ...
Maybe could design a tube composed of many small sub sections and spot weld a certain number of them together with a laser for some sort of match. Still pretty iffy.
For audio, maybe use an Op Amp to servo control the screen grid V to maintain some kind of real time monitored performance.
Fortunately for an RIAA pre-amp, the signals are so small that linearity probably isn't a big deal for the 1st stage.
Like cathode area and surface roughness, heater power, knee voltage, grid alignment, gm, Ri, Mu, screen current ...
Maybe could design a tube composed of many small sub sections and spot weld a certain number of them together with a laser for some sort of match. Still pretty iffy.
For audio, maybe use an Op Amp to servo control the screen grid V to maintain some kind of real time monitored performance.
Fortunately for an RIAA pre-amp, the signals are so small that linearity probably isn't a big deal for the 1st stage.
Last edited:
Sort of.
Sort of.
The trick is to hide the IC error correction inside the chassis and you can still call it a tube amplifier.
See the LM4562 data sheet, read the discussion around fig.109. http://www.ti.com/lit/ds/symlink/lm4562.pdf
DT
Sort of.
The trick is to hide the IC error correction inside the chassis and you can still call it a tube amplifier.
See the LM4562 data sheet, read the discussion around fig.109. http://www.ti.com/lit/ds/symlink/lm4562.pdf
DT
Yes, I've seen the TI THD test method, but what is going on inside the IC? (probably the input differential stage is losing its gain with the resistor attached)
One common monitor/control scheme is to put an out of band signal into the amplifier and monitor it's amplitude at the amplifier output. Then that gets subsequently filtered out of the output. Its monitored level would control some gain control parameter like a tail current. Usually that would be some HF out of band signal.
With two (separate) low noise tube amplifiers arranged for a balanced input/output, we could turn that around.
We put in a small DC common mode signal and monitor the output "DC" common mode versus differential AC signal operation. Constant gain would mean constant DC CM output, so provides a means of monitoring dynamic amplifier gain. The common mode signal can be deleted by a subsequent stage for audio purposes. But we would use that "DC" common mode (internal) output signal to dynamically control the gain of the amplifier. Maybe by screen grid V control or stage operating bias etc. (so we get a CM gain control loop) This should flatten the gain and some of the noise. Although it is not -directly- sensitive to differential gain variation. Matched amplifier channel components (tubes) assure the common mode gain very nearly matches the differential mode gain (for small AC signals which are differential mode). (an ultra low noise Op Amp would be used for the servo control, although nothing says we couldn't use a tube Op Amp for that, just performance level gain/dist. and noise)
.
One common monitor/control scheme is to put an out of band signal into the amplifier and monitor it's amplitude at the amplifier output. Then that gets subsequently filtered out of the output. Its monitored level would control some gain control parameter like a tail current. Usually that would be some HF out of band signal.
With two (separate) low noise tube amplifiers arranged for a balanced input/output, we could turn that around.
We put in a small DC common mode signal and monitor the output "DC" common mode versus differential AC signal operation. Constant gain would mean constant DC CM output, so provides a means of monitoring dynamic amplifier gain. The common mode signal can be deleted by a subsequent stage for audio purposes. But we would use that "DC" common mode (internal) output signal to dynamically control the gain of the amplifier. Maybe by screen grid V control or stage operating bias etc. (so we get a CM gain control loop) This should flatten the gain and some of the noise. Although it is not -directly- sensitive to differential gain variation. Matched amplifier channel components (tubes) assure the common mode gain very nearly matches the differential mode gain (for small AC signals which are differential mode). (an ultra low noise Op Amp would be used for the servo control, although nothing says we couldn't use a tube Op Amp for that, just performance level gain/dist. and noise)
.
Last edited:
Divder Magic
Hello,
It not what is going on inside the Op-Amp but what is happening with the 2-resistor divider network that wraps around the Op-Amp. The Op-Amp output, signal plus distortion, is looped back to the Op-Amp inverting input. The original AP Analyzer output and amplified signal completely cancel each out at the Op-Amp output. The only thing that remains is 100 times the Op-Amp distortion for the AP Analyzer to measure. That is the inner workings of fig. 109 in the LM4562 datasheet.
Using the same resistor divider network thinking and adding a second Op-Amp an error correction amplifier, Op-Amp loop, can be added to a vacuum tube amplifier. This time it is the distortion that is canceled out, it is only signal that remains.
DT
Hello,
It not what is going on inside the Op-Amp but what is happening with the 2-resistor divider network that wraps around the Op-Amp. The Op-Amp output, signal plus distortion, is looped back to the Op-Amp inverting input. The original AP Analyzer output and amplified signal completely cancel each out at the Op-Amp output. The only thing that remains is 100 times the Op-Amp distortion for the AP Analyzer to measure. That is the inner workings of fig. 109 in the LM4562 datasheet.
Using the same resistor divider network thinking and adding a second Op-Amp an error correction amplifier, Op-Amp loop, can be added to a vacuum tube amplifier. This time it is the distortion that is canceled out, it is only signal that remains.
DT
Attachments
Hope This Helps
According to 6J52P = D3A = E810F = 7788 tube. Gold grid ― Hi-End vacuum tubes, sockets, capacitors, nixie. Retail and wholesale to worldwide. the 6J52P, E810F, & 7788 are all replacements for the D3a tube. Hope that helps in some small way...
I'm listening to Late One Night by Dan Siegal
Thetubeguy1954 (Tom)
According to 6J52P = D3A = E810F = 7788 tube. Gold grid ― Hi-End vacuum tubes, sockets, capacitors, nixie. Retail and wholesale to worldwide. the 6J52P, E810F, & 7788 are all replacements for the D3a tube. Hope that helps in some small way...
I'm listening to Late One Night by Dan Siegal
Thetubeguy1954 (Tom)
...mostly...
Read my #23 post there:
https://www.diyaudio.com/forums/tubes-valves/320418-45-amp-build-direct-coupled-3.html#post5383141
- Status
- Not open for further replies.