H2

dang if that isn't tempting
 

Attachments

  • temp.jpg
    temp.jpg
    64 KB · Views: 2,235
You mean like this ?
Actually I was not that extreme. I was thinking about something more like these.
This is a very interesting part. I think many SET users are using SE preamp or SE driver stage, but how do they deal with H2 in their whole signal chain? Must be a black art.
Advanced calibrated black art, banned from Hogwart stuff. :D
 

Attachments

  • SPE-270-Hibon-6__40869.1512406469.1280.1280.png
    SPE-270-Hibon-6__40869.1512406469.1280.1280.png
    93.4 KB · Views: 2,140
Nelson Pass said:
...The first option is to buy some Triodes (for example low voltage 6922's) and some other parts plus the equipment to build and calibrate a single-ended gain stage with just the right amount of 2nd harmonic....
...I personally think that this is a trick best played only once in the audio chain, so if you already have a SET amp or some other component with lots of 2nd harmonic, you want to temper your expectations when used with this also. Or maybe you want to hear what your SET sound like with the 2nd harmonic cancelled......
This is a very interesting part. I think many SET users are using SE preamp or SE driver stage, but how do they deal with H2 in their whole signal chain? Must be a black art.
As an addition, the late Eduardo de Lima of Audiopax noted in his paper Whysingle-endedtubeamplifiers.pdf that :
In figures 1, 2 and 3 we can see what happens with a loudspeaker driven by a low distortion amplifier and by a single ended amplifier of high 2nd harmonic distortion with both polarities. In this case the speaker alone has around 2% of distortion between 100 Hz and 1000 Hz with 2.83 Vrms in the input (that corresponds to 1W in an 8 ohm load). We can see this in figure 1. In figure 2 we connect it to a SE amplifier with an average 0.8% 2nd harmonic distortion at this output level. It is easy to see the distortion addition between 150 Hz and 300 Hz and subtraction around 900Hz. We could say that, as an average, the distortion has increased. Figure 3 shows the same set up with the other polarity. We can see that the whole 150 Hz till 300 Hz region has reduced distortion and that around 900 Hz we have an increase in it. The overall effect seems to be a reasonable decrease in distortion.

I just want to remind those interested in effects of H2 and blessed to attend the BAF 2018 that in addition to our beloved Papa and Wayne, another wizard versed in the magic of Hollow State electronics by the name of Roger Modjeski is scheduled to come. So prepare your questions and be totally wowed by the "black art". May the coming Triwizard meet unveils lots of Spooky level entertainment for us greedy boyz. :)
 
The one and only
Joined 2001
Paid Member
You can adjust the voltage as you like, and the range of Jfet characteristics
is wide enough to need that larger range. However once you have found
the "sweet spot" you will find that a much smaller range will cover the
practical variations of interest. I mention the +/- 0.3V as a guide for those
who only have a DC voltmeter.

:wiz:
 

PRR

Member
Joined 2003
Paid Member
Remember that harmonic distortion of simple devices depends on signal level.

This basic JFET may give 5% at high level but 0.5% at lower levels. In many systems the lower levels and % may be what you get.

2nd harmonic rises nearly(?) as the signal level, 3rd rises like the square of level. A small reduction of level makes a small reduction of 2nd but a large reduction of 3rd.

Mr Pass' plan is so bent that I suspect "no 3rd harmonic" until the 2nd harmonic is totally saturated and the stage clips. With careful bias, it could batter half the wave and still not show much 3rd.

Instead of signal level, you *may* be able to adjust supply voltage. In typical hi-fi use, crank it up to 100V (with a 100V JFET!) and it will be "no distortion"; turn it down to 2V and you have gobs of distortion even at low levels.
 
The one and only
Joined 2001
Paid Member
Mr Pass' plan is so bent that I suspect "no 3rd harmonic" until the 2nd harmonic is totally saturated and the stage clips. With careful bias, it could batter half the wave and still not show much 3rd.

Instead of signal level, you *may* be able to adjust supply voltage. In typical hi-fi use, crank it up to 100V (with a 100V JFET!) and it will be "no distortion"; turn it down to 2V and you have gobs of distortion even at low levels.

I believe this was covered in the article I posted at FW, showing the
distortion vs output voltage, the 3rd vs 2nd, and comments on voltage
adjustment.

With this circuit you can get a more pure 2nd vs 3rd by raising the
supply voltage and operating at the opposite phase 2nd, but it then
clips earlier, and you would usually want to flip the phase before the
Jfet.
 
That "sweet spot" between the relative amount of H2, H3 and successive for an expected variation of the signal of, for example, 10 to 15 dB of difference between strident and less strident passages, can only occur for a defined level, but not for that entire range. Is that "sweet spot" maintained with signals of very different duration in time or with very different cadences? I raise that because there are parameters that can be affected by the establishment and variation of the temperature on the chip with the presence of the signal and with the shape of the envelope of the amplitude of the same signal.
 
The one and only
Joined 2001
Paid Member
If we run distortion vs level tests with the previous signal having been
at high levels vs low level, we don't see a significant difference, so I
conclude that it isn't much of an issue.

The time constant on this is pretty long - it takes about 2 minutes for the
Jfet to warm up on turn-on.
 
If we run distortion vs level tests with the previous signal having been
at high levels vs low level, we don't see a significant difference, so I
conclude that it isn't much of an issue.

The time constant on this is pretty long - it takes about 2 minutes for the
Jfet to warm up on turn-on.

Thank you for your clarification, Mr. Pass. Since a difference between musical passages of 10 dB supposes an alteration of similar amount of 10 dB between both H2, but of 20 dB between both H3. If we have fixed a "sweet spot" ratio of 18 dB between the H2 and the H3 at an output, for example, of 0.5 V, we see that if that same output decreases at a level of 0.15 V, the ratio between the new H2 and new H3 is now 28 dB and not 18 dB.

Best regards
 
I'm curious about the SIT-1 where the phase of H2 can be adjusted. How was this accomplished? Big variable power supply?

I had a couple J112s so I built the H2. Interesting how low the THD can go when the 2nd harmonic cancels. Yes, you need a limited adjustment range on the power supply around the sweet spot. When I first read the article, I was unsure exactly what points on the waveform were "in phase". Looking at the Channel and Reading outputs of my AP System One with a dual trace scope and triggering on the + slope of the Channel output, it finally became clear.

I wonder about the "no degeneration" directive. This seems to discount the work of Dimitri Danyuk on Triode Vacuum tube emulation. Link: Triode Emulator by Dimitri Danyuk | Field Effect Transistor | Amplifier

I tried a trim pot to add a little degeneration and didn't see a rise in higher order harmonics above -80dB, but I didn't move the spectrum analyzer input to the measurement output.