Welcome! Need support, you got it. Or share youe ideas and experiences.
Return to K&K Audio / Lundahl Transformers
In Reply to: My observations posted by IslandPink on June 04, 2004 at 08:32:26:
There's been a quiet discussion offline about circuits that are transitional between the conventional cap-bypassed Class A PP of the Amity/Aurora/Karna and the current-sourced Allen Wright approach. Oddly enough, Western Electric got there first - in 1936, and called it the "Harmonic Equalizer."The Harmonic Equalizer only affects 3rd, 5th, 7th etc. harmonics, and is not what they used to call "reverse feedback", which was reserved for the SE 91A amplifier.
Anyway, to advantage of the Western Electric-style Harmonic Equalizer, insert a resistor with a value somewhere between 20 and 80 ohms in series with the CT/filament shunt cap. In the ECC99 driver circuit, use a resistor with value somewhere between 40 and 160 ohms in series with the CT/cathode shunt cap.
What does this do? Well, the Harmonic Equalizer reduces 3rd harmonic by 10dB, and 5th by 15 to 20dB. There is *no* effect on 2nd and 4th, which isn't really so bad, since they are sonically more benign, and are at pretty low levels anyway due to circuit balance. Oh yes, clipping is improved, as well.
Since there is very little power going through this small resistor, you can twiddle with a cheapo pot just to assess the effect, then replace with a good-quality wirewound resistor. Remember, the Harmonic Equalizer is nothing more than a resistor in series with the usual bypass cap, so it works equally well in a conventional Class A PP cathode-bypassed circuit.
Follow Ups:
How does this affect output impedance? For instance with an 80 ohme resistor is it increased by (mu +1)*80 in series with internal impedance and dc resistance of OPT?
For a SE circuit, yes, there would be a rise in output impedance (damping factor), but this is Class A PP circuit, so increasing the common cathode impedance actually has no effect on output Z at all - although there are minor secondary effects related to clipping, etc. when comparing zero impedance (normal cap-bypass) to near-infinite (current source).The re-discovery of the WE Harmonic Equalizer, lost since 1936, is prompting a lot of off-line discussion amongst the PP triode folks. The reduction of 3rd and 5th harmonics is real, measurable, and exists over a range of signal levels from 5 watts down to a few milliwatts. (Measured on a $15,000 Audio Precision analyzer with a noise floor at -120dB, so we're not talking test-equipment limitations here.)
Hi Lynn,Thanks for the article. It seems that this is a refinement of the Ultrapath idea. What is the advantage of using the series resistor to attenuate the grunge on the CT rather than to take the whole grunge to the cathode?
Dave
By John Atwood of One Electron (see URL below). He's the one with the Audio Precision test set, a test amplifier, and links to the original WE articles.
The actual test amplifier has independently biased triode-connected 6CK4's (each with its own adjustable current source), and a pair of 1000uF electrolytics in series between the two cathodes. The center-point of the two caps can be returned to ground, left open-circuit, or connected to a 500-ohm pot. If it is returned to ground, that is equivalent to normal cap-bypassed Class A PP amplifier; if it left open, that's equivalent to the Allen Wright current-source approach; and if the variable resistor is chosen, that's similar to the WE Harmonic Equalizer.Note the test amp is not designed to be listened to, but make a series of measurements. I expect John will re-test with battery-heated 2A3's at some point in the future, although these results are similar to what other people have seen with 300B's.
Although there is no change in output Z with common cathode source impedance, the distortion nulling, although broad, has some sensitivity to the load the output tubes see. For a 2K pri impedance (very low), 20 ohms is about right; for 10K or higher, 100 ohms is about right. But the 3rd and 5th-harmonic nulls are quite broad, so a 2:1 or 3:1 mismatch error in resistor value is not significant.
So it doesn't really make any difference if the 20 to 100 ohms series resistance is added to a conventional cathode-bypass capacitor, a WE-style shunt cap, or the cathode resistor is split in two (450 + 50 ohms, for example) and the cap connected to the center point of the two resistors.
Close examination of the WE86 (visit the Nutshellhifi.com Library for PDF) schematic also discloses an unusual network attached to the grid circuit - we're still discussing what that does.
nt
Ok Lynn,Thanks for that , I will print it out and take away to mull over on my business trip to Dundee .
I was thinking , a while back , of putting up a message on TubeDIY , subject 'Do we need 2nd harmonic' , but thought it might be asking for trouble with the culture there at the moment. I might still do that though .More later ,
This post is made possible by the generous support of people like you and our sponsors: