The published data for that UL of operation includes THD curves that barely
exceed 1% below clipping - with NO feedback applied .
Well, a number of things affect published data which we always
must question about conditions for the test unless fully stated.
** See:
http://www.r-type.org/pdfs/el34.pdf
37W = 1.3% THD class AB ( just )
20W = 0.8% THD class A
Schem and curves are shown half way down the same page.
-------------------------------------------------------------------
However, Dynaco claim virtually the same THD figures, with circa 20dB of
feedback applied ??
I searched around for data on MkIV and I didn't find
any info saying THD was the same with or without NFB.
** Of course you did not, cos that is NOT what my post says !!!
I am aware that your reference was not to what Dynaco did / didn't say, but referred to Mullard data or equal or same from 1954.
Is the 7199's pentode gain stage mainly responsible for this anomaly ??
One fact is that Dynaco has 7199 pentode in SE mode, so a mix of odd+even H are produced. The absence of a balanced LTP drive amp in MkIV means pentode even H are not reduced by the balanced action of an LTP. At least some reduction of 2H in in EF86 in Quad-II is achieved.
For comparison, the Quad II use two, low cost, EF86 pentodes in
differential mode to provide gain and phase inversion. The output stage uses
local (cathode) feedback, plus overall feedback but is otherwise
conventional PP with KT66s or 6L6GCs performing equally.
Agreed, except that now, who cares about EF86 being cheap?
The Quad II achieves 0.1% THD at rated power and 0.05% THD at circa 10 watts
or less.
Well all that depends on load conditions, and perfect tube matching, and
almost no Quad-II amps I've tested achieve these wonderful figures you
quote.
** Really??
IME they all do.
I think not all,because tube properties affect THD. In any typical Quad-II, swapping positions of KT66 can make a big difference to THD because of 2H cancelling, and also when changing positions of EF86. Some working serviceable Quad-II I have tested have THD way above the Data Quad crow about. Peter Walker knew all about the limitations of his amps and the effects of aging. But 99% of Quad-II users would not hear any indication whatever that THD at say 3Watts was 10 times higher than the THD curves for Quad-II might indicate. Most ppl listened at less than average of 1Watt, so discussions of THD become academic value for most. And by listeners I include maybe hundreds of professional guys in recording studios monitoring what they did between 1950s and 1970s, by which time the tube gear was largely banished, replaced by SS with lower THD but which may have sounded no better, if we are to listen to what some audio enthusiasts say.
My example tests 0.03% at 10W with rather worn Sylvania 6L6GCs and very old
EL84s.
?? I think you mean "worn Sylvania 6L6GCs and very old EF86"
I have a UL PP amp in my shed which has the first OPT I ever wound in 1994 and it had ancient worn Sylvania 6CA7 in PP with 50% UL taps but has 12AT7 paralleled for input, 6CG7 driver in LTP. The B+ is a regulated +400V and and I have fixed bias, so no mis-biasing because of Ek rise during class AB which usually makes all AB amps display much higher THD than with fixed bias.
With 15dB GNFB I got 0.125% THD at 32 Watts, AB1, and THD was better than proportional to Vo, so that at 10W where Po was all class A1 I got about 0.06%,
and about what anyone could expect. If I put on 6dB more GNFB, THD would be half the above figures.
Over time the 6CA7 gave up, and began overheating when g1 bias failed to keep the Ia under control due to tube gas increasing. So I changed to KT88 and amp gives over 50W max and very good THD figures. UL IS GOOD, no doubt about it, but well done CFB is better.
I share Quad's belief that the CFB from OPT very much linearizes all output pentodes and tetrodes, and works better than the lazier use of UL taps and because the CFB is applied around a linear external loop including the higher gm of grid 1, not just the slightly non linear action of g2 which has a low gm and cannot do much when the RLa-a goes low, as it surely does when designers don't provide outlets of OPT straps to allow a match to 4r0. Many ppl use 4r0 speakers with Quad-II strapped for the lowest recommended strapping for 8r0, and then the THD becomes quite high.
When I removed the OPT bell ends from Dynaco MkIV OPTs, I found enough sec windings to allow the sec turns to be changed to get 8k6 : 8r0 and 8k6 : 3r5 and because I've used KT88 and all triode drive amp, I can boast that the MkIV will outperform any Quad-II in all ways that can be measured.
During the re-engineering process of MkIV, I did consider using a choke with grounded CT and between KT88 cathodes and with cross coupling by elcaps from cathodes to OPT screen connections. It effectively doubles the RLa-a, without changing the sec windings. And in class AB, more than 1/2 the OPT primary is used. The MkIV has 33% UL taps, so CFB% would have been high but about 80Vrms needs to be applied to each output grid. But under the MkIV chassis there just wasn't room for extra L and C parts so I settled for a plain UL schematic similar to one channel of a 5050 amp you'll find at my website.
High CFB% yields very low output stage THD but then driver design has to be very good, because the total amount of GNFB + CFB should not exceed 20dB, so there is a point where increasing CFB and then lowering GNFB means that driver amp THD is not reduced by the lowered amount of GNFB. McIntosh take matters further with 50% CFB, and when loaded to give pure class A, they are remarkably linear.
I much like Olde Quads, but its so easy to make them better. I have a pair which soon will get full treatment. I have a pair of new OPTs I had specially wound by a Chinese guy in Sydney who really knows what he's doing. Old Quad-II OPTs will go into another amp to replace far worse OPTs I had scrounged from somewhere. The tube rectifier will go and the screen choke, and that's where driver triodes will go because new OPT occupies where old one was plus where EF86 were placed.
The arguments which have been most relevant to my ears for last 20 years have been about what sounds better, and generally, I got better reports for my gear I'd made if I used better techniques than found in original Quad-II.
In my SE32 amp, I have a single 13E1, it uses some g2 FB, and 33% CFB, and I get 32W and very low THD for an SE amp. Same goes for SE35, with 4 x EL34..
It ain't what you got that counts, its how ya use it.
All past amplifier techniques may be questioned, lest we have to put up with mediocre crap which could not perform optimally in 1955 because they didn't have the diodes and caps and low prices in real terms that we have now.
Patrick Turner.
..... Phil