Re: 211 A2 amp


I said...
Well, I don't think a tube amp of any kind using global NFB and which
has been unconditionally stabilised
is going to suffer "deterioration of other parameters", which presumably
are imaging and
"openess of sound", whatever that is.

I increased the stability margin by several dB, and so with a little
less NFB the THD was marginally higher.

The best stability margin you can have is with no global NFB.

Then as you add NFB, you can draw a plot of the margin of stability if
you really must,
and by the time you have say 30dB applied, and the test with 0.22uF
across the
output causes inevitable HF oscillations no matter what you do to
the compensation networks, or, you get LF oscillations that cannot be
cured with
gain tweaker networks, then you have reached the point of maximal
applied FB for that
amp to get the wanted bandwidth say of 10Hz to 50kHz.
Its possible to tweak open loop phase shift even further if you settle
for less closed loop bandwidth
but you shouldn't, because you'll end up with an amp with
maybe 35dB NFB applied at say between 500Hz and 5kHz, and each side of
the that
the efectively applied NFB will reduce because of the reduction of OLG
with networks. With such reductions of OLG, the reduction of distortion
products
at above 5kHz becomes less, and sound goes hard, most probably.

30dB of NFB was applied around a large number of amps back in the 1960s
to get typical
THD < 0.05% at a dB below clip at 1kHz, R load only.
So at 3 watts typical was <0.02%. Companies liked to take part in wars
of Watts,
and battles of THD to get sales.


In my mind there is no need for such low THD/IMD figures simply
because the products formed in class A working of tube amps is much more
bearable
than say the same amount but higher numbered H in a class B SS amp that
has not been designed very well.
In fact 0.1% THD in any SET amp at 3 watts is fine, because its all 2H,
and IMD results are not as appalling as one might suspect.

But I get samples of SET amps made by DIYers and some companies like
Silk Amps from Thailand, and they have maybe 5% THD at 2 watts from a
very poorly
conceived circuit, and this is way too high.

The use of a pair of KT88, KT90 or 6550 of 300B for the output stage
means
anode idle dissipation is naturally higher than lesser capable tubes
such as
EL34, 6L6, KT66, 6V6, EL84 etc, so therefore the class A PO before class
AB action commences
is higher.

And if you ever draw a graph of the distortion against output voltage
you'll see that THD is very low while the amp is in class A, and then it
increases
more rapidly above the A to AB threshold, but which really only need to
handle transients
such as drum beats and cannon shots.

The maximum amount of class A available depends on the RLa-a load,
anode dissipation, ie, Ea x Ia at idle, and maximum anode efficiency for
class A
which is about 45% for multigrids and up to about 30% for the same
trioded
or for 300B.
More PO is available in A2.


If you carefully read all my pages on load matching to triodes and beam
tetrodes
you will see what load values required for pure class A and how the load
affects the maximum PO and balance between A and AB PO.
And if anyone is allergic to NFB, then to get anywhere decent ENOUGH
measurements that
the best sound from tubes depends, they must use triodes and a high
value RLa-a,
say 12k with KT88 and Ea = 500V, Ia = 60mA per output tube, ( only with
multi grids,
not 300B )

This means they'll select a Hammond 1650P which has 6k6 : 4, 8, 16 ohms
and
they will never use the 8 and 16 ohm outlet config,
and with 8 ohm speakers the tube RLa-a is 13.2k, and PO ALL class A.
OPT winding losses become lower, as you increase speaker Z
with a give outlet value.


Then it will be found that no matter what tricks are played, the output
stage
has around 1% THD mainly 3H at a dB below clipping, with about 20 watts
available.
If more PO is wanted, halve RLa-a and use 4 output tubes.
The 40 watts of pure class A will be sublime, but nobody bothers
to waste so much electricity, and they back off the bias to 35mA even
with only
2 output tubes.
Then class A PO drops but 90% of what they listen to is covered by pure
class A
and the distortions rise only slightly for the class A region although
its not unusual to get 2% THD or more with class AB at full PO.
Suitable input/driver stages with low THD are shown at my website.
At 3 watts in a 20 watt PP pure class A amp, expect 0.3%,
and 0.15% in the 40 watter.
Pure class A has THD dropping linearly with VO, not PO.

The least THD should be where there is no RL connected, and the load
line is a
nearly horizontal line, and maximal Va occurs with almost no Ia change.
But you will still see up to 1% THD just under clipping.

The internally applied local NFB action within the triodes is maximal
because gain is maximal.

Things are a lot worse with tetrodes or pentodes used as such.
The UL connection is a great help, but the CFB connection is better,
and can give lower THD than triodes at the same PO,
and maximal A1 PO is also greater.

I have not tried the 211 in a PP circuit.
Suppose Pda = 60 watts per tube.
At max class efficiency of say 25%, expect 30 watts of pure class A,
which is 948Vrms across the 30k OPT primary.

The OPT is a devil to wind compared to making one for 6k,
and to get the same wide BW from 10Hz to at least 50kHz.

845 under similar dissipation amounts would make 30% efficiency,
and give 36 watts, but one can run them with Pda = 75 watts each
and go into a little A2 and get 60 watts class A from a pair in either
SE or PP.
The PP load needs to be 24k, and the OPT is a difficult thing to wind
to get the bandwidth expected today.
Its possible that where the bandwidth is difficult to get,
you have two OPT, one for below 1kHz, and the other cap coupled to the
anodes for the above 1 kHz F.
global NFB arrangements if used become fiddly.

A circlotron with 845 could be done, but is entirely unecessary
because Rout and THD/IMD is all low enough with the anode loaded tubes
and no
loop NFB at all.

The 805 of some other high µ large triode may be suited to a circlotron
because you need a lotta FB with 805.
And you'd need a shirt and trouser load of driver voltage, which is
bleedin difficult,
wheras global NFB applied back to an input triode cathode is so much
easier.



Patrick Turner.
















Regards to all
Iain
.

Relevant Pages

  • Re: 211 A2 amp
    ... The best stability margin you can have is with no global NFB. ... Indeed but a little NFB does wonders for a PP amp I find. ... at above 5kHz becomes less, and sound goes hard, most probably. ... THD < 0.05% at a dB below clip at 1kHz, ...
    (rec.audio.tubes)
  • Re: Distortion in amplifiers.
    ... they set their levels so THD is 15% minimum most days... ... But Cheever's treatise includes the effect of using NFB in an amp making ... The past examiners of this phenomena have concluded that where open loop ...
    (rec.audio.tubes)
  • Re: Negative Feedback and the Criminal Assets Recovery Bureau
    ... You definitely have to keep perspective with NFB. ... My best amp ever, the T113 ... Planned distortion cancellation is a chimera. ... The thd is nearly proportional to output voltage in most amps. ...
    (rec.audio.tubes)
  • Re: 211 A2 amp
    ... The best stability margin you can have is with no global NFB. ... Indeed but a little NFB does wonders for a PP amp I find. ... THD < 0.05% at a dB below clip at 1kHz, ...
    (rec.audio.tubes)
  • Re: Distortion in amplifiers.
    ... The R3 590k bias resistor needs to be a much lower value. ... The dedicated experimentor would find the best Ec for lowest THD at the ... and 22k from its emitter output to -40V. ... Then i'd have a 680 ohm Q1 emitter resistor to give local current NFB ...
    (rec.audio.tubes)