Single-ended triumphs

[Andre Jute[_2_]]

The triumph of single-ended can be seen even in bikes:
http://www.cannondale.com/onbike

Andre Jute
Visit Jute on Bicycles at
http://www.audio-talk.co.uk/fiultra/BICYCLE%20%26%20CYCLING.html

[Patrick Turner]

On Apr 4, 10:18*am, Andre Jute > wrote:
> The triumph of single-ended can be seen even in bikes:
> *http://www.cannondale.com/onbike

I huess you have seen the latest thing from the mighty BMW company - a
bike with a shaft drive to the rear wheel in a swinging-sprung-damped
tube.

http://www.bmw.com./newbike/uppenhillen

Meanwhile, the most serious single ended cyclist is Micheal Milton...

http://www.michaelmilton.com/default.aspx?d=457759

And meanwhile, we might ask what's wrong with SE pentodes and
tetrodes.
I did a test today to have a look at the linearity of the gm of an
EL34 with IaQ = 68mA, Ea = 350V.
Its easy to do this one in an existing amp. Just connect a link across
OPT from anode to B+.
Then have a 10R between cathode and 0V and put enough grid signal
voltage in to get signal V readings across the 10 ohms between say 0V
and 0.5Vrms. This indicates a peak current reading of +/- 0.70mA, and
this can be analysed for THD.
I got 7% THD of mainly 2H when Ia cuts off, and 2.2% at the half
voltage way between 0V cut off onset.

The gm at around the Q point was about 10.7mA/V, and quite healthy.
The only way to avoid high signal THD in pentodes is to make the Ea
swing high as possibe and Ia swing low as possible to have the load
line traverse a smallas psossible change in gm during the Ia swing.
Its rather difficult to do because we expect an SE tube to have a
linear Ia change between 0mA and twice the IaQ.

In practice with a load and changing anode load voltage with fixed
Eg2, the gm seems to vary more thaqn my experiment indicates and
create more harmonics. So the only way to get rid of the harmonics is
to add current FB or voltage FB loops.
Current FB in an OP stage where the Rk would have to be at least 1/10
of RL to achieve anything much means that the Ra would become much
higher. Typical EL35 pentode at above conditons might be 20k, and with
500 ohms for Rk unbypassed Ra' = 80k, and considerably more voltage FB
must be applied to get Ra low for a good DF.
So that leaves just the voltage FB, and lots must be applied to get
the pentode harmonics to be as blame free as a real triode.
OR, one might make a non linear cathode bypass circuit using a PNP
mosfet source follower powered by an SE transistor so that a sample of
the output load voltage generates a cathode correction voltage which
has complementary current behavoir to the pentode. Its far too much
trouble though of course and pigs might fly before it'd work properly
without adding more side effect shyte to a signal.

But my revised SE35 amps with 4 x EL34 have very nice THD character
with very little H above 2H once the load is above 4 ohms. At 4W, THD
< 0.1% above 4.5 ohms, and dipping to 0.008% at 10 ohms.
What I found hard to explain was that with 2 ohms the effect of adding
the global NFB reduced THD by 1/2.56, at 5 ohms by 1/4.5, and by 16
ohms by 1/6. And this despite no change in the input-drive amp gain.
Measurements of gain and calculation of distortion after FB was done
using the formula D' = D / [1 + (A x ß) ].
This showed D' should be reduced about 1/3.5 times with 2 ohms and 1/4
times at 16 ohms.
I am not sure why the observed and measured THD results did not tally
with calculated results except for the fact that when THD becomes
highish as it is with a mismatch load like 2 ohms, then second order
products are formed in addition to the reduced THD, so they add to
what is measured. But without any GNFB, my amp makes 1% THD at 4W into
2 ohms. The correct centre value load for max PO is about 4.5 ohms. I
don't know any possible reason for the the higher than calculated THD
reduction where the load is higher the centre value load at say 16
ohms.
I doubt my voltemeter is at fault because I'm always watching for
measurement errors and the inconsistencies in observations they cause.
Probably, using a simulator program would be totally pointless unless
the program had a very accurate model of the pentode or beam tetrode
and their complex harmonic production.

Anyway, despite the slight incomprehensible BS served up to me by
these EL34 tubes, their overall kitchen preparation of edible music is
OK.

Patrick Turner.


>
> Andre Jute
> Visit Jute on Bicycles at
> *http://www.audio-talk.co.uk/fiultra/BICYCLE%20%26%20CYCLING.html

[Andre Jute[_2_]]

On Apr 5, 9:23*am, Patrick Turner > wrote:
> On Apr 4, 10:18*am, Andre Jute > wrote:
>
> > The triumph of single-ended can be seen even in bikes:
> > *http://www.cannondale.com/onbike
>
> I huess you have seen the latest thing from the mighty BMW company - a
> bike with a shaft drive to the rear wheel in a swinging-sprung-damped
> tube.
>
> http://www.bmw.com./newbike/uppenhillen

That URL leads to an error, even after correction by removing the stop
after com.

> Meanwhile, the most serious single ended cyclist is Micheal Milton...
>
> http://www.michaelmilton.com/default.aspx?d=457759

Inspiring.

> And meanwhile, we might ask what's wrong with SE pentodes and
> tetrodes.

Eh? I don't have to ask. I've built SE EL34 for donkey's years. Sweet
little amp. My fave cheapskate amp is an SE EL34. You can see my
design here:
-- http://www.audio-talk.co.uk/fiultra/Jute-EL34-SEntry.jpg --
and here are the loudspeakers I designed to go with it
-- http://www.audio-talk.co.uk/fiultra/KISS%20195%20The%20Impresario.jpg
--

Been there, done that, published the schemo while Jesus was a
teenager.

Andre Jute
Visit Jute on Amps at
http://www.audio-talk.co.uk/fiultra/
"wonderfully well written and reasoned information for the tube audio
constructor"
John Broskie TubeCAD & GlassWare
"an unbelievably comprehensive web site containing vital gems of
wisdom"
Stuart Perry Hi-Fi News & Record Review


> I did a test today to have a look at the linearity of the gm of an
> EL34 with IaQ = 68mA, Ea = 350V.
> Its easy to do this one in an existing amp. Just connect a link across
> OPT from anode to B+.
> Then have a 10R between cathode and 0V and put enough grid signal
> voltage in to get signal V readings across the 10 ohms between say 0V
> and 0.5Vrms. This indicates a peak current reading of +/- 0.70mA, and
> this can be analysed for THD.
> I got 7% THD of mainly 2H when Ia cuts off, and 2.2% at the half
> voltage way between 0V cut off onset.
>
> The gm at around the Q point was about 10.7mA/V, and quite healthy.
> The only way to avoid high signal THD in pentodes is to make the Ea
> swing high as possibe and Ia swing low as possible to have the load
> line traverse a smallas psossible change in gm during the Ia swing.
> Its rather difficult to do because we expect an SE tube to have a
> linear Ia change between 0mA and twice the IaQ.
>
> In practice with a load and changing anode load voltage with fixed
> Eg2, the gm seems to vary more thaqn my experiment indicates and
> create more harmonics. So the only way to get rid of the harmonics is
> to add current FB or voltage FB loops.
> Current FB in an OP stage where the Rk would have to be at least 1/10
> of RL to achieve anything much means that the Ra would become much
> higher. Typical EL35 pentode at above conditons might be 20k, and with
> 500 ohms for Rk unbypassed Ra' = 80k, and considerably more voltage FB
> must be applied to get Ra low for a good DF.
> So that leaves just the voltage FB, and lots must be applied to get
> the pentode harmonics to be as blame free as a real triode.
> OR, one might make a non linear cathode bypass circuit using a PNP
> mosfet source follower powered by an SE transistor so that a sample of
> the output load voltage generates a cathode correction voltage which
> has complementary current behavoir to the pentode. Its far too much
> trouble though of course and pigs might fly before it'd work properly
> without adding more side effect shyte to a signal.
>
> But my revised SE35 amps with 4 x EL34 have very nice THD character
> with very little H above 2H once the load is above 4 ohms. At 4W, THD
> < 0.1% above 4.5 ohms, and dipping to 0.008% at 10 ohms.
> What I found hard to explain was that with 2 ohms the effect of adding
> the global NFB reduced THD by 1/2.56, at 5 ohms by 1/4.5, and by 16
> ohms by 1/6. And this despite no change in the input-drive amp gain.
> Measurements of gain and calculation of distortion after FB was done
> using the formula D' = D / [1 + (A x ß) ].
> This showed D' should be reduced about 1/3.5 times with 2 ohms and 1/4
> times at 16 ohms.
> I am not sure why the observed and measured THD results did not tally
> with calculated results except for the fact that when THD becomes
> highish as it is with a mismatch load like 2 ohms, then second order
> products are formed in addition to the reduced THD, so they add to
> what is measured. But without any GNFB, my amp makes 1% THD at 4W into
> 2 ohms. The correct centre value load for max PO is about 4.5 ohms. I
> don't know any possible reason for the the higher than calculated THD
> reduction where the load is higher the centre value load at say 16
> ohms.
> I doubt my voltemeter is at fault because I'm always watching for
> measurement errors and the inconsistencies in observations they cause.
> Probably, using a simulator program would be totally pointless unless
> the program had a very accurate model of the pentode or beam tetrode
> and their complex harmonic production.
>
> Anyway, despite the slight incomprehensible BS served up to me by
> these EL34 tubes, their overall kitchen preparation of edible music is
> OK.
>
> Patrick Turner.
>
>
>
>
>
> > Andre Jute
> > Visit Jute on Bicycles at
> > *http://www.audio-talk.co.uk/fiultra/BICYCLE%20%26%20CYCLING.html

[Patrick Turner]

On Apr 8, 6:40*am, Andre Jute > wrote:
> On Apr 5, 9:23*am, Patrick Turner > wrote:
>
> > On Apr 4, 10:18*am, Andre Jute > wrote:
>
> > > The triumph of single-ended can be seen even in bikes:
> > > *http://www.cannondale.com/onbike
>
> > I huess you have seen the latest thing from the mighty BMW company - a
> > bike with a shaft drive to the rear wheel in a swinging-sprung-damped
> > tube.
>
> >http://www.bmw.com./newbike/uppenhillen
>
> That URL leads to an error, even after correction by removing the stop
> after com.

Well I kinda thought you might have troubles ;-)........
>
> > Meanwhile, the most serious single ended cyclist is Micheal Milton...
>
> >http://www.michaelmilton.com/default.aspx?d=457759
>
> Inspiring.

I've been on a ride with 20 people one sunday and he was in the bunch.
There's a 4km climb up Mt Stromlo here at about 5% average slope and
he was quite strong. If both my legs were as good as just one of his
I'd win the TDF quite easily.
>
> > And meanwhile, we might ask what's wrong with SE pentodes and
> > tetrodes.
>
> Eh? I don't have to ask. I've built SE EL34 for donkey's years. Sweet
> little amp. My fave cheapskate amp is an SE EL34. You can see my
> design here:
> --http://www.audio-talk.co.uk/fiultra/Jute-EL34-SEntry.jpg--

That's about the simplest amplifier anyone could build.

I use a single EL34 in triode for my kitchen radio. Sounds good to me.
Speaker is one 12" Rola deluxe from 1953 and a 1974 dome tweeter from
1974, maybe 96dB efficient.
Reflex box about 3CuFt works fine and bass seems to make the whole
kitchen vibrate.

> and here are the loudspeakers I designed to go with it
> --http://www.audio-talk.co.uk/fiultra/KISS%20195%20The%20Impresario.jpg
> --
>
> Been there, done that, published the schemo while Jesus was a
> teenager.

Don't you mean Motzart? Jesus wasn't really interested in music afawk.
Of course I love many many things Jesus said and but he's all too
complex a being for me especially when we have to think about the
Father, Holy Ghost, Virgin births and Rising from the Dead and all
that heavy mystical religious spiritual stuff. All Jesus had to say
was "be good man" and if you are good, then you have a good life apart
from being kicked around by Fate and ****ed up bigtime by Lady Luck.

Basically, Jesus did say "be good man" and he reputably backed it all
up with miracles. Perhaps Jesus would have easily understood your
schematic. He should have because we were taught he was God amoung
other things. Probably he would have looked at you and drawn you aside
from the crowd and said,
"Now listen here Jutey, your'e a bit early tryin' to tempt me with
triodes and I haven't gorn upstairs yet to arrange a God Of Triodes
who will be inventing the laws of electronics so if yer wouldn't mind,
keep that there schemo under yer hat and whatever ya do, don't show
the darn thing to the Romans!"

Patrick Turner.

>
> Andre Jute
> Visit Jute on Amps at
> *http://www.audio-talk.co.uk/fiultra/
> "wonderfully well written and reasoned information for the tube audio
> constructor"
> John Broskie TubeCAD & GlassWare
> "an unbelievably comprehensive web site containing vital gems of
> wisdom"
> Stuart Perry Hi-Fi News & Record Review
>
>
>
> > I did a test today to have a look at the linearity of the gm of an
> > EL34 with IaQ = 68mA, Ea = 350V.
> > Its easy to do this one in an existing amp. Just connect a link across
> > OPT from anode to B+.
> > Then have a 10R between cathode and 0V and put enough grid signal
> > voltage in to get signal V readings across the 10 ohms between say 0V
> > and 0.5Vrms. This indicates a peak current reading of +/- 0.70mA, and
> > this can be analysed for THD.
> > I got 7% THD of mainly 2H when Ia cuts off, and 2.2% at the half
> > voltage way between 0V cut off onset.
>
> > The gm at around the Q point was about 10.7mA/V, and quite healthy.
> > The only way to avoid high signal THD in pentodes is to make the Ea
> > swing high as possibe and Ia swing low as possible to have the load
> > line traverse a smallas psossible change in gm during the Ia swing.
> > Its rather difficult to do because we expect an SE tube to have a
> > linear Ia change between 0mA and twice the IaQ.
>
> > In practice with a load and changing anode load voltage with fixed
> > Eg2, the gm seems to vary more thaqn my experiment indicates and
> > create more harmonics. So the only way to get rid of the harmonics is
> > to add current FB or voltage FB loops.
> > Current FB in an OP stage where the Rk would have to be at least 1/10
> > of RL to achieve anything much means that the Ra would become much
> > higher. Typical EL35 pentode at above conditons might be 20k, and with
> > 500 ohms for Rk unbypassed Ra' = 80k, and considerably more voltage FB
> > must be applied to get Ra low for a good DF.
> > So that leaves just the voltage FB, and lots must be applied to get
> > the pentode harmonics to be as blame free as a real triode.
> > OR, one might make a non linear cathode bypass circuit using a PNP
> > mosfet source follower powered by an SE transistor so that a sample of
> > the output load voltage generates a cathode correction voltage which
> > has complementary current behavoir to the pentode. Its far too much
> > trouble though of course and pigs might fly before it'd work properly
> > without adding more side effect shyte to a signal.
>
> > But my revised SE35 amps with 4 x EL34 have very nice THD character
> > with very little H above 2H once the load is above 4 ohms. At 4W, THD
> > < 0.1% above 4.5 ohms, and dipping to 0.008% at 10 ohms.
> > What I found hard to explain was that with 2 ohms the effect of adding
> > the global NFB reduced THD by 1/2.56, at 5 ohms by 1/4.5, and by 16
> > ohms by 1/6. And this despite no change in the input-drive amp gain.
> > Measurements of gain and calculation of distortion after FB was done
> > using the formula D' = D / [1 + (A x ß) ].
> > This showed D' should be reduced about 1/3.5 times with 2 ohms and 1/4
> > times at 16 ohms.
> > I am not sure why the observed and measured THD results did not tally
> > with calculated results except for the fact that when THD becomes
> > highish as it is with a mismatch load like 2 ohms, then second order
> > products are formed in addition to the reduced THD, so they add to
> > what is measured. But without any GNFB, my amp makes 1% THD at 4W into
> > 2 ohms. The correct centre value load for max PO is about 4.5 ohms. I
> > don't know any possible reason for the the higher than calculated THD
> > reduction where the load is higher the centre value load at say 16
> > ohms.
> > I doubt my voltemeter is at fault because I'm always watching for
> > measurement errors and the inconsistencies in observations they cause.
> > Probably, using a simulator program would be totally pointless unless
> > the program had a very accurate model of the pentode or beam tetrode
> > and their complex harmonic production.
>
> > Anyway, despite the slight incomprehensible BS served up to me by
> > these EL34 tubes, their overall kitchen preparation of edible music is
> > OK.
>
> > Patrick Turner.
>
> > > Andre Jute
> > > Visit Jute on Bicycles at
> > > *http://www.audio-talk.co.uk/fiultra/BICYCLE%20%26%20CYCLING.html- Hide quoted text -
>
> - Show quoted text -