I just posted a schematic of a new way to
bias class AB tube amp output stages at
ABSE.
The idea uses a pair of transistors to shunt
the higher portion of non linear signal currents which
will charge up the cathode bypass caps used on
cathode biased amps.
The cathode voltage will remain constant with the right
choice of transistor shunt device, and value of base R,
Thus the circuit will have a constant cathode bias voltage, like
fixed bias, up to clipping.
Without such AC regulation of the DC, the rise in cathode voltage
can be up to twice the normal quiescent level, which reduces po, and
increases
thd enormously. This method avoids both problems of cathode bias,
but retains the simplicity and self regulation of cathode bias under
idle conditions.

Under conditions of normal music use, there is some small continual
drift
of the cathode voltages as the cathode currents change to try to
maintain equilibrium.

These drifts are caused by transients causing non linear charge currents
to alter
the cathode voltage level.
The use of the AC transistor bypass arrangements reduce this drift.
Note the 1k and 0.02 uF network into the base of the transistors.
This limits base current, and any F above 10 kHz has less effect
than LF, so if the amp gets fed some RF, the caps will charge up,
biasing the tubes off,
and protecting the tubes.

The 750 ohm cathode resistors still need to be connected to a
DC detector, so that excessive cathode current will trip
the protect circuits.
A doubling of idle DC in an output tube will raise the idle voltage on
the
10 ohms from say 0.5v to 1v, which barely starts to turn on the
transistors,
which have a non linear turn on character.
The AC flow from 200 peak volts into a load of 2 kohms at the anode
will produce a peak current of 0.1 amps, resulting in a rise of peak
voltage across the 10 ohms from 0.5v to say 1.5 volts.
But the instantaneous rise in 10 ohm voltage is limited by the shunting
of the
cathode current excess during the positive voltage peaks of the input
voltage.

I measured the thd, and it was similar to using fixed bias with a sine
wave.

A class A2 or AB2 could use a similar bias arrangement.

Patrick Turner.

How's this for a challenge, Patrick, world ?

Give me a response in the hungry rat post

will discuss *any* aspect of that circuit there

got you hooked ?
Rudy

"Patrick Turner" wrote in message
...
: I just posted a schematic of a new way to
: bias class AB tube amp output stages at
: ABSE.
: The idea uses a pair of transistors to shunt
: the higher portion of non linear signal currents which
: will charge up the cathode bypass caps used on
: cathode biased amps.
: The cathode voltage will remain constant with the right
: choice of transistor shunt device, and value of base R,
: Thus the circuit will have a constant cathode bias voltage, like
: fixed bias, up to clipping.
: Without such AC regulation of the DC, the rise in cathode voltage
: can be up to twice the normal quiescent level, which reduces po, and
: increases
: thd enormously. This method avoids both problems of cathode bias,
: but retains the simplicity and self regulation of cathode bias under
: idle conditions.
:
: Under conditions of normal music use, there is some small continual
: drift
: of the cathode voltages as the cathode currents change to try to
: maintain equilibrium.
:
: These drifts are caused by transients causing non linear charge currents
: to alter
: the cathode voltage level.
: The use of the AC transistor bypass arrangements reduce this drift.
: Note the 1k and 0.02 uF network into the base of the transistors.
: This limits base current, and any F above 10 kHz has less effect
: than LF, so if the amp gets fed some RF, the caps will charge up,
: biasing the tubes off,
: and protecting the tubes.
:
: The 750 ohm cathode resistors still need to be connected to a
: DC detector, so that excessive cathode current will trip
: the protect circuits.
: A doubling of idle DC in an output tube will raise the idle voltage on
: the
: 10 ohms from say 0.5v to 1v, which barely starts to turn on the
: transistors,
: which have a non linear turn on character.
: The AC flow from 200 peak volts into a load of 2 kohms at the anode
: will produce a peak current of 0.1 amps, resulting in a rise of peak
: voltage across the 10 ohms from 0.5v to say 1.5 volts.
: But the instantaneous rise in 10 ohm voltage is limited by the shunting
: of the
: cathode current excess during the positive voltage peaks of the input
: voltage.
:
: I measured the thd, and it was similar to using fixed bias with a sine
: wave.
:
: A class A2 or AB2 could use a similar bias arrangement.
:
: Patrick Turner.
:
:
:
:

"Patrick Turner" wrote in message
...
I just posted a schematic of a new way to
bias class AB tube amp output stages at
ABSE.
The idea uses a pair of transistors to shunt
the higher portion of non linear signal currents which
will charge up the cathode bypass caps used on
cathode biased amps.
The cathode voltage will remain constant with the right
choice of transistor shunt device, and value of base R,
Thus the circuit will have a constant cathode bias voltage, like
fixed bias, up to clipping.
Without such AC regulation of the DC, the rise in cathode voltage
can be up to twice the normal quiescent level, which reduces po, and
increases
thd enormously. This method avoids both problems of cathode bias,
but retains the simplicity and self regulation of cathode bias under
idle conditions.

Under conditions of normal music use, there is some small continual
drift
of the cathode voltages as the cathode currents change to try to
maintain equilibrium.

These drifts are caused by transients causing non linear charge currents
to alter
the cathode voltage level.
The use of the AC transistor bypass arrangements reduce this drift.
Note the 1k and 0.02 uF network into the base of the transistors.
This limits base current, and any F above 10 kHz has less effect
than LF, so if the amp gets fed some RF, the caps will charge up,
biasing the tubes off,
and protecting the tubes.

The 750 ohm cathode resistors still need to be connected to a
DC detector, so that excessive cathode current will trip
the protect circuits.
A doubling of idle DC in an output tube will raise the idle voltage on
the
10 ohms from say 0.5v to 1v, which barely starts to turn on the
transistors,
which have a non linear turn on character.
The AC flow from 200 peak volts into a load of 2 kohms at the anode
will produce a peak current of 0.1 amps, resulting in a rise of peak
voltage across the 10 ohms from 0.5v to say 1.5 volts.
But the instantaneous rise in 10 ohm voltage is limited by the shunting
of the
cathode current excess during the positive voltage peaks of the input
voltage.

I measured the thd, and it was similar to using fixed bias with a sine
wave.

A class A2 or AB2 could use a similar bias arrangement.

Patrick Turner.




Hiya, Patrick.
Hobby Boy here.. What does the 0.47uF link do between the HT and the two
transistors ?
regards
jim

jim wrote:

"Patrick Turner" wrote in message
...
I just posted a schematic of a new way to
bias class AB tube amp output stages at
ABSE.
The idea uses a pair of transistors to shunt
the higher portion of non linear signal currents which
will charge up the cathode bypass caps used on
cathode biased amps.
The cathode voltage will remain constant with the right
choice of transistor shunt device, and value of base R,
Thus the circuit will have a constant cathode bias voltage, like
fixed bias, up to clipping.
Without such AC regulation of the DC, the rise in cathode voltage
can be up to twice the normal quiescent level, which reduces po, and
increases
thd enormously. This method avoids both problems of cathode bias,
but retains the simplicity and self regulation of cathode bias under
idle conditions.

Under conditions of normal music use, there is some small continual
drift
of the cathode voltages as the cathode currents change to try to
maintain equilibrium.

These drifts are caused by transients causing non linear charge currents
to alter
the cathode voltage level.
The use of the AC transistor bypass arrangements reduce this drift.
Note the 1k and 0.02 uF network into the base of the transistors.
This limits base current, and any F above 10 kHz has less effect
than LF, so if the amp gets fed some RF, the caps will charge up,
biasing the tubes off,
and protecting the tubes.

The 750 ohm cathode resistors still need to be connected to a
DC detector, so that excessive cathode current will trip
the protect circuits.
A doubling of idle DC in an output tube will raise the idle voltage on
the
10 ohms from say 0.5v to 1v, which barely starts to turn on the
transistors,
which have a non linear turn on character.
The AC flow from 200 peak volts into a load of 2 kohms at the anode
will produce a peak current of 0.1 amps, resulting in a rise of peak
voltage across the 10 ohms from 0.5v to say 1.5 volts.
But the instantaneous rise in 10 ohm voltage is limited by the shunting
of the
cathode current excess during the positive voltage peaks of the input
voltage.

I measured the thd, and it was similar to using fixed bias with a sine
wave.

A class A2 or AB2 could use a similar bias arrangement.

Patrick Turner.




Hiya, Patrick.
Hobby Boy here.. What does the 0.47uF link do between the HT and the two
transistors ?
regards
jim

Its a bit tricky, that part.
But its normal good practice to place a polypropylene
cap between the CT on the OPT, and 0V,
which is what is shown on my schematic.

Patrick Turner.

"Patrick Turner" wrote in message
...


jim wrote:

"Patrick Turner" wrote in message
...
I just posted a schematic of a new way to
bias class AB tube amp output stages at
ABSE.
The idea uses a pair of transistors to shunt
the higher portion of non linear signal currents which
will charge up the cathode bypass caps used on
cathode biased amps.
The cathode voltage will remain constant with the right
choice of transistor shunt device, and value of base R,
Thus the circuit will have a constant cathode bias voltage, like
fixed bias, up to clipping.
Without such AC regulation of the DC, the rise in cathode voltage
can be up to twice the normal quiescent level, which reduces po, and
increases
thd enormously. This method avoids both problems of cathode bias,
but retains the simplicity and self regulation of cathode bias under
idle conditions.

Under conditions of normal music use, there is some small continual
drift
of the cathode voltages as the cathode currents change to try to
maintain equilibrium.

These drifts are caused by transients causing non linear charge
currents
to alter
the cathode voltage level.
The use of the AC transistor bypass arrangements reduce this drift.
Note the 1k and 0.02 uF network into the base of the transistors.
This limits base current, and any F above 10 kHz has less effect
than LF, so if the amp gets fed some RF, the caps will charge up,
biasing the tubes off,
and protecting the tubes.

The 750 ohm cathode resistors still need to be connected to a
DC detector, so that excessive cathode current will trip
the protect circuits.
A doubling of idle DC in an output tube will raise the idle voltage on
the
10 ohms from say 0.5v to 1v, which barely starts to turn on the
transistors,
which have a non linear turn on character.
The AC flow from 200 peak volts into a load of 2 kohms at the anode
will produce a peak current of 0.1 amps, resulting in a rise of peak
voltage across the 10 ohms from 0.5v to say 1.5 volts.
But the instantaneous rise in 10 ohm voltage is limited by the
shunting
of the
cathode current excess during the positive voltage peaks of the input
voltage.

I measured the thd, and it was similar to using fixed bias with a sine
wave.

A class A2 or AB2 could use a similar bias arrangement.

Patrick Turner.




Hiya, Patrick.
Hobby Boy here.. What does the 0.47uF link do between the HT and the
two
transistors ?
regards
jim

Its a bit tricky, that part.
But its normal good practice to place a polypropylene
cap between the CT on the OPT, and 0V,
which is what is shown on my schematic.

Patrick Turner.



Yeah, OK, it's actually across HT to earth. Shunts the HT line
electrolytics at HF.. Nice thoughts, Patrick. The technology moves on a bit.
kind regards
jim

Yeah, OK, it's actually across HT to earth. Shunts the HT line
electrolytics at HF.. Nice thoughts, Patrick. The technology moves on a bit.
kind regards
jim

Good to see the crew on HMS RAT is still awake,
and not sailing backwards across the Irish Sea.

Patrick Turner.

"Patrick Turner" wrote in message
...


Yeah, OK, it's actually across HT to earth. Shunts the HT line
electrolytics at HF.. Nice thoughts, Patrick. The technology moves on a
bit.
kind regards
jim

Good to see the crew on HMS RAT is still awake,
and not sailing backwards across the Irish Sea.

Patrick Turner.



I've just been on to Julie Andrews.... Gave him the word about Neddy and
HMS Pianola... You remember ???? Don't try and squirm out of that one,
.....you....you ..... Australian person !!!

You still owe me a ****in' piano or 72 AUSD compensation.... !!!

Went down in the Indian Ocean with Neddy and both hands.. ?

Julie has let Phooly Jullison know and, together, they will now be
assaulting you with endangering shipping, loss of our national heritage,
fraud, deception, being Australian, and wearing brown shoes at the
weekend...

You deserve everything you get !!!

jim

jim wrote:

"Patrick Turner" wrote in message
...


Yeah, OK, it's actually across HT to earth. Shunts the HT line
electrolytics at HF.. Nice thoughts, Patrick. The technology moves on a
bit.
kind regards
jim

Good to see the crew on HMS RAT is still awake,
and not sailing backwards across the Irish Sea.

Patrick Turner.



I've just been on to Julie Andrews.... Gave him the word about Neddy and
HMS Pianola... You remember ???? Don't try and squirm out of that one,
....you....you ..... Australian person !!!

You still owe me a ****in' piano or 72 AUSD compensation.... !!!

Went down in the Indian Ocean with Neddy and both hands.. ?

Julie has let Phooly Jullison know and, together, they will now be
assaulting you with endangering shipping, loss of our national heritage,
fraud, deception, being Australian, and wearing brown shoes at the
weekend...

You deserve everything you get !!!

jim

C'mon man, give me a break,
why we let youse win the world Rugby Cup,
and youse ought to be grateful and happy, and feelin fully compensated,
and now yer winging about old debts, pianos, etc, etc;
bloody hell.....

Patrick Turner.

"Patrick Turner" wrote in message
...


jim wrote:

"Patrick Turner" wrote in message
...


Yeah, OK, it's actually across HT to earth. Shunts the HT line
electrolytics at HF.. Nice thoughts, Patrick. The technology moves
on a
bit.
kind regards
jim

Good to see the crew on HMS RAT is still awake,
and not sailing backwards across the Irish Sea.

Patrick Turner.



I've just been on to Julie Andrews.... Gave him the word about Neddy
and
HMS Pianola... You remember ???? Don't try and squirm out of that one,
....you....you ..... Australian person !!!

You still owe me a ****in' piano or 72 AUSD compensation.... !!!

Went down in the Indian Ocean with Neddy and both hands.. ?

Julie has let Phooly Jullison know and, together, they will now be
assaulting you with endangering shipping, loss of our national
heritage,
fraud, deception, being Australian, and wearing brown shoes at the
weekend...

You deserve everything you get !!!

jim

C'mon man, give me a break,
why we let youse win the world Rugby Cup,

Err, sorry, what ?? You mean 15 big, gruff, British boys came over and
kicked the **** out of your home-grown Nancy Boys !!!!! Didn't disturb
their hairstyles or cosmetics did we ?? Sorry if we spoilt somebody's lip
gloss or mascara !!!

and youse ought to be grateful and happy, and feelin fully compensated,
and now yer winging about old debts, pianos, etc, etc;
bloody hell.....

Wait till I tell Julie Phooly and Andrew Anderson !!!!!

INCITING RACIAL HATRED !!!!!!!

You are in for it now, Turner

Regards
jim

Hi Patrick !

Interresting idea, I suggest to do not duplicate the circuit for each
cathode since with
your arragement one tube cannot "block" the other !
(Sorry, but more than 30 years engeenering in industry make me an
incorrigible bean counter !)
A single zener could do the same job ... with less tunning flexibility ..
Cheers, Yves.

"Patrick Turner" wrote in message
...
I just posted a schematic of a new way to
bias class AB tube amp output stages at
ABSE.
The idea uses a pair of transistors to shunt
the higher portion of non linear signal currents which
will charge up the cathode bypass caps used on
cathode biased amps.
The cathode voltage will remain constant with the right
choice of transistor shunt device, and value of base R,
Thus the circuit will have a constant cathode bias voltage, like
fixed bias, up to clipping.
Without such AC regulation of the DC, the rise in cathode voltage
can be up to twice the normal quiescent level, which reduces po, and
increases
thd enormously. This method avoids both problems of cathode bias,
but retains the simplicity and self regulation of cathode bias under
idle conditions.

Under conditions of normal music use, there is some small continual
drift
of the cathode voltages as the cathode currents change to try to
maintain equilibrium.

These drifts are caused by transients causing non linear charge currents
to alter
the cathode voltage level.
The use of the AC transistor bypass arrangements reduce this drift.
Note the 1k and 0.02 uF network into the base of the transistors.
This limits base current, and any F above 10 kHz has less effect
than LF, so if the amp gets fed some RF, the caps will charge up,
biasing the tubes off,
and protecting the tubes.

The 750 ohm cathode resistors still need to be connected to a
DC detector, so that excessive cathode current will trip
the protect circuits.
A doubling of idle DC in an output tube will raise the idle voltage on
the
10 ohms from say 0.5v to 1v, which barely starts to turn on the
transistors,
which have a non linear turn on character.
The AC flow from 200 peak volts into a load of 2 kohms at the anode
will produce a peak current of 0.1 amps, resulting in a rise of peak
voltage across the 10 ohms from 0.5v to say 1.5 volts.
But the instantaneous rise in 10 ohm voltage is limited by the shunting
of the
cathode current excess during the positive voltage peaks of the input
voltage.

I measured the thd, and it was similar to using fixed bias with a sine
wave.

A class A2 or AB2 could use a similar bias arrangement.

Patrick Turner.

Earth to Mantle ... Mantle to Core ... HT to Ground ... Ground to Aurora!

: Yeah, OK, it's actually across HT to earth. Shunts the HT line
: electrolytics at HF.. Nice thoughts, Patrick. The technology moves on a bit.
: kind regards
: jim
:
: Good to see the crew on HMS RAT is still awake,
: and not sailing backwards across the Irish Sea.
:
: Patrick Turner.

Cue Gilbert & Sullivan .... with Neddy Seagoon on sinking pianoforte ...

A plea !
Can I actually get a copy of this diagram, which hasn't yet as ideogram,
crossed the Tasman Sea, so far as I can see? Thanks, if so, from me;-)
(tee-hee-hee-hee, la-doh!)

RdM

[partaking of a sacrificial lamb appres midnight feast, and reminded of:]
"I love to cook with wine. Sometimes I even put it in the food."
[taking in BBC World ... I'm OK now ...no worries:=]

Primitiv Dict
AURORA:
Noun: Aurora [aurorae]
1. Goddess of the dawn; counterpart of Greek Eos
Noun: Aurora [aurorae]
1. The first light of day
2. An atmospheric phenomenon consisting of bands of light caused by
charged solar particles following the earth's magnetic lines of force

3. New superweapons utilising Tesla's technology tested in latter century?

Yves Monmagnon wrote:

Hi Patrick !

Interresting idea, I suggest to do not duplicate the circuit for each
cathode since with
your arragement one tube cannot "block" the other !
(Sorry, but more than 30 years engeenering in industry make me an
incorrigible bean counter !)
A single zener could do the same job ... with less tunning flexibility ..
Cheers, Yves.

Nope, the zener idea don't work nearly as well as this one.
Yes, two transistors are required, no boubt about it.

I have tried the idea of using zeners as at
http://www.turneraudio.com.au/htmlwe...evisedst70.htm

The problem with the zeners is that you must allow some headroom so the zeners
are not
conducting when the tube is in the idle condition, or else its just fixed
bias.
The zener voltage has to be above the idle voltage of the cathode, about
5 volts is about right.
The zener circuit stops the cathode voltage rising above 5 volts above the
idle voltage.
Meanwhile, with music, the cathode voltage don't rise very much at all,
but it does wander up and down a bit like all cathode bias amps
due to transients, and probably to even order harmonics in the
large LF signals.

The use of transistors with their low voltage range of around
0.3 to 0.7 volts for full turn on means that the AC signal in the 10 ohms
can be used to bypass the large 2H currents through the transistor,
rather than allow them to charge up the cathode cap like a rectifier diode in
a power
supply.
At low levels of signal, when the
output tubes are in class A, there is still 2H present in the current flow of
the cathode,
increasing from near nothing, to a considerable amount at the cut off point
and above.
The normal bias current might be 50 mA.
OK, so you have 0.5v DV at the 10 ohms at idle.
The class A limit is reached when the current change in the tube is
+/- 50 mA. above this point its class AB.
The voltage signal is +/- 0.5volts either side of 0.5v,
so the peak voltage even while the amp is in class A is +1v,
and when the transistor is guided to begin conduction.
Its hfe varies during its turn on range, and hence its input Z is
a non linear value of input impedance, and so the transistor does some work
even when the amp is in class A, and gone over say 10 watts of power.
Music isn't a single sine wave.
Its lots of combined sine waves of varying amplitude,
and its somewhat difficult to express what is there
with music unless you couldd understand a formula covering
50 pages.
The transistors don't have to do much at 5 watts,
but at signals rising transiently above
5 watts, the transistors start working, and they limit the positive
peak voltage rise in the 10 ohms, because the rise in voltage
dynamically turns on the transistor, and shunts the very current
causing the voltage rise in the 10 ohms.
The caps are spared the charge up currents.
The 10 ohm R also have the effect of providing current FB
in the output stage, and they also limit the charge current flowing
into and out of the 1,000 uF.
The wave form on the 10 ohm resistors is quite distorted normally,
due to cut off in AB, and instead of the usual flat bottomed wave with
arched peaks, the waves look like a compressed sine wave
of much lower amplitude than with no transistors present when the transistors
are
working at high AB power.
But at class A conditions, the positive voltage peaks at the 10 ohm R
are just gently compressed, and thd is not increased in an un-benign manner.

The 10 ohm R acts to increase the Ra of the tube by U x Rk,
and in the case of KT88, or some UL octal tube,
its as much as 250 ohms. In a triode amp it'd be around 50 ohms.
In either case, the increase is negligible,
But the current sensing R of 10 ohms doesn't want to be
very large, since large Rk values give rise to
little thd reductions.

It would be possible to use 1 ohm for Rk,
and have a low impedance transformer to couple the
transistor base, so the clipped signal in the current sensing
won't harm the signal, even at high outputs.
The transformer coupled transistor could *never*
be turned on due to excessive DC tube flow.

A transformer of 1:1 could be used for each 10 ohm R
as shown, but i found a simple R did the trick, to feed the transistor.

The transitors don't want to be over excited by to much bass current.
If the base is turned on too hard by a low value of series base R,
the amp tries to lower its own cathode voltage as input signal is increased.
In other words, it'd be like having some device which
reduces the bias voltage, and increases the fixed bias current with increasing
signal.
This could be dangerous in an AB amp, since you just don't want the tubes to
be subject to increased dissipation too much.

Identifying the correct Rb value is done by
plotting the graph of Ek above the idle condition against increasing sine wave
signal.
The Ek should rise slightly from idle, to about 2volts above idle at clipping
onset.

The 47 ohms collector load R is there to limit the current through
the transistor.
Its peak rated current of this particular BD239B is 2 amps,
in a TO220 package.
So if the idle voltage Ek is 50v, then the current max is 50v.
Many various power transistors with a Vce rating of at least 100v could be
used,
maybe with an Imax of 5 amps.
You won't blow up such transistors.
I just tried 2SD426 in a pair of 300 watt amps,
with 6 x 6550 each side of the PP circuit.
It worked fine, and saves the use of 12 pots for bias adjustment
on each tube.

In my latest cathode FB circuit version of the 300 watt amps I also
have an Rk of half the usual value for cathode bias R, and I have an applied
fixed bias of -17v to each grid. Screens are fed from a regulated fixed 400v.
The Rk are 500 ohms each, and Ek at 23v, so Ia is fairly low,
yet the first 100 watts with a sine wave remains at low thd.
The lower than usual Rk still provides enough R to self regulate the
anode current in the tube, because the output tubes are
operating as cathode followers feeding Rk at extremely low F, or at DV.
With a regulated screen voltage, rise and fall of mains supplies causes
negligible change in anode currents, because at DV, the effective
Ra is the official data Ra plus U x Rk, and with 6550, that'd be
18k plus U x 0.5k, which is heaps.
The use of cathode bypass caps with organic salts made by Nichicon,
Panasonic, Roedestien is recommended.
I like *large* value bypass caps, but its easy to find
63 volt rated 1,000 uF caps.
Some folks use 4,700 uF.

With the active AC regulation of the cathode bias, the thd
profile is the same as with a fixed bias amp.
The main body of the thd when the amp travels into AB is due
to the cut off behaviour of the tubes, something
one cannot avoid, and only minimise with careful selection of electrode
supply voltages, which should be maintained at a constant level
until clipping, when its probably a good thing that screen voltages can sag
easily,
preventing fried screens.

Triode operation gives low power, but the triode cut off is less sharp,
but the transistor limiters still work well.

Patrick Turner.






"Patrick Turner" wrote in message
...
I just posted a schematic of a new way to
bias class AB tube amp output stages at
ABSE.
The idea uses a pair of transistors to shunt
the higher portion of non linear signal currents which
will charge up the cathode bypass caps used on
cathode biased amps.
The cathode voltage will remain constant with the right
choice of transistor shunt device, and value of base R,
Thus the circuit will have a constant cathode bias voltage, like
fixed bias, up to clipping.
Without such AC regulation of the DC, the rise in cathode voltage
can be up to twice the normal quiescent level, which reduces po, and
increases
thd enormously. This method avoids both problems of cathode bias,
but retains the simplicity and self regulation of cathode bias under
idle conditions.

Under conditions of normal music use, there is some small continual
drift
of the cathode voltages as the cathode currents change to try to
maintain equilibrium.

These drifts are caused by transients causing non linear charge currents
to alter
the cathode voltage level.
The use of the AC transistor bypass arrangements reduce this drift.
Note the 1k and 0.02 uF network into the base of the transistors.
This limits base current, and any F above 10 kHz has less effect
than LF, so if the amp gets fed some RF, the caps will charge up,
biasing the tubes off,
and protecting the tubes.

The 750 ohm cathode resistors still need to be connected to a
DC detector, so that excessive cathode current will trip
the protect circuits.
A doubling of idle DC in an output tube will raise the idle voltage on
the
10 ohms from say 0.5v to 1v, which barely starts to turn on the
transistors,
which have a non linear turn on character.
The AC flow from 200 peak volts into a load of 2 kohms at the anode
will produce a peak current of 0.1 amps, resulting in a rise of peak
voltage across the 10 ohms from 0.5v to say 1.5 volts.
But the instantaneous rise in 10 ohm voltage is limited by the shunting
of the
cathode current excess during the positive voltage peaks of the input
voltage.

I measured the thd, and it was similar to using fixed bias with a sine
wave.

A class A2 or AB2 could use a similar bias arrangement.

Patrick Turner.

"jim" wrote in message
...
:
: "Patrick Turner" wrote in message
: ...
:
:
: Yeah, OK, it's actually across HT to earth. Shunts the HT line
: electrolytics at HF.. Nice thoughts, Patrick. The technology moves on
a
: bit.
: kind regards
: jim
:
: Good to see the crew on HMS RAT is still awake,
: and not sailing backwards across the Irish Sea.
:
: Patrick Turner.
:
:
:
: I've just been on to Julie Andrews.... Gave him the word about Neddy and
: HMS Pianola... You remember ???? Don't try and squirm out of that one,
: ....you....you ..... Australian person !!!
:
: You still owe me a ****in' piano or 72 AUSD compensation.... !!!
:
: Went down in the Indian Ocean with Neddy and both hands.. ?
:
: Julie has let Phooly Jullison know and, together, they will now be
: assaulting you with endangering shipping, loss of our national heritage,
: fraud, deception, being Australian, and wearing brown shoes at the
: weekend...
:
: You deserve everything you get !!!
:
: jim
:
hehe,
well you know what is thought about your "true" origin
in Helsinki... some of the dutch are now also having 2nd thoughts..
great stuff there,
loved it
Rudy

What schematic I cant find it?



"Patrick Turner" wrote in message
...


Yves Monmagnon wrote:

Hi Patrick !

Interresting idea, I suggest to do not duplicate the circuit for each
cathode since with
your arragement one tube cannot "block" the other !
(Sorry, but more than 30 years engeenering in industry make me an
incorrigible bean counter !)
A single zener could do the same job ... with less tunning flexibility
...
Cheers, Yves.

Nope, the zener idea don't work nearly as well as this one.
Yes, two transistors are required, no boubt about it.

I have tried the idea of using zeners as at
http://www.turneraudio.com.au/htmlwe...evisedst70.htm

The problem with the zeners is that you must allow some headroom so the
zeners
are not
conducting when the tube is in the idle condition, or else its just fixed
bias.
The zener voltage has to be above the idle voltage of the cathode, about
5 volts is about right.
The zener circuit stops the cathode voltage rising above 5 volts above the
idle voltage.
Meanwhile, with music, the cathode voltage don't rise very much at all,
but it does wander up and down a bit like all cathode bias amps
due to transients, and probably to even order harmonics in the
large LF signals.

The use of transistors with their low voltage range of around
0.3 to 0.7 volts for full turn on means that the AC signal in the 10 ohms
can be used to bypass the large 2H currents through the transistor,
rather than allow them to charge up the cathode cap like a rectifier diode
in
a power
supply.
At low levels of signal, when the
output tubes are in class A, there is still 2H present in the current flow
of
the cathode,
increasing from near nothing, to a considerable amount at the cut off
point
and above.
The normal bias current might be 50 mA.
OK, so you have 0.5v DV at the 10 ohms at idle.
The class A limit is reached when the current change in the tube is
+/- 50 mA. above this point its class AB.
The voltage signal is +/- 0.5volts either side of 0.5v,
so the peak voltage even while the amp is in class A is +1v,
and when the transistor is guided to begin conduction.
Its hfe varies during its turn on range, and hence its input Z is
a non linear value of input impedance, and so the transistor does some
work
even when the amp is in class A, and gone over say 10 watts of power.
Music isn't a single sine wave.
Its lots of combined sine waves of varying amplitude,
and its somewhat difficult to express what is there
with music unless you couldd understand a formula covering
50 pages.
The transistors don't have to do much at 5 watts,
but at signals rising transiently above
5 watts, the transistors start working, and they limit the positive
peak voltage rise in the 10 ohms, because the rise in voltage
dynamically turns on the transistor, and shunts the very current
causing the voltage rise in the 10 ohms.
The caps are spared the charge up currents.
The 10 ohm R also have the effect of providing current FB
in the output stage, and they also limit the charge current flowing
into and out of the 1,000 uF.
The wave form on the 10 ohm resistors is quite distorted normally,
due to cut off in AB, and instead of the usual flat bottomed wave with
arched peaks, the waves look like a compressed sine wave
of much lower amplitude than with no transistors present when the
transistors
are
working at high AB power.
But at class A conditions, the positive voltage peaks at the 10 ohm R
are just gently compressed, and thd is not increased in an un-benign
manner.

The 10 ohm R acts to increase the Ra of the tube by U x Rk,
and in the case of KT88, or some UL octal tube,
its as much as 250 ohms. In a triode amp it'd be around 50 ohms.
In either case, the increase is negligible,
But the current sensing R of 10 ohms doesn't want to be
very large, since large Rk values give rise to
little thd reductions.

It would be possible to use 1 ohm for Rk,
and have a low impedance transformer to couple the
transistor base, so the clipped signal in the current sensing
won't harm the signal, even at high outputs.
The transformer coupled transistor could *never*
be turned on due to excessive DC tube flow.

A transformer of 1:1 could be used for each 10 ohm R
as shown, but i found a simple R did the trick, to feed the transistor.

The transitors don't want to be over excited by to much bass current.
If the base is turned on too hard by a low value of series base R,
the amp tries to lower its own cathode voltage as input signal is
increased.
In other words, it'd be like having some device which
reduces the bias voltage, and increases the fixed bias current with
increasing
signal.
This could be dangerous in an AB amp, since you just don't want the tubes
to
be subject to increased dissipation too much.

Identifying the correct Rb value is done by
plotting the graph of Ek above the idle condition against increasing sine
wave
signal.
The Ek should rise slightly from idle, to about 2volts above idle at
clipping
onset.

The 47 ohms collector load R is there to limit the current through
the transistor.
Its peak rated current of this particular BD239B is 2 amps,
in a TO220 package.
So if the idle voltage Ek is 50v, then the current max is 50v.
Many various power transistors with a Vce rating of at least 100v could be
used,
maybe with an Imax of 5 amps.
You won't blow up such transistors.
I just tried 2SD426 in a pair of 300 watt amps,
with 6 x 6550 each side of the PP circuit.
It worked fine, and saves the use of 12 pots for bias adjustment
on each tube.

In my latest cathode FB circuit version of the 300 watt amps I also
have an Rk of half the usual value for cathode bias R, and I have an
applied
fixed bias of -17v to each grid. Screens are fed from a regulated fixed
400v.
The Rk are 500 ohms each, and Ek at 23v, so Ia is fairly low,
yet the first 100 watts with a sine wave remains at low thd.
The lower than usual Rk still provides enough R to self regulate the
anode current in the tube, because the output tubes are
operating as cathode followers feeding Rk at extremely low F, or at DV.
With a regulated screen voltage, rise and fall of mains supplies causes
negligible change in anode currents, because at DV, the effective
Ra is the official data Ra plus U x Rk, and with 6550, that'd be
18k plus U x 0.5k, which is heaps.
The use of cathode bypass caps with organic salts made by Nichicon,
Panasonic, Roedestien is recommended.
I like *large* value bypass caps, but its easy to find
63 volt rated 1,000 uF caps.
Some folks use 4,700 uF.

With the active AC regulation of the cathode bias, the thd
profile is the same as with a fixed bias amp.
The main body of the thd when the amp travels into AB is due
to the cut off behaviour of the tubes, something
one cannot avoid, and only minimise with careful selection of electrode
supply voltages, which should be maintained at a constant level
until clipping, when its probably a good thing that screen voltages can
sag
easily,
preventing fried screens.

Triode operation gives low power, but the triode cut off is less sharp,
but the transistor limiters still work well.

Patrick Turner.






"Patrick Turner" wrote in message
...
I just posted a schematic of a new way to
bias class AB tube amp output stages at
ABSE.
The idea uses a pair of transistors to shunt
the higher portion of non linear signal currents which
will charge up the cathode bypass caps used on
cathode biased amps.
The cathode voltage will remain constant with the right
choice of transistor shunt device, and value of base R,
Thus the circuit will have a constant cathode bias voltage, like
fixed bias, up to clipping.
Without such AC regulation of the DC, the rise in cathode voltage
can be up to twice the normal quiescent level, which reduces po, and
increases
thd enormously. This method avoids both problems of cathode bias,
but retains the simplicity and self regulation of cathode bias under
idle conditions.

Under conditions of normal music use, there is some small continual
drift
of the cathode voltages as the cathode currents change to try to
maintain equilibrium.

These drifts are caused by transients causing non linear charge
currents
to alter
the cathode voltage level.
The use of the AC transistor bypass arrangements reduce this drift.
Note the 1k and 0.02 uF network into the base of the transistors.
This limits base current, and any F above 10 kHz has less effect
than LF, so if the amp gets fed some RF, the caps will charge up,
biasing the tubes off,
and protecting the tubes.

The 750 ohm cathode resistors still need to be connected to a
DC detector, so that excessive cathode current will trip
the protect circuits.
A doubling of idle DC in an output tube will raise the idle voltage on
the
10 ohms from say 0.5v to 1v, which barely starts to turn on the
transistors,
which have a non linear turn on character.
The AC flow from 200 peak volts into a load of 2 kohms at the anode
will produce a peak current of 0.1 amps, resulting in a rise of peak
voltage across the 10 ohms from 0.5v to say 1.5 volts.
But the instantaneous rise in 10 ohm voltage is limited by the
shunting
of the
cathode current excess during the positive voltage peaks of the input
voltage.

I measured the thd, and it was similar to using fixed bias with a sine
wave.

A class A2 or AB2 could use a similar bias arrangement.

Patrick Turner.