[Patrick Turner]

John Stewart mentioned........

Altho several successful vacuum tube amplifiers incorporating a
bootstrapped driver have been built & marketed, there are some of us
still concerned with the positive feedback used. Turns out, not much
occurs.

But could it lead to instability?

There is a simple method to calculate the maximum this positive
feedback
would be before even picking up a soldering iron. Anyone know how?

Cheers, John


I don't know the formula which must apply to work out how much loop
PFB can be applied before instability will occur within a wanted
bandwidth of an amplifier.

In tube amps bootstrapping is used by a number of makers from the
past, McIntosh, EAR, and others.
In solid state amps it was used very commonly to boost the gain of the
VAS amp before the emitter follower output stage.

John's schematic at the link shows two LTPs in cascade with what might
be 6AS7 output tubes.
The schematic has minimal information and I suggest John re-post his
schematic with ALL maximum operating signal voltages at -1dB below
clipping so that everyone will be informed rather than bored stiff or
confused.

Let us assume the grid signal max at each output tube anode is
200Vrms, and at the taps on the OPT shown 1/2 way along the anode
winding it is 100Vrms, and that output tube grid signal is 120Vrms
because a 6AS7 has a gain of less than 2.0 which is the µ of the 6AS7.

The point is, for tubes like 6AS7, a rather high grid drive signal is
needed. From what I said above the signal voltage across the The 27k
between 6SN7 and OPT is 20Vrms, so Iac = 0.74mArms, so at the 6SN7
whch may be producing 120Vrms the load ohms is 120V / 0.074 = 162k.
This is in parallel to the bias cap coupled biasing R of 270k, so
total RL for the 6SN7 = 101k.
Therefore the gain of the 6SN7 is quite high, = 20 x 101 / ( 101 +
10 ) = 18.2, approx.
Therefore the THD of the 6SN7 could be expected to be lower than if
there was no bootstrapping and the dc to the 6SN7 anode was brought
via resistances only to a B+ supply, because such R would make the
anode load for 6SN7 be rather low, maybe 30k, and for 120Vmrs, that's
4mArms, and idle
current in 6SN7 has to be high, and so the B+ would have to be higher
than available in John's schematic.

I have tried many experiments with bootstrapping, and found that in a
circuit like John's there is only a very marginal reduction of THD due
to bootstrapping because the PFB involved increases THD which works
against the lowering of THD because of the raising of RL. The
principle reason for bootstrapping is to allow a wide driver stage
voltage swing. I found that bootstrapping can also reduce margins of
stability, sliightly reduce BW, and increase Rout of the amp. But
these bad effects of PFB are not too bad where the OP stage has almost
no gain, like John's, or the McIntosh etc.
The global NFB counter the downside of bootstrapping.

If one Includes a voltage doubler in the B+ part of the power supply
it is easy to get a B+rail of say +700Vdc and from this you feed the
6SN7 anodes with 5mA for say +250Vdc at the anodes and Rdc = 90k, and
the total load for each 6SN7 triode = 68k, and this is about 6 x Ra,
and load line analysis will show you should get 120Vrms - WITHOUT PFB
FROM BOOTSTRAPPING.

I have a plain UL amp with 6 x 6CA7 OP tubes driven with a 6CG7 on
each side of an LTP and with the elevated B+ supply, and the THD
production of the driver stage is about 1/3 of what every other half
baked designer has achieved.

I don't like bootstrapping drivers as it is done in the circuits
discussed. I also don't much like needing to provide a high B+ just
for the driver stage.

Far better is to use the driver stage as shown at
http://www.turneraudio.com.au/300w-1...tput-jan06.htm

There is a nice big fat page which fully explains everything about the
drive amp shown.

With the bootstrapping as shown in John's schematic, any THD voltage
at the OPT is divided by the R divider formed by 27k and the Ra of the
6SN7 triode, about 10k, so +D at OPT becomes +0.27D at the 6SN7 and
this is amplified by say 1.5 to appear as 0.4D at the 6AS7 anode, with
0.2D adding to what already appears at the OPT tap from where the
bootstrap 27k connects. The effective increase in THD via the
bootstrapping is probably about +6dB, there is an sunstantial
increase, and the whole idea of feeding back THD to add to itself,
send it all around the loop again, and then add global NFB to
counteract the effect is like taking 3 steps backward while also
making 5 steps forward at the same to get 2 steps ahead.
Using my choke+resistance loading & DC supply to the driver tubes and
using 6BQ5/EL84 in triode makes for a far better driver so you get 5
steps ahead without any backward steps.
The EL84 in triode gives Ra = 2k2, about 1/4 that of 6SN7. The grid
bias resistors of the outpt stage may be reduced in value to better
control the bias voltage. 270k is too high if the output tube is 1/2
way through its service life and has begun to get the "positive grid
disease" where one notices that there may be a volt or two Vdc across
the biasing R. I like to keep Rg bias R to 120k or less for all large
octal tubes.

Patrick Turner.

[John L Stewart]

Originally Posted by Patrick Turner

John Stewart mentioned........

Altho several successful vacuum tube amplifiers incorporating a
bootstrapped driver have been built & marketed, there are some of us
still concerned with the positive feedback used. Turns out, not much
occurs.

But could it lead to instability?

There is a simple method to calculate the maximum this positive
feedback
would be before even picking up a soldering iron. Anyone know how?

Cheers, John


I don't know the formula which must apply to work out how much loop
PFB can be applied before instability will occur within a wanted
bandwidth of an amplifier.

In tube amps bootstrapping is used by a number of makers from the
past, McIntosh, EAR, and others.
In solid state amps it was used very commonly to boost the gain of the
VAS amp before the emitter follower output stage.

John's schematic at the link shows two LTPs in cascade with what might
be 6AS7 output tubes.
The schematic has minimal information and I suggest John re-post his
schematic with ALL maximum operating signal voltages at -1dB below
clipping so that everyone will be informed rather than bored stiff or
confused.

Let us assume the grid signal max at each output tube anode is
200Vrms, and at the taps on the OPT shown 1/2 way along the anode
winding it is 100Vrms, and that output tube grid signal is 120Vrms
because a 6AS7 has a gain of less than 2.0 which is the µ of the 6AS7.

The point is, for tubes like 6AS7, a rather high grid drive signal is
needed. From what I said above the signal voltage across the The 27k
between 6SN7 and OPT is 20Vrms, so Iac = 0.74mArms, so at the 6SN7
whch may be producing 120Vrms the load ohms is 120V / 0.074 = 162k.
This is in parallel to the bias cap coupled biasing R of 270k, so
total RL for the 6SN7 = 101k.
Therefore the gain of the 6SN7 is quite high, = 20 x 101 / ( 101 +
10 ) = 18.2, approx.
Therefore the THD of the 6SN7 could be expected to be lower than if
there was no bootstrapping and the dc to the 6SN7 anode was brought
via resistances only to a B+ supply, because such R would make the
anode load for 6SN7 be rather low, maybe 30k, and for 120Vmrs, that's
4mArms, and idle
current in 6SN7 has to be high, and so the B+ would have to be higher
than available in John's schematic.

I have tried many experiments with bootstrapping, and found that in a
circuit like John's there is only a very marginal reduction of THD due
to bootstrapping because the PFB involved increases THD which works
against the lowering of THD because of the raising of RL. The
principle reason for bootstrapping is to allow a wide driver stage
voltage swing. I found that bootstrapping can also reduce margins of
stability, sliightly reduce BW, and increase Rout of the amp. But
these bad effects of PFB are not too bad where the OP stage has almost
no gain, like John's, or the McIntosh etc.
The global NFB counter the downside of bootstrapping.

If one Includes a voltage doubler in the B+ part of the power supply
it is easy to get a B+rail of say +700Vdc and from this you feed the
6SN7 anodes with 5mA for say +250Vdc at the anodes and Rdc = 90k, and
the total load for each 6SN7 triode = 68k, and this is about 6 x Ra,
and load line analysis will show you should get 120Vrms - WITHOUT PFB
FROM BOOTSTRAPPING.

I have a plain UL amp with 6 x 6CA7 OP tubes driven with a 6CG7 on
each side of an LTP and with the elevated B+ supply, and the THD
production of the driver stage is about 1/3 of what every other half
baked designer has achieved.

I don't like bootstrapping drivers as it is done in the circuits
discussed. I also don't much like needing to provide a high B+ just
for the driver stage.

Far better is to use the driver stage as shown at
http://www.turneraudio.com.au/300w-1...tput-jan06.htm

There is a nice big fat page which fully explains everything about the
drive amp shown.

With the bootstrapping as shown in John's schematic, any THD voltage
at the OPT is divided by the R divider formed by 27k and the Ra of the
6SN7 triode, about 10k, so +D at OPT becomes +0.27D at the 6SN7 and
this is amplified by say 1.5 to appear as 0.4D at the 6AS7 anode, with
0.2D adding to what already appears at the OPT tap from where the
bootstrap 27k connects. The effective increase in THD via the
bootstrapping is probably about +6dB, there is an sunstantial
increase, and the whole idea of feeding back THD to add to itself,
send it all around the loop again, and then add global NFB to
counteract the effect is like taking 3 steps backward while also
making 5 steps forward at the same to get 2 steps ahead.
Using my choke+resistance loading & DC supply to the driver tubes and
using 6BQ5/EL84 in triode makes for a far better driver so you get 5
steps ahead without any backward steps.
The EL84 in triode gives Ra = 2k2, about 1/4 that of 6SN7. The grid
bias resistors of the outpt stage may be reduced in value to better
control the bias voltage. 270k is too high if the output tube is 1/2
way through its service life and has begun to get the "positive grid
disease" where one notices that there may be a volt or two Vdc across
the biasing R. I like to keep Rg bias R to 120k or less for all large
octal tubes.

Patrick Turner.

This response is an example of how things can quickly go off the rails when we don't read someones post carefully. I didn't mention full loop feedback anywhere.

All of the information required to lead to a solution is given the schema. If one needs to know the voltages they are easily calculated from the supply voltages, resitors & tubes used.

I will leave this open for now. Perhaps Phil A or others may take a run at the simple solution asked for in my original post. And if you like Patrick, look carefully & go for it!

And here is another schema for all to puzzle over.

Cheers, John

[John L Stewart]

[quote=Alex;928172]I would set the taps at 1/mu (1/2 in this case), so that 6SN7s run at
constant plate current. In this case, even if you pull 6SN7 out of the
socket, the output stage will not turn into a nice multivibrator. It will be
just at the brink of instability.

Another funny feature of this circuit: Why do they waste 130V on automatic
bias whilst having minus 150V rail at their disposal?

Alex

The 6AS7 /6080 family does not like fixed bias at all, not recommended by the manufacturers. That does create a problem where lots of heat from the cathode resistors needs to be dissipated, a real problem if inside the chassis.

There are are number of SS auto bias circuits published as follows-

Valve Amplifiers, Morgan Jones
Priciples Of Power, Kevin O'Connor
& others.

Each cathode would need one of these circuits, so lots of complication. If I try another I will get 25W resistors that can be connected thru the chassis to a heat sink as we often see with power transistors. Trying to keep things simple.

I did try a common cathode resistor at one point. Then I watched as the 6080 burned a hole thru one of the grids. Thermal instabilty, just like in SS.

Cheers, John

[Alex]

I would set the taps at 1/mu (1/2 in this case), so that 6SN7s run at
constant plate current. In this case, even if you pull 6SN7 out of the
socket, the output stage will not turn into a nice multivibrator. It will be
just at the brink of instability.

Another funny feature of this circuit: Why do they waste 130V on automatic
bias whilst having minus 150V rail at their disposal?

Alex


"Patrick Turner" wrote in message
...
John Stewart mentioned........

Altho several successful vacuum tube amplifiers incorporating a
bootstrapped driver have been built & marketed, there are some of us
still concerned with the positive feedback used. Turns out, not much
occurs.

But could it lead to instability?

There is a simple method to calculate the maximum this positive
feedback
would be before even picking up a soldering iron. Anyone know how?

Cheers, John


I don't know the formula which must apply to work out how much loop
PFB can be applied before instability will occur within a wanted
bandwidth of an amplifier.

In tube amps bootstrapping is used by a number of makers from the
past, McIntosh, EAR, and others.
In solid state amps it was used very commonly to boost the gain of the
VAS amp before the emitter follower output stage.

John's schematic at the link shows two LTPs in cascade with what might
be 6AS7 output tubes.
The schematic has minimal information and I suggest John re-post his
schematic with ALL maximum operating signal voltages at -1dB below
clipping so that everyone will be informed rather than bored stiff or
confused.

Let us assume the grid signal max at each output tube anode is
200Vrms, and at the taps on the OPT shown 1/2 way along the anode
winding it is 100Vrms, and that output tube grid signal is 120Vrms
because a 6AS7 has a gain of less than 2.0 which is the µ of the 6AS7.

The point is, for tubes like 6AS7, a rather high grid drive signal is
needed. From what I said above the signal voltage across the The 27k
between 6SN7 and OPT is 20Vrms, so Iac = 0.74mArms, so at the 6SN7
whch may be producing 120Vrms the load ohms is 120V / 0.074 = 162k.
This is in parallel to the bias cap coupled biasing R of 270k, so
total RL for the 6SN7 = 101k.
Therefore the gain of the 6SN7 is quite high, = 20 x 101 / ( 101 +
10 ) = 18.2, approx.
Therefore the THD of the 6SN7 could be expected to be lower than if
there was no bootstrapping and the dc to the 6SN7 anode was brought
via resistances only to a B+ supply, because such R would make the
anode load for 6SN7 be rather low, maybe 30k, and for 120Vmrs, that's
4mArms, and idle
current in 6SN7 has to be high, and so the B+ would have to be higher
than available in John's schematic.

I have tried many experiments with bootstrapping, and found that in a
circuit like John's there is only a very marginal reduction of THD due
to bootstrapping because the PFB involved increases THD which works
against the lowering of THD because of the raising of RL. The
principle reason for bootstrapping is to allow a wide driver stage
voltage swing. I found that bootstrapping can also reduce margins of
stability, sliightly reduce BW, and increase Rout of the amp. But
these bad effects of PFB are not too bad where the OP stage has almost
no gain, like John's, or the McIntosh etc.
The global NFB counter the downside of bootstrapping.

If one Includes a voltage doubler in the B+ part of the power supply
it is easy to get a B+rail of say +700Vdc and from this you feed the
6SN7 anodes with 5mA for say +250Vdc at the anodes and Rdc = 90k, and
the total load for each 6SN7 triode = 68k, and this is about 6 x Ra,
and load line analysis will show you should get 120Vrms - WITHOUT PFB
FROM BOOTSTRAPPING.

I have a plain UL amp with 6 x 6CA7 OP tubes driven with a 6CG7 on
each side of an LTP and with the elevated B+ supply, and the THD
production of the driver stage is about 1/3 of what every other half
baked designer has achieved.

I don't like bootstrapping drivers as it is done in the circuits
discussed. I also don't much like needing to provide a high B+ just
for the driver stage.

Far better is to use the driver stage as shown at
http://www.turneraudio.com.au/300w-1...tput-jan06.htm

There is a nice big fat page which fully explains everything about the
drive amp shown.

With the bootstrapping as shown in John's schematic, any THD voltage
at the OPT is divided by the R divider formed by 27k and the Ra of the
6SN7 triode, about 10k, so +D at OPT becomes +0.27D at the 6SN7 and
this is amplified by say 1.5 to appear as 0.4D at the 6AS7 anode, with
0.2D adding to what already appears at the OPT tap from where the
bootstrap 27k connects. The effective increase in THD via the
bootstrapping is probably about +6dB, there is an sunstantial
increase, and the whole idea of feeding back THD to add to itself,
send it all around the loop again, and then add global NFB to
counteract the effect is like taking 3 steps backward while also
making 5 steps forward at the same to get 2 steps ahead.
Using my choke+resistance loading & DC supply to the driver tubes and
using 6BQ5/EL84 in triode makes for a far better driver so you get 5
steps ahead without any backward steps.
The EL84 in triode gives Ra = 2k2, about 1/4 that of 6SN7. The grid
bias resistors of the outpt stage may be reduced in value to better
control the bias voltage. 270k is too high if the output tube is 1/2
way through its service life and has begun to get the "positive grid
disease" where one notices that there may be a volt or two Vdc across
the biasing R. I like to keep Rg bias R to 120k or less for all large
octal tubes.

Patrick Turner.

[``ZACK``]

"John L Stewart" wrote in message
...

Patrick Turner;928123 Wrote:
John Stewart mentioned........

Altho several successful vacuum tube amplifiers incorporating a
bootstrapped driver have been built & marketed, there are some of us
still concerned with the positive feedback used. Turns out, not much
occurs.

But could it lead to instability?

There is a simple method to calculate the maximum this positive
feedback
would be before even picking up a soldering iron. Anyone know how?

Cheers, John


I don't know the formula which must apply to work out how much loop
PFB can be applied before instability will occur within a wanted
bandwidth of an amplifier.

In tube amps bootstrapping is used by a number of makers from the
past, McIntosh, EAR, and others.
In solid state amps it was used very commonly to boost the gain of the
VAS amp before the emitter follower output stage.

John's schematic at the link shows two LTPs in cascade with what might
be 6AS7 output tubes.
The schematic has minimal information and I suggest John re-post his
schematic with ALL maximum operating signal voltages at -1dB below
clipping so that everyone will be informed rather than bored stiff or
confused.

Let us assume the grid signal max at each output tube anode is
200Vrms, and at the taps on the OPT shown 1/2 way along the anode
winding it is 100Vrms, and that output tube grid signal is 120Vrms
because a 6AS7 has a gain of less than 2.0 which is the µ of the 6AS7.

The point is, for tubes like 6AS7, a rather high grid drive signal is
needed. From what I said above the signal voltage across the The 27k
between 6SN7 and OPT is 20Vrms, so Iac = 0.74mArms, so at the 6SN7
whch may be producing 120Vrms the load ohms is 120V / 0.074 = 162k.
This is in parallel to the bias cap coupled biasing R of 270k, so
total RL for the 6SN7 = 101k.
Therefore the gain of the 6SN7 is quite high, = 20 x 101 / ( 101 +
10 ) = 18.2, approx.
Therefore the THD of the 6SN7 could be expected to be lower than if
there was no bootstrapping and the dc to the 6SN7 anode was brought
via resistances only to a B+ supply, because such R would make the
anode load for 6SN7 be rather low, maybe 30k, and for 120Vmrs, that's
4mArms, and idle
current in 6SN7 has to be high, and so the B+ would have to be higher
than available in John's schematic.

I have tried many experiments with bootstrapping, and found that in a
circuit like John's there is only a very marginal reduction of THD due
to bootstrapping because the PFB involved increases THD which works
against the lowering of THD because of the raising of RL. The
principle reason for bootstrapping is to allow a wide driver stage
voltage swing. I found that bootstrapping can also reduce margins of
stability, sliightly reduce BW, and increase Rout of the amp. But
these bad effects of PFB are not too bad where the OP stage has almost
no gain, like John's, or the McIntosh etc.
The global NFB counter the downside of bootstrapping.

If one Includes a voltage doubler in the B+ part of the power supply
it is easy to get a B+rail of say +700Vdc and from this you feed the
6SN7 anodes with 5mA for say +250Vdc at the anodes and Rdc = 90k, and
the total load for each 6SN7 triode = 68k, and this is about 6 x Ra,
and load line analysis will show you should get 120Vrms - WITHOUT PFB
FROM BOOTSTRAPPING.

I have a plain UL amp with 6 x 6CA7 OP tubes driven with a 6CG7 on
each side of an LTP and with the elevated B+ supply, and the THD
production of the driver stage is about 1/3 of what every other half
baked designer has achieved.

I don't like bootstrapping drivers as it is done in the circuits
discussed. I also don't much like needing to provide a high B+ just
for the driver stage.

Far better is to use the driver stage as shown at
http://tinyurl.com/4dkgywm

There is a nice big fat page which fully explains everything about the
drive amp shown.

With the bootstrapping as shown in John's schematic, any THD voltage
at the OPT is divided by the R divider formed by 27k and the Ra of the
6SN7 triode, about 10k, so +D at OPT becomes +0.27D at the 6SN7 and
this is amplified by say 1.5 to appear as 0.4D at the 6AS7 anode, with
0.2D adding to what already appears at the OPT tap from where the
bootstrap 27k connects. The effective increase in THD via the
bootstrapping is probably about +6dB, there is an sunstantial
increase, and the whole idea of feeding back THD to add to itself,
send it all around the loop again, and then add global NFB to
counteract the effect is like taking 3 steps backward while also
making 5 steps forward at the same to get 2 steps ahead.
Using my choke+resistance loading & DC supply to the driver tubes and
using 6BQ5/EL84 in triode makes for a far better driver so you get 5
steps ahead without any backward steps.
The EL84 in triode gives Ra = 2k2, about 1/4 that of 6SN7. The grid
bias resistors of the outpt stage may be reduced in value to better
control the bias voltage. 270k is too high if the output tube is 1/2
way through its service life and has begun to get the "positive grid
disease" where one notices that there may be a volt or two Vdc across
the biasing R. I like to keep Rg bias R to 120k or less for all large
octal tubes.

Patrick Turner.

This response is an example of how things can quickly go off the rails
when we don't read someones post carefully. I didn't mention full loop
feedback anywhere.

All of the information required to lead to a solution is given the
schema. If one needs to know the voltages they are easily calculated
from the supply voltages, resitors & tubes used.

I will leave this open for now. Perhaps Phil A or others may take a run
at the simple solution asked for in my original post. And if you like
Patrick, look carefully & go for it!

And here is another schema for all to puzzle over.

Cheers, John

have a look at this site, scroll down
you will see the driver you want.
http://www.lenardaudio.com/education...ve_amps_3.html

[John L Stewart]

[ This response is an example of how things can quickly go off the rails
when we don't read someones post carefully. I didn't mention full loop
feedback anywhere.

All of the information required to lead to a solution is given the
schema. If one needs to know the voltages they are easily calculated
from the supply voltages, resitors & tubes used.

I will leave this open for now. Perhaps Phil A or others may take a run
at the simple solution asked for in my original post. And if you like
Patrick, look carefully & go for it!

And here is another schema for all to puzzle over.

Cheers, John

[/i][/color]
have a look at this site, scroll down
you will see the driver you want.
http://www.lenardaudio.com/education...ve_amps_3.html[/quote]

----------------------------------------------------------------

Hey Zack, that is a slick circuit mod but gets you only half way there for those OP stages requiring large grid swings. That would include McIntosh, Circlotron, Norman Crowhursts Twin Coupled Amp & that 6AS7/6080 Amp I built. In all, the CF needs to be bootstrapped as well. Otherwise needs a somewhat higher plate supply.

But otherwise nice to drive KTnn, 6L6nn, 7027, ETC, you pick your favourite.

Cheers, John