Splitted from this thread.
Darius Loftin White
Hi Sheldon,
thanks for your interest. Please let us make it step by step.
First, imagine that C2 has infinity capacitance no losses and it is charged.
In this case there will be no signal voltage drop at the cathode resistor.
Thus there is no signal feedback caused by "Chb".
In this fundamental schematic
the green line shows the signal current, the red line shows the signal path.
The left circuit is no Loftin White, it is simply direct coupling.
The right circuit is my triode Loftin White topology.
Kind regards,
Darius
Darius Loftin White
Hi Sheldon,
thanks for your interest. Please let us make it step by step.
First, imagine that C2 has infinity capacitance no losses and it is charged.
In this case there will be no signal voltage drop at the cathode resistor.
Thus there is no signal feedback caused by "Chb".
In this fundamental schematic
the green line shows the signal current, the red line shows the signal path.
The left circuit is no Loftin White, it is simply direct coupling.
The right circuit is my triode Loftin White topology.
Kind regards,
Darius
LW cycles through only one cap, instead of down to ground and back
up though a bypass cap in series. Thus the one cap may be smaller?
Or did I miss the whole point?
And if I may imagine a "blue highlighted path" between stages,
might loop up from cathode to plate of the input triode down
through grid to cathode of the output triode, and back through
your CHB path, what is that, "Cathode High Frequency Bypass"?
Not sure at 28K+, why would CHB be significant compared to
the path 4.7K to GND and back up the 560ohm... Again, I may
have missed the point.
up though a bypass cap in series. Thus the one cap may be smaller?
Or did I miss the whole point?
And if I may imagine a "blue highlighted path" between stages,
might loop up from cathode to plate of the input triode down
through grid to cathode of the output triode, and back through
your CHB path, what is that, "Cathode High Frequency Bypass"?
Not sure at 28K+, why would CHB be significant compared to
the path 4.7K to GND and back up the 560ohm... Again, I may
have missed the point.
#2 C2 --> infinit
Hi,
what does LW cycles mean and what is a one cap?
This amp is an audio frequency amp not radio frequency.
What does your "blue path" mean? An additional current loop?
There is no current going from the grid to the cathode.
The grid is negative biased against the cathode.
Since no signal drops at the cathode of the output triode
no signal is on the Chb path. Thus Chb has no influence in
the signal.
Kind regards,
Darius
Hi,
what does LW cycles mean and what is a one cap?
This amp is an audio frequency amp not radio frequency.
What does your "blue path" mean? An additional current loop?
There is no current going from the grid to the cathode.
The grid is negative biased against the cathode.
Since no signal drops at the cathode of the output triode
no signal is on the Chb path. Thus Chb has no influence in
the signal.
Kind regards,
Darius
oldeurope said:
Ah, OK. The green AC current path means that the signal current should not be seen by the cathode resistor, assuming a perfect capacitor for C2. So, the AC connection to the input grid is only correcting for an imperfect C2 capacitor? Thanks for your patience.
Sheldon
Edit: I notice in your 300b design, that you have added a noise cancellation circuit - nice. Have you ever checked out this site?: http://members.aol.com/sbench101 He also likes to explore unusual topologies.
Re: Re: Re: Darius Loftin White explained
Ah, yes. I didn't see the English version, and I didn't spot those. I have a version of Steve's 801 resistor coupled amp. Sweet little amp. If I get ambitious, I may wind up some output transformers and try it as a L/W amp.
Sheldon
oldeurope said:
Ah, yes. I didn't see the English version, and I didn't spot those. I have a version of Steve's 801 resistor coupled amp. Sweet little amp. If I get ambitious, I may wind up some output transformers and try it as a L/W amp.
Sheldon
Re: #2 C2 --> infinit
LW was supposed to read "Loftin White"
One cap, C2 - 24uF.
Cycle meaning a loop, like the green one you drew.
Audio, not Radio? Would never have figured that out.
Imaginary Blue Path as I fully described it, not highlighted
in your original drawing. Relevant only to asking a question.
Intending to describe a loop of interstage signal voltage
drops, summing back to unity as it returns to the orgin.
No current from Grid to Cathode? No, of course not.
I did not assume current when you draw a Red line
from Grid to Plate, Grid to Plate, then across an OPT
into the Speaker. If current loop makes no sense, why
assume I am asking about a current loop?
CHB Path... Does nothing? OK, whats it there for then?
I think it must do something, but what???
I see 220K/560, voltage gain of x393 or Mu of the input
Triode, whichever of the two gain factors is the lower...
Most likely Mu is setting input gain, not the resistors.
Relevant to my next statement/question.
The Output Cathode may follow "without signal drop?"
But it is following interstage signal of at least Mu Gain.
And coupled thru CHB path back to the input cathode
follower of Unity Gain or less. These two cathodes are
not of the same AC signal level, and of inverted phase!
Again, what exactly does CHB do if not conduct an
alternating current in proportion to the AC voltage
drop across it.? (Vin*(Mu+1))/Zchb)
oldeurope said:
LW was supposed to read "Loftin White"
One cap, C2 - 24uF.
Cycle meaning a loop, like the green one you drew.
Audio, not Radio? Would never have figured that out.
Imaginary Blue Path as I fully described it, not highlighted
in your original drawing. Relevant only to asking a question.
Intending to describe a loop of interstage signal voltage
drops, summing back to unity as it returns to the orgin.
No current from Grid to Cathode? No, of course not.
I did not assume current when you draw a Red line
from Grid to Plate, Grid to Plate, then across an OPT
into the Speaker. If current loop makes no sense, why
assume I am asking about a current loop?
CHB Path... Does nothing? OK, whats it there for then?
I think it must do something, but what???
I see 220K/560, voltage gain of x393 or Mu of the input
Triode, whichever of the two gain factors is the lower...
Most likely Mu is setting input gain, not the resistors.
Relevant to my next statement/question.
The Output Cathode may follow "without signal drop?"
But it is following interstage signal of at least Mu Gain.
And coupled thru CHB path back to the input cathode
follower of Unity Gain or less. These two cathodes are
not of the same AC signal level, and of inverted phase!
Again, what exactly does CHB do if not conduct an
alternating current in proportion to the AC voltage
drop across it.? (Vin*(Mu+1))/Zchb)
I am guessing that all signals around the green output loop are
common mode to the power rail noise, including the cathode!!!
As are the rLoad and Plate of the input stage. Power hum can
have no effect in the output Triode circuit when all points are
common mode with it.
But the input plate is not common mode to the input signal, nor
to ground. And triodes will respond to plate voltage hum as an
input that modulates current gain. It cannot be ignored, and is
probably corrected in this circuit somehow, somewhere...
Perhaps CHB is for cancellation of hum in the input stage???
I see no great effect it has upon the output stage... At least
not directly...
common mode to the power rail noise, including the cathode!!!
As are the rLoad and Plate of the input stage. Power hum can
have no effect in the output Triode circuit when all points are
common mode with it.
But the input plate is not common mode to the input signal, nor
to ground. And triodes will respond to plate voltage hum as an
input that modulates current gain. It cannot be ignored, and is
probably corrected in this circuit somehow, somewhere...
Perhaps CHB is for cancellation of hum in the input stage???
I see no great effect it has upon the output stage... At least
not directly...
#8 300B hum cancellation...
Hello kenpeter,
I am sorry but I think I am not able to explain you this thing.
Absolutely no chance, I am very sorry.
Hello agent.5
In this blog you can see a conventional SE and the triode Loftin White.
The conventional SE is made sensitive for hum at the supply rail.
In this conventional SE the hum at the supply rail and the hum
caused by AC heating of the 300B are compensating each other.
(Please note that the component values are specially selected
this way that phase and level are ok for a hum cancellation.)
The Loftin White topology has an excellent supply voltage riple
rejection because riples at the positive supply rail are coupled to the cathode
of the power tube by C2. Chb makes the output of the driver stage
floating on the AC component at the cathode of the power tube.
It would be possible to generate a hum by miss adjusting P and
if phase and level are ok there would be a compensation of the hum
caused by the the double mains frequency hum caused by the low thermal
inertia. In this case the points of hum cancellation and
compensating an imperfect C2 capacitor would be different.
This is why I generated a hum signal in the power supply
and feed it to the cathode of the driver stage to compensate the hum
caused by AC heating of the 300B. Hum cancellation and compensating
an imperfect C2 are independent this way.
Have a look at the Oscilloscope screen shots.
Kind regards,
Darius
Hello kenpeter,
I am sorry but I think I am not able to explain you this thing.
Absolutely no chance, I am very sorry.
Hello agent.5
In this blog you can see a conventional SE and the triode Loftin White.
The conventional SE is made sensitive for hum at the supply rail.
In this conventional SE the hum at the supply rail and the hum
caused by AC heating of the 300B are compensating each other.
(Please note that the component values are specially selected
this way that phase and level are ok for a hum cancellation.)
The Loftin White topology has an excellent supply voltage riple
rejection because riples at the positive supply rail are coupled to the cathode
of the power tube by C2. Chb makes the output of the driver stage
floating on the AC component at the cathode of the power tube.
It would be possible to generate a hum by miss adjusting P and
if phase and level are ok there would be a compensation of the hum
caused by the the double mains frequency hum caused by the low thermal
inertia. In this case the points of hum cancellation and
compensating an imperfect C2 capacitor would be different.
This is why I generated a hum signal in the power supply
and feed it to the cathode of the driver stage to compensate the hum
caused by AC heating of the 300B. Hum cancellation and compensating
an imperfect C2 are independent this way.
Have a look at the Oscilloscope screen shots.
Kind regards,
Darius
Re: #8 300B hum cancellation...
Practically, I don't know if it matters, but this author discounts thermal inertia as the main cause of hum from AC powered DHT filaments: http://www.geocities.com/dmitrynizh/filament-ac-harmonic.htm#abstract
Sheldon
oldeurope said:
Practically, I don't know if it matters, but this author discounts thermal inertia as the main cause of hum from AC powered DHT filaments: http://www.geocities.com/dmitrynizh/filament-ac-harmonic.htm#abstract
Sheldon
My misunderstanding was that the output cathode follows the
interstage Mu. And that CHB coupled both hum and feedback.
C2 completely bypasses the output cathode back to the top rail.
There is no follower copy of interstage Mu in the presence of the
strong C2 bypass cap, only power line hum as intended...
You give up pig wrestling too easy. I was not wresting against
your efforts. Only with my own inability to describe what was
confusing to me. I never doubted you were right for a second.
interstage Mu. And that CHB coupled both hum and feedback.
C2 completely bypasses the output cathode back to the top rail.
There is no follower copy of interstage Mu in the presence of the
strong C2 bypass cap, only power line hum as intended...
You give up pig wrestling too easy. I was not wresting against
your efforts. Only with my own inability to describe what was
confusing to me. I never doubted you were right for a second.
#11
Hi Sheldon,
this is interesting.
But read the last sentences of the conclusion.
Filaments don't have comparable mass. The thinner a filament
is, the lower is the thermal inertia. One inch of a 2A3 cathode
has more mass than one inch of a 6A3 cathode.
There are other points i disagree ...
Since this hum can be cancelled, it is not that important for me
where it comes from. Maybe the truth is between both theories.
Kind regards,
Darius
Hi Sheldon,
this is interesting.
But read the last sentences of the conclusion.
Filaments don't have comparable mass. The thinner a filament
is, the lower is the thermal inertia. One inch of a 2A3 cathode
has more mass than one inch of a 6A3 cathode.
There are other points i disagree ...
Since this hum can be cancelled, it is not that important for me
where it comes from. Maybe the truth is between both theories.
Kind regards,
Darius
Re: #11
Yes, I agree. In the context of this thread it's not important. I have seen other discussions using higher frequency supplies, on the assumption that filament mass was the dominant issue. If harmonic distortion is a significant contributor, then all that would do is shift the hum frequency higher, and in the case of square waves, generate lots of high order harmonics. If the frequency is high enough, maybe it doesn't matter there either. Or maybe it does, I don't know.
Sheldon
oldeurope said:
Yes, I agree. In the context of this thread it's not important. I have seen other discussions using higher frequency supplies, on the assumption that filament mass was the dominant issue. If harmonic distortion is a significant contributor, then all that would do is shift the hum frequency higher, and in the case of square waves, generate lots of high order harmonics. If the frequency is high enough, maybe it doesn't matter there either. Or maybe it does, I don't know.
Sheldon
Darius,
You got me interested in this Loftin-White stuff, so I looked a little on the web. I find lots of schematics called Loftin-White, but none that have the output current loop as shown in the original design, or in your version. But you knew that I guess. Everyone seems to have focussed on direct coupling, but ignored the really cool part of the design - at least to my poorly trained eyes. Surprisingly, even John Broskie of TubeCad has got it wrong.
If I get to this down the road on my 801 amp, I may use a mosfet follower in place of the second triode (no room to add another, the FET's are smaller).
Sheldon
You got me interested in this Loftin-White stuff, so I looked a little on the web. I find lots of schematics called Loftin-White, but none that have the output current loop as shown in the original design, or in your version. But you knew that I guess. Everyone seems to have focussed on direct coupling, but ignored the really cool part of the design - at least to my poorly trained eyes. Surprisingly, even John Broskie of TubeCad has got it wrong.
If I get to this down the road on my 801 amp, I may use a mosfet follower in place of the second triode (no room to add another, the FET's are smaller).
Sheldon
One thread I came across when searching the term Loftin-White, is this one: http://www.vintage-radio.net/forum/showthread.php?t=20849
Now, it's clear that the one poster did not understand the difference between feedback and what is going on in this circuit. I must confess that I was confused too. Pretty interesting how that happens. I think I'm beginning to see the difference now. Feedback implies that the input signal is compared with the output and the output corrected, based on the difference. I see now that the signal is not compared to the output.
In the Loftin-White, we are looking only keeping the power supply constant. The error signal that is compensated is only this one and has nothing to do with the input signal, except as it cancels ripple in the supply, whether caused by output demands or by residual supply ripple. You are right, it's pretty darn clever.
Here's an amp I intend to convert at some point:
Now, it's clear that the one poster did not understand the difference between feedback and what is going on in this circuit. I must confess that I was confused too. Pretty interesting how that happens. I think I'm beginning to see the difference now. Feedback implies that the input signal is compared with the output and the output corrected, based on the difference. I see now that the signal is not compared to the output.
In the Loftin-White, we are looking only keeping the power supply constant. The error signal that is compensated is only this one and has nothing to do with the input signal, except as it cancels ripple in the supply, whether caused by output demands or by residual supply ripple. You are right, it's pretty darn clever.
Here's an amp I intend to convert at some point:
Attachments
- Status
- This old topic is closed. If you want to reopen this topic, contact a moderator using the "Report Post" button.