I read tube amp project articles in my down time, and have always had this little hole in my knowledge as a newbie. These articles will often describe the power supply after describing the amp circuit, usually with a separate PS schematic...
In the PS description If I see the author say something like... "You'll need a 270-0-270 transformer and with this filter circuit your B+ should come to about 315 volts"... If no other info is given... Should I assume the 315 volts number pertains to the power supply measured completely unloaded, or should I routinely make the assumption that I will have 315 volts after the supply is hooked up to the amp? In this case the amp schematic didn't have an explicit B+ value named on it, class A. Similarly when I see an amp schematic that names 250 volts as the B+, this means I have to design the PS to supply 250 volts at max expected load correct? So a supply that measures 250 with no load would not be appropriate, I'd have to make sure my supply settles at 250 with the max load, correct?
In the PS description If I see the author say something like... "You'll need a 270-0-270 transformer and with this filter circuit your B+ should come to about 315 volts"... If no other info is given... Should I assume the 315 volts number pertains to the power supply measured completely unloaded, or should I routinely make the assumption that I will have 315 volts after the supply is hooked up to the amp? In this case the amp schematic didn't have an explicit B+ value named on it, class A. Similarly when I see an amp schematic that names 250 volts as the B+, this means I have to design the PS to supply 250 volts at max expected load correct? So a supply that measures 250 with no load would not be appropriate, I'd have to make sure my supply settles at 250 with the max load, correct?
The original article or designer would typically have aimed for and tested with the stated B+ volts during idle operation, or using a stated transformer secondary voltage and any described type of rectifier/filter.
I have rarely if ever seen an article make the comment of 'do not exceed' a certain B+ level although that is a value that should be discussed as to why limits may need to be applied. Often any such limit may be related to filter and coupling capacitor voltage ratings, as they are typically the 'limit' case, although the rectifier diode PIV may also be, as well as max design centre ratings for screen or plate voltages, or that certain components like output stage output tubes operate at no more than a certain dissipation level that then relates back to B+. Certainly if a B+ level turns out to be higher than has been used for an existing amp design, then the constructor really should assess what are the limits and risks (rather than wait for collateral damage after some part fails).
Power transformers don't always come with the same secondary voltage ratings - either in the rated voltage, or the allowed regulation - so there needs to be a fair amount of leeway afforded any construction in B+ level, with the main underlying rating being that the heaters operate at their rating and within tolerance (which given the possibility of mains voltage variations, means trying to aim for <5% tolerance for nominal mains voltage imho).
Any sag in a B+ level is really up to the amp designer or constructor to vacillate over. Of course some amp designs don't impose large power supply current swings (eg. class A, or where music has only transient peak levels rather than sustained peak levels), and some power supply configurations have negligible B+ variation (eg. choke input filter style), and some amp designs don't show any change due to B+ variation due to their use of feedback to alleviate all sorts of internal amp variations and distortions.
I have rarely if ever seen an article make the comment of 'do not exceed' a certain B+ level although that is a value that should be discussed as to why limits may need to be applied. Often any such limit may be related to filter and coupling capacitor voltage ratings, as they are typically the 'limit' case, although the rectifier diode PIV may also be, as well as max design centre ratings for screen or plate voltages, or that certain components like output stage output tubes operate at no more than a certain dissipation level that then relates back to B+. Certainly if a B+ level turns out to be higher than has been used for an existing amp design, then the constructor really should assess what are the limits and risks (rather than wait for collateral damage after some part fails).
Power transformers don't always come with the same secondary voltage ratings - either in the rated voltage, or the allowed regulation - so there needs to be a fair amount of leeway afforded any construction in B+ level, with the main underlying rating being that the heaters operate at their rating and within tolerance (which given the possibility of mains voltage variations, means trying to aim for <5% tolerance for nominal mains voltage imho).
Any sag in a B+ level is really up to the amp designer or constructor to vacillate over. Of course some amp designs don't impose large power supply current swings (eg. class A, or where music has only transient peak levels rather than sustained peak levels), and some power supply configurations have negligible B+ variation (eg. choke input filter style), and some amp designs don't show any change due to B+ variation due to their use of feedback to alleviate all sorts of internal amp variations and distortions.
Last edited:
Thanks! Idle current, makes sense. Unloaded or max makes less sense now that I think of it. You may not know the max or how the amp will be used. And even the worst PS will measure correctly unloaded. Semantically I'll assume they mean at idle, makes total sense.
There's a great, free, PS design program called PSUD2. Although not mentioned on the linked page, if you go to the download page there is also a Mac version.
PSUD2
AC voltage is a sine wave. Transformer outputs sine wave to rectifier. Capacitor bank averages this rectified voltage to a DC voltage.
Transformer output * 1.414 (sine wave integral) - rectifier Vforward dropoff = B+ voltage.
The "sag" is the Vforward of the rectifier
Transformer output * 1.414 (sine wave integral) - rectifier Vforward dropoff = B+ voltage.
The "sag" is the Vforward of the rectifier
Using a center tapped secondary, with the center tap grounded:
Cap Input filter:
(1.414 x rms of the 1/2 secondary) - (minus) the tube rectifier drop, if you use a tube rectifier.
Choke Input filter:
(0.9 x rms of the 1/2 secondary) - (minus) the tube rectifier drop, if you use a tube rectifier.
To get the 0.9 factor, the choke has to have At Least the Critical Inductance value:
350/Load mA draw. Example, 100mA load, use At Least 3.5H choke (better yet use a 5H or 7H choke).
Solid state diode drops are much lower than tube rectifier drops.
The final B+ you get will be lower than the above 1.414x or 0.9x numbers.
That is because of the Primary DCR x Step Up Ratio; 1/2 Secondary DCR;
Choke DCR, and any resistors from cap to choke, choke to cap, cap to cap, etc.
Those are in series with the load, Volts drop = total resistance x load current.
The effects of the DCR caused voltage drops before a cap input;
And . . .
The effects of the Rectifier drop;
Are larger for Cap input filters,
Versus what they are for Choke input filters.
That is because the Cap input filters have large current spikes (and so large DCR and rectifier R drops);
And Choke input filters have Average current, so much lower DCR and rectifier R drops).
That means with Cap input filters you get less than 1.414x;
and with Choke input filters you may get close to up to 0.9x.
Some people design B+ using Software.
I use manual calculations to design to the desired B+ voltage, and the maximum ripple voltage that I will allow.
Cap Input filter:
(1.414 x rms of the 1/2 secondary) - (minus) the tube rectifier drop, if you use a tube rectifier.
Choke Input filter:
(0.9 x rms of the 1/2 secondary) - (minus) the tube rectifier drop, if you use a tube rectifier.
To get the 0.9 factor, the choke has to have At Least the Critical Inductance value:
350/Load mA draw. Example, 100mA load, use At Least 3.5H choke (better yet use a 5H or 7H choke).
Solid state diode drops are much lower than tube rectifier drops.
The final B+ you get will be lower than the above 1.414x or 0.9x numbers.
That is because of the Primary DCR x Step Up Ratio; 1/2 Secondary DCR;
Choke DCR, and any resistors from cap to choke, choke to cap, cap to cap, etc.
Those are in series with the load, Volts drop = total resistance x load current.
The effects of the DCR caused voltage drops before a cap input;
And . . .
The effects of the Rectifier drop;
Are larger for Cap input filters,
Versus what they are for Choke input filters.
That is because the Cap input filters have large current spikes (and so large DCR and rectifier R drops);
And Choke input filters have Average current, so much lower DCR and rectifier R drops).
That means with Cap input filters you get less than 1.414x;
and with Choke input filters you may get close to up to 0.9x.
Some people design B+ using Software.
I use manual calculations to design to the desired B+ voltage, and the maximum ripple voltage that I will allow.
Last edited:
How come I never see designs with a choke from the secondary center tap to the capacitor virtual ground while using the rectified full secondary as B+.
I personally really really dig that psu design the most.
I personally really really dig that psu design the most.
brightcity,
Good Old Terman . . .
Is the schematic in Post # 8 what you are talking about?
If so, that uses only 1/2 of the secondary per alternation.
(2 alternations per cycle).
Yes, it uses the whole secondary, but only 1/2 at a time.
When I talk about calculating B+ voltage on a center tapped secondary,
and I mention 1/2 of the secondary, that is what I am talking about . . .
Full Wave Rectification, one alternation at a time (1/2 secondary at a time).
If you want to use the Whole secondary voltage of a center tapped secondary all at the same time, then insulate the center tap, and then connect a bridge rectifier across the whole secondary. You will get twice the B+ voltage.
One reason to put the choke in the center tap lead, is to be able to use a lower voltage rated choke. In that circuit, it will have less voltage from the coils to the choke laminations (but it needs to be Cap input, if it is Choke input, it will have lots more voltage across the choke and one end of its coils).
Details, Details, Details.
Does that explain the differences?
Good Old Terman . . .
Is the schematic in Post # 8 what you are talking about?
If so, that uses only 1/2 of the secondary per alternation.
(2 alternations per cycle).
Yes, it uses the whole secondary, but only 1/2 at a time.
When I talk about calculating B+ voltage on a center tapped secondary,
and I mention 1/2 of the secondary, that is what I am talking about . . .
Full Wave Rectification, one alternation at a time (1/2 secondary at a time).
If you want to use the Whole secondary voltage of a center tapped secondary all at the same time, then insulate the center tap, and then connect a bridge rectifier across the whole secondary. You will get twice the B+ voltage.
One reason to put the choke in the center tap lead, is to be able to use a lower voltage rated choke. In that circuit, it will have less voltage from the coils to the choke laminations (but it needs to be Cap input, if it is Choke input, it will have lots more voltage across the choke and one end of its coils).
Details, Details, Details.
Does that explain the differences?
Last edited:
> I never see designs with a choke from the secondary center tap to the capacitor virtual ground while using the rectified full secondary as B+.
I'm gonna need a picture of that.
"Virtual ground" rarely appears in tube literature.
I'm gonna need a picture of that.
"Virtual ground" rarely appears in tube literature.
PRR,
I overlooked that part of brightcity's statements ('virtual ground').
One sign I made and put on my cubical at my work was:
"Grounds Are Commonly Misunderstood"
One example:
We had a customer using very large diameter ground conductors, 10 feet long, and with hundreds of amp pulses that had sub nano-second rise times.
They would ask, why is there 100V at 1 foot from the 'bottom' of the ground conductor?
I overlooked that part of brightcity's statements ('virtual ground').
One sign I made and put on my cubical at my work was:
"Grounds Are Commonly Misunderstood"
One example:
We had a customer using very large diameter ground conductors, 10 feet long, and with hundreds of amp pulses that had sub nano-second rise times.
They would ask, why is there 100V at 1 foot from the 'bottom' of the ground conductor?
Last edited:
Maybe something like this, but than with a choke between the centre-tap and "+1" (if that is possible). The voltage in this arrangement is about 150 V for "+1", and a little more than 300 V for "+4".
No "virtual ground" though.
No "virtual ground" though.
brightcity,
I guess you were wanting Details of my Post # 9.
All of what I described there are classical design power supplies.
The one with the choke in the center tap is less used anymore. But when someone is building a 1250V B+ for an 845 Triode,
and all he has is a 500V choke, he puts the choke in the center tap lead (caution, it needs to be with a cap input supply, and in that topology, the cap plus lead also has to go to the center tap.
A 500V choke can not swing the 1250V of the 1250V supply of a choke input supply, even though it is in the center tap circuit.
But all those are described in much literature; and they also appear in countless schematics in threads and posts in the Tubes / Valves part of this Forum.
Many in the Tubes / Valves have suggested many very good books to give understanding of the basic electronic fundamentals, and also more details.
I personally like the Radio Amateurs Handbook published every year by the ARRL.
Look it up on the web, and order one (or check with your local library, they may have a copy). I started with the 1956 version. Lots of tube stuff there.
Even the most recent version will be very helpful.
Learn resistors, capacitors, inductors, transformers, tubes, solid state, impedance, resonance, power supplies, oscillators, amplifiers, and much more.
Excellent theory, excellent implementation, and easy to understand.
I guess you were wanting Details of my Post # 9.
All of what I described there are classical design power supplies.
The one with the choke in the center tap is less used anymore. But when someone is building a 1250V B+ for an 845 Triode,
and all he has is a 500V choke, he puts the choke in the center tap lead (caution, it needs to be with a cap input supply, and in that topology, the cap plus lead also has to go to the center tap.
A 500V choke can not swing the 1250V of the 1250V supply of a choke input supply, even though it is in the center tap circuit.
But all those are described in much literature; and they also appear in countless schematics in threads and posts in the Tubes / Valves part of this Forum.
Many in the Tubes / Valves have suggested many very good books to give understanding of the basic electronic fundamentals, and also more details.
I personally like the Radio Amateurs Handbook published every year by the ARRL.
Look it up on the web, and order one (or check with your local library, they may have a copy). I started with the 1956 version. Lots of tube stuff there.
Even the most recent version will be very helpful.
Learn resistors, capacitors, inductors, transformers, tubes, solid state, impedance, resonance, power supplies, oscillators, amplifiers, and much more.
Excellent theory, excellent implementation, and easy to understand.
Last edited:
Thank you for your opinion! As this is one of my favorite psu designs, I will continue to discuss.
My issue is with the assumption that the center tap choke can me a small fractional voltage rating of a B+ choke.
This is simply quite the opposite.
There are two factors to take into account, adjacent asynchronous pairs and low current draw.
As the Center tap is between both ends of the secondary, the voltage incident at that point is a cross product of the delayed integrals at both end points. As the delay exists and the endpoints are not cojacent and with modern rectifiers, the factor can safely be considered to be 2 times the B+ voltage and rarely more.
Also the low current draw at the center tap will yield oscilitory forces, especially in the case of a virtual ground. This will be determined by the mains transformer with low correlation to the choke and could be damaging.
My issue is with the assumption that the center tap choke can me a small fractional voltage rating of a B+ choke.
This is simply quite the opposite.
There are two factors to take into account, adjacent asynchronous pairs and low current draw.
As the Center tap is between both ends of the secondary, the voltage incident at that point is a cross product of the delayed integrals at both end points. As the delay exists and the endpoints are not cojacent and with modern rectifiers, the factor can safely be considered to be 2 times the B+ voltage and rarely more.
Also the low current draw at the center tap will yield oscilitory forces, especially in the case of a virtual ground. This will be determined by the mains transformer with low correlation to the choke and could be damaging.
brightcity,
I apologize. I made a mistake.
In Post # 8, the first example schematic that Terman gives is actually a Choke input B+ filter.
That is because C1 is the stray capacitance, and not a large enough capacitance.
It does not have low capacitive reactance at 50Hz or 60Hz, nor at the 100Hz or 120Hz full wave rectified frequency.
That makes that circuit effectively a Choke input B+ filter.
In that case, the first choke has to swing a very large voltage. You can not use an inexpensive low voltage choke there, even though one end goes through the other filter elements to ground.
I am very curious about your "Virtual Ground".
And I am very curious about "using the whole secondary" (do you mean using the whole secondary during each alternation?
Would you please draw up a schematic of what you are talking about, and then please post it.
Thanks!
I apologize. I made a mistake.
In Post # 8, the first example schematic that Terman gives is actually a Choke input B+ filter.
That is because C1 is the stray capacitance, and not a large enough capacitance.
It does not have low capacitive reactance at 50Hz or 60Hz, nor at the 100Hz or 120Hz full wave rectified frequency.
That makes that circuit effectively a Choke input B+ filter.
In that case, the first choke has to swing a very large voltage. You can not use an inexpensive low voltage choke there, even though one end goes through the other filter elements to ground.
I am very curious about your "Virtual Ground".
And I am very curious about "using the whole secondary" (do you mean using the whole secondary during each alternation?
Would you please draw up a schematic of what you are talking about, and then please post it.
Thanks!
Last edited: