Long time reader, first time poster, need help interpreting power supply

[El Arte]

Hello!

I am looking at tube amplifiers of different kinds as a hobby.

I have much to learn.

I just document the attached power supply section of an amp I am dissecting.

It looks fairly familiar, but a few resistors are in unusual (to me) locations:

R1, I assume, is to bleed C4 on power down.

R3 must execute a necessary voltage drop, but I don't understand why it's not after this section entirely.

Similarly, I am not sure why R4 is where it is.

Can someone educate me or send me to some material that could educate me?

[kagliostro]

Give a look to all but answers to your questions are on The Power Supply section


http://www.valvewizard.co.uk


Franco

[El Arte]

Give a look to all but answers to your questions are on The Power Supply section


http://www.valvewizard.co.uk


Franco

Great, thank you!

[El Arte]

Darn, I see a 150uF reservoir capacitor after a Russian 5Ц4С rectifier. According to Valve Wizard, this could cause arcing, or strain the PT.
And yet, it appears to work fine in practice.
Although he also says "Single-ended amps may benefit from more capacitance because they don't reject hum like push-pull amps do."

More to read...

[El Arte]

According to ValveWizard, R2 should be 330K, not 220K (according to the 50/C rule he suggests).

[sluckey]

Your capacitor values are unrealistic for a tube rectifier. R1 and R2 bleeder resistor values are fine just as they are. This slightly redrawn schematic may be easier to understand.

[El Arte]

Your capacitor values are unrealistic for a tube rectifier. R1 and R2 bleeder resistor values are fine just as they are. This slightly redrawn schematic may be easier to understand.

Thank you! I like your picture better, as it resembles most filtering networks I have seen after a tube rectifier.

By unrealistic, I assume you mean there is no way the tube rectifier will ever be able to charge and keep those capacitors fully charged? Is it a waste of capacitor values or could it cause dire problems, such as those ValveWizard talks about in terms of causing the rectifier to arc?

Furthermore, in the schematic I posted above, R1 was a stand-in resistor (I added myself) designed to close the circuit for simulation in LTSpice. The actual amp build I am looking at doesn't have a 200K resistor in parallel with C6. The net marked +B2 in the schematic goes straight to the plate of some tube. Maybe the plate resistance of that tube plays the same role?

If these questions are silly, please forgive me, I am still new at this.

[PRR]

Please take a realistic example from any well-designed tube amp.

Fender AA-Champ, caps are in dozens-uFd, drop resistors are thousands Ohm.

470uFd and 10r are more like transistor circuits.

[El Arte]

Please take a realistic example from any well-designed tube amp.

Fender AA-Champ, caps are in dozens-uFd, drop resistors are thousands Ohm.

470uFd and 10r are more like transistor circuits.

Thank you for sharing!

It's certainly over-the-top (maybe following the silly "more is better" mantra), but will it hurt anything?
I am giving a listen to said amp and it's quite remarkable...there is no noise floor to speak of and one would swear it's performing as intended.
I am just trying to figure out the rules of these filter networks.
In theory, such oversized capacitors simply drive the cut-off frequency towards zero.
I also learned they force the PT and rectifier tube to bear a high current load.

So, the way I see it right now, it's either wasteful or it is not. It's either bad for the longevity of the PT and/or rectifier tube, or it is not.

What's not in doubt is that it performs admirably.

[sluckey]

It's certainly over-the-top (maybe following the silly "more is better" mantra), but will it hurt anything?

I don't understand why the rectifier has not sparked a small fire. Must not be pushing it hard?

[HotBluePlates]

Fender AA-Champ, caps are in dozens-uFd, drop resistors are thousands Ohm.

470uFd and 10r are more like transistor circuits.

It's certainly over-the-top (maybe following the silly "more is better" mantra), but will it hurt anything? ...

I don't understand why the rectifier has not sparked a small fire. Must not be pushing it hard?

High capacitance results in large current pulses to charge the too-large 470µF filter caps.

If there is not a bunch of series resistance between the PT and rectifier tube, those current pulses will cause the rectifier to arc plate-to-cathode.

When the rectifier arcs, it will pass negative volts as well as the normal (rectified) positive volts.

When the filter caps have negative volts applied, they will (literally) explode.  BANG!!  My first electronics teacher connected small electrolytic caps across outlet terminals to demonstrate this fact (which left a strong impression).

__________________________

Echoing Sluckey, it is a miracle your rectifier tube hasn't popped yet. 

[El Arte]

Fender AA-Champ, caps are in dozens-uFd, drop resistors are thousands Ohm.

470uFd and 10r are more like transistor circuits.

It's certainly over-the-top (maybe following the silly "more is better" mantra), but will it hurt anything? ...

I don't understand why the rectifier has not sparked a small fire. Must not be pushing it hard?

High capacitance results in large current pulses to charge the too-large 470µF filter caps.

If there is not a bunch of series resistance between the PT and rectifier tube, those current pulses will cause the rectifier to arc plate-to-cathode.

When the rectifier arcs, it will pass negative volts as well as the normal (rectified) positive volts.

When the filter caps have negative volts applied, they will (literally) explode.  BANG!!  My first electronics teach connected small electrolytic caps across outlet terminals to demonstrate this fact (which left a strong impression).

__________________________

Echoing Sluckey, it is a miracle your rectifier tube hasn't popped yet.

There are no resistors before the rectifier.
So, it must be something to do with the specs of the PT+rectifier. The rectifier is a Russian 5C4S, which can pass 122 mA according to the data sheet. Maybe that is beyond what the PT secondary can deliver to begin with, in which case it would work?

[pdf64]

With rectifiers as with any component, you’re free to breach their ratings.
They aren’t somehow going to step in and save themselves.

The 5C4S isn’t going limit current to 122mA.
It’s the designer’s responsibility to ensure that the circuit limits current and voltage to levels within the ratings of all the components involved.

[acheld]

Since you look like you're a Spice guy (from your posted schematics only), why not run a sim to see what your peak current is likely to be?   Does that agree with what your rectifier tube data sheet says?   

While it is common for (guitar) tube amps to push the boundaries of various tubes, such as exceeding voltage limits, it is NOT common to see experienced builders push the boundaries of tube rectifiers -- as all too often the consequences of the rectifier failure is literally a disaster.   

IF you need more reservoir and filtering capacitance, then it's time to move on to solid state rectifiers.

[El Arte]

Since you look like you're a Spice guy (from your posted schematics only), why not run a sim to see what your peak current is likely to be?   Does that agree with what your rectifier tube data sheet says?   

While it is common for (guitar) tube amps to push the boundaries of various tubes, such as exceeding voltage limits, it is NOT common to see experienced builders push the boundaries of tube rectifiers -- as all too often the consequences of the rectifier failure is literally a disaster.   

IF you need more reservoir and filtering capacitance, then it's time to move on to solid state rectifiers.

You’re right!

I put the supply side and the amp side in 2 different schematics, but I’ll try that as operating circumstances must explain why it works.

[HotBluePlates]

... The rectifier is a Russian 5C4S, which can pass 122 mA according to the data sheet. ...

The Russian data sheet says "140mA" for rectified current, but this is steady-state DC.  A little below that, it says "Peak Plate Current   1.4A".

5Ц4M data sheet


What does that mean?

During the brief window when the incoming rectified voltage is higher than the cap's voltage, there is a brief, large current pulse, as shown below:

... If there is not a bunch of series resistance between the PT and rectifier tube, those current pulses will cause the rectifier to arc plate-to-cathode. ...

There are no resistors before the rectifier.
So, it must be something to do with the specs of the PT+rectifier.

The winding resistance of the PT is also "series resistance" because it is in-series with the loop from

Cap Negative --> Ground --> PT Center-tap --> PT winding-end --> Rectifier Plate --> Rectifier Cathode --> Cap Positive

______________________________________

What governs the current?

Step back to Ohm's Law first:  Volts / Resistance = Current

Now consider that the size of the Peak Current Pulse is proportional to the difference between "discharged volts" and "rectifier output volts."  Or "how low does the Red Line dip before the Yellow Charging Window?" 

Say it drops 50v, and the PT's winding resistance is 35Ω:  50v / 35Ω = 1.43A


Now let's say you haven't attached this power supply to any load, you have simply assembled it on a breadboard.

With no load, the cap volts are not pulled down, and there is no "Peak Current Pulse" because the cap does not need recharging.  0v / Small Resistance = Zero Current ----> 0v / 35Ω = 0A


But later when you use it to power a real amp, the too-big caps will pull too high a peak charging current (unless PT winding resistance limits the current).

Small caps (small "C") require a smaller-charge (small "Q" and smaller-current (small "I") to reach a given voltage.

[El Arte]

... The rectifier is a Russian 5C4S, which can pass 122 mA according to the data sheet. ...

The Russian data sheet says "140mA" for rectified current, but this is steady-state DC.  A little below that, it says "Peak Plate Current   1.4A".

5Ц4M data sheet


What does that mean?

During the brief window when the incoming rectified voltage is higher than the cap's voltage, there is a brief, large current pulse, as shown below:

... If there is not a bunch of series resistance between the PT and rectifier tube, those current pulses will cause the rectifier to arc plate-to-cathode. ...

There are no resistors before the rectifier.
So, it must be something to do with the specs of the PT+rectifier.

The winding resistance of the PT is also "series resistance" because it is in-series with the loop from

Cap Negative --> Ground --> PT Center-tap --> PT winding-end --> Rectifier Plate --> Rectifier Cathode --> Cap Positive

______________________________________

What governs the current?

Step back to Ohm's Law first:  Volts / Resistance = Current

Now consider that the size of the Peak Current Pulse is proportional to the difference between "discharged volts" and "rectifier output volts."  Or "how low does the Red Line dip before the Yellow Charging Window?" 

Say it drops 50v, and the PT's winding resistance is 35Ω:  50v / 35Ω = 1.43A


Now let's say you haven't attached this power supply to any load, you have simply assembled it on a breadboard.

With no load, the cap volts are not pulled down, and there is no "Peak Current Pulse" because the cap does not need recharging.  0v / Small Resistance = Zero Current ----> 0v / 35Ω = 0A


But later when you use it to power a real amp, the too-big caps will pull too high a peak charging current (unless PT winding resistance limits the current).

Small caps (small "C") require a smaller-charge (small "Q" and smaller-current (small "I") to reach a given voltage.

So, the LTSpice simulation says peak inrush current happens at 0.1 seconds and is below 1A.

At 5 seconds, current pulses peak at ~0.2A per cycle.

Now, I trust LTSpice to be optimistic. I wish there were 1-ohm resistors before the rectifier plates, so I could verify what happens in the real world.

[El Arte]

The winding resistance of the PT is also "series resistance" because it is in-series with the loop from

Cap Negative --> Ground --> PT Center-tap --> PT winding-end --> Rectifier Plate --> Rectifier Cathode --> Cap Positive

I also realized that due to the way I simulate the PT in LTSpice, the winding resistance is not accounted for.

[HotBluePlates]

So, the LTSpice simulation says peak inrush current happens at 0.1 seconds and is below 1A.

At 5 seconds, current pulses peak at ~0.2A per cycle.

Did you have a "load" pulling current in-parallel to C4 and/or C5?

     An actual Fender Champ has a supply voltage of 350vdc, and an idle plate current of about 37mA.

     That means it looks like a resistance of  350/0.037A = ~9.5kΩ in parallel with C4 and R2.

     There is an additional ~3mA drawn by the screen, whose voltage would still be 349vdc at C5.  That load looks like 349v/0.003A = 117kΩ in parallel with C5.


If you build a single-ended amp with a tube bigger than a 6V6, the value of the effective "load resistor" will go down.
If you build a push-pull amp, the value of the effective load resistor will be halved.
If you build a Class AB amp, the idle current is smaller than the full-power current, so the latter must be the basis for estimating the effect on the power supply.

[El Arte]

So, the LTSpice simulation says peak inrush current happens at 0.1 seconds and is below 1A.

At 5 seconds, current pulses peak at ~0.2A per cycle.

Did you have a "load" pulling current in-parallel to C4 and/or C5?

     An actual Fender Champ has a supply voltage of 350vdc, and an idle plate current of about 37mA.

     That means it looks like a resistance of  350/0.037A = ~9.5kΩ in parallel with C4 and R2.

     There is an additional ~3mA drawn by the screen, whose voltage would still be 349vdc at C5.  That load looks like 349v/0.003A = 117kΩ in parallel with C5.


If you build a single-ended amp with a tube bigger than a 6V6, the value of the effective "load resistor" will go down.
If you build a push-pull amp, the value of the effective load resistor will be halved.
If you build a Class AB amp, the idle current is smaller than the full-power current, so the latter must be the basis for estimating the effect on the power supply.

Please don't hate me, but I think I totally misled you by omission. I was so keen on understanding the supply side, I neglected to tell you what the supply feeds.

This is not a guitar amp. This is a preamp that doubles as a headphones amp.

There is a single 12AX7 in the input stage, with each triode handling a channel.
The second stage is a 6C19P cathode follower stage (2 X 6C19P), which is probably there to reduce output impedance for headphones.

Regardless of what LTSpice says, I think I should actually take live measurements while a 1 kHz tone is playing at 1V, so I can measure what should be my level of worry going forward. From what LTSpice says, the biggest worry is from cold start, but if that peak current handling data is correct, it's quite possible the scheme is safe-ish.

Also, clarification: This is not my design, of course, and I am not defending this design. I just drew the schematic by reverse engineering the actual hardware, and I was mostly curious about the placement of those resistors.

[HotBluePlates]

... This is a preamp that doubles as a headphones amp.

There is a single 12AX7 in the input stage, with each triode handling a channel.
The second stage is a 6C19P cathode follower stage (2 X 6C19P), which is probably there to reduce output impedance for headphones. ...

That loading is so light that it is the same as "no load" near-enough.

With that in mind, I would ditch the rectifier tube because it's not doing anything but helping to warm the room in winter.  With some solid-state diodes, the whole issue of filter capacitance is mostly moot.

[El Arte]

... This is a preamp that doubles as a headphones amp.

There is a single 12AX7 in the input stage, with each triode handling a channel.
The second stage is a 6C19P cathode follower stage (2 X 6C19P), which is probably there to reduce output impedance for headphones. ...

That loading is so light that it is the same as "no load" near-enough.

With that in mind, I would ditch the rectifier tube because it's not doing anything but helping to warm the room in winter.  With some solid-state diodes, the whole issue of filter capacitance is mostly moot.

Thanks!

Any experience with these?

https://www.tubedepot.com/products/solid-state-rectifier

[sluckey]

I put one in the last amp I built.

[El Arte]

I put one in the last amp I built.

Thank you!

Is that a Dynaco kind of machine?

And where did you get that nice bench contraption that holds the chassis while you work on it?

[sluckey]

http://sluckeyamps.com/phoenix/phoenix.htm

http://sluckeyamps.com/cradle/cradle.htm