• WARNING: Tube/Valve amplifiers use potentially LETHAL HIGH VOLTAGES.
    Building, troubleshooting and testing of these amplifiers should only be
    performed by someone who is thoroughly familiar with
    the safety precautions around high voltages.

Help to understand and adapt Baby Ongaku PS to suit my power transformer...

I'm new to reading schematics (and designing power supplies) and curious to know what is happening in the Baby Ongaku PS. I should note that my ultimate goal is to use a transformer I have on hand that doesn't have a 6.3V CT.

So, from what I can tell, the amp has what is more or less a choke input featuring 22uF and 100uF filter caps followed by a 100k bleeder resistor. Does that sound about right?

When it comes to adapting the schematic to suit my PT, would I just run the 0V and the 6.3V from the PT to the 12AT7 and put the 22uF cap and the 100k bleeder resistor directly to ground?

Many thanks!!

Screen Shot 2023-07-25 at 11.25.19 PM.png
 
Search "heater elevation". Lack of a CT isn't a problem - see the below very good website.
Thank you for your insight! I was reading that WAY wrong... as to be expected at this point. I took a quick look at the site you sent before work and it's very helpful in identifying the elevation resistor and cap and learning about virtual center taps.

It seems with my transformer, I would use a virtual CT on the 6.3V secondary. That said in a stereo configuration - with two 12AT7s - would this be repeated on both 6.3V circuits?

I'm still a little confused as to how the 390k resistor works in relation to the heater elevation portion of the circuit. Is it a dropping resistor? Wouldn't this be where a bleeder resistor is inserted?

I would like to implement a choke input in the circuit since my transformer's high voltage secondary is a little high. In this case, I think I could eliminate the .68uF cap and play with some RC filtering after the choke. Is it correct that the .68uF cap is there to adjust the B+ voltage?

To practice (and learn) I've been trying to model this in PSUD; however, this one is a little tricky, and while a great piece of software I can't quite get this PS to be fully replicated, especially considering the elevated heater.
 
If you are just using one transformer, there is no need to duplicate the virtual point for the heater supply.

The 390k in conjunction with the 100k form a potential divider from b+ elevating the heater supply virtual centre tap by about 77v.
There is already a choke on the b+ supply and the .68uf is just a smoothing cap on the ht rectifier output, then comes the choke and then the 100uf.

If you want to drop the b+ a resistor (+ 100uf cap) after the choke would suffice. Just check the wattage requirement. The b+ voltage is not usually too critical within reason but the 7.5k resistor in the anode of the upper triode may need changing.
 
Thank you for your insight! I was reading that WAY wrong... as to be expected at this point. I took a quick look at the site you sent before work and it's very helpful in identifying the elevation resistor and cap and learning about virtual center taps.

It seems with my transformer, I would use a virtual CT on the 6.3V secondary. That said in a stereo configuration - with two 12AT7s - would this be repeated on both 6.3V circuits?

I'm still a little confused as to how the 390k resistor works in relation to the heater elevation portion of the circuit. Is it a dropping resistor? Wouldn't this be where a bleeder resistor is inserted?

I would like to implement a choke input in the circuit since my transformer's high voltage secondary is a little high. In this case, I think I could eliminate the .68uF cap and play with some RC filtering after the choke. Is it correct that the .68uF cap is there to adjust the B+ voltage?

To practice (and learn) I've been trying to model this in PSUD; however, this one is a little tricky, and while a great piece of software I can't quite get this PS to be fully replicated, especially considering the elevated heater.

The .68uF cap is there to limit the B+. Remember that this amp was designed to use available Magnequest transformers, so that cap is intentionally small to keep the B+ lower.
What are the specs of your power transformer? It's possible that a change of rectifier tube could lower your B+. Using the 5V4 is not a requirement here.
 
If you are just using one transformer, there is no need to duplicate the virtual point for the heater supply.
I spaced and was thinking my PT had two 6.3V windings. It only has one so will implement a virtual center tap on the one 6.3V secondary.

The 390k in conjunction with the 100k form a potential divider from b+ elevating the heater supply virtual centre tap by about 77v.
I'm still not clear on why this is done - I'm sure there is a great reason, just new to me. Why would you need to increase the heater by 77V? I'm probably (definitely) missing something but don't the filaments/heaters need just 6.3V AC?

What are the specs of your power transformer? It's possible that a change of rectifier tube could lower your B+. Using the 5V4 is not a requirement here.
It's a Hashimoto transformer and the specs are below and was thinking about using the 420V to feed a 5U4GB for more voltage drop across the rectifier. I'm not sure I would need the .68uF in this case, correct? It should drop further with the addition of an RC stage or two after the choke.

Primary 120V
Secondaries
480V-420V-240V-0-240V-420V-480V AC 170mA
0-2.5V-5V 5.5A 2 circuits
6.3V 3A
5V 3A
 
Every indirectly heated tube has what is called a heater-to-cathode limit. This is the limit on the voltage difference between the heater, which usually rests at ground potential, and the cathode, which is often biased many volts above ground. In this amp, half of the 12AT7 sits "on top" of the other half, an SRPP arrangement. The "top" half of the 12AT7 has it's cathode sitting at about 100 volts. This creates a very high voltage difference between the heater at ground and the cathode at 100 volts. This can lead to breakdown of the heater-cathode insulation and cause hum and eventually failure. The heater-cathode voltage limit of the 12AT7 is 90 volts. To reduce this voltage potential, you lift the filament supply from ground and connect it to a positive voltage source. The voltage divider off the B+ provides a positive voltage source to reference the filament supply to. In this case it looks like about 40-50 volts.

The 480-0-480 taps with a choke-input filter with a 5U4 rectifier would get you close to 350vdc. You still need a very small cap, like .47uF, after the rectifier, since many modern ungapped chokes cannot tolerate the initial ripple directly from the rectifier. They'll buzz. A hefty bleeder resistor, like 50K/10 watts, would also be adviseable after the second cap to ensure there is always some current draw on the supply, to prevent the voltage from shooting up in case there's a tube failure. The small first cap should limit the voltage, but I would still add a bleeder.
 
A pure choke input filter reduces the voltage, about 0.9 x Vrms. Other voltage drops include primary DCR; secondary DCR; tube rectifier, choke DCR, and the resistor that comes from the choke output and first capacitor, to the next capacitor.
All of the above depends on the Load current.

The 0.9 voltage reduction is based on the choke having at least the required critical inductance.
Critical inductance for 60Hz power mains and full wave rectification (120Hz) is 350/DC Load in mA
Critical inductance for 50Hz power mains and full wave rectification (100Hz)is 420/DC Load in mA

A monoblock with 50mA output stage, and perhaps input stage 5mA, and a bleeder/heater voltage offset circuit current 3mA, gives about 58mA.
In the US (60Hz / 120Hz full wave) . . . 350 / 58 = 6 Henry critical inductance.
A quality 10Henry 100mA choke should work well.
If you build a stereo amplifier, then the higher load current lets you use my favorite choke, the Hammond 193H, 5Henry 200mA.
I no longer build stereo amplifiers, I do not want a 4th Hernia, I build Mono-Block amplifiers (and I can use the 193H in all of them, I use more current).

The most often used 2A3 classical setup is: 250V plate to filament, 60mA, 2.5k OPT primary, and 45V Self Bias (295V not including the voltage drop in the output transformer primary. That works good. But then . . .

I prefer Gordon Rankin's design with the 2A3 at: 300V plate to filament, 50mA, 2.5k OPT primary, and 50V Self Bias (350V, not including the DCR voltage drop in the output transformer's primary. That works really good.

I actually prefer using a 3k, 3.5k, or 5k output transformer primary versus Gordon's 2.5k selection.
(A little less power, you will not notice; But with the advantages of lower distortion and higher damping factor; those you will notice).

A cap before the choke Increses the DCV at the choke output.

Not including the other voltage drops, 480Vrms x 0.9 = 432VDC
Not including the other voltage drops, 420Vrms x 0.9 = 378VDC
If the choke has critical inductance, then with a small capacitor before the choke, you get between 432VDC and 378VDC (Less than that when you include all the voltage drops that I mentioned above).
With some high voltage HEXFRED rectifiers, you can easily use the 420V secondary taps, and gain back the rectifier tap.
I need to get more HEXFREDs, I like the ones I have in one of my amplifiers.

Caution: If the B+ ever becomes unloaded, 480Vrms may rise to almost 670VDC.
That is a lot of voltage for the B+ filter caps, and is almost 1340 Peak Inverse Volts across the rectifier.
420Vrms is a little more friendly, 588VDC, and 1176 Peak Inverse Volts across the rectifier.
An additional 50K 25 Watt bleeder resistor can help to reduce the "unloaded" B+ voltage.
Put the bleeder After the choke.

Happy designing, building, and listening!
 
Last edited:
  • Like
Reactions: 1 user
Every indirectly heated tube has what is called a heater-to-cathode limit. This is the limit on the voltage difference between the heater, which usually rests at ground potential, and the cathode, which is often biased many volts above ground. In this amp, half of the 12AT7 sits "on top" of the other half, an SRPP arrangement. The "top" half of the 12AT7 has it's cathode sitting at about 100 volts. This creates a very high voltage difference between the heater at ground and the cathode at 100 volts. This can lead to breakdown of the heater-cathode insulation and cause hum and eventually failure. The heater-cathode voltage limit of the 12AT7 is 90 volts. To reduce this voltage potential, you lift the filament supply from ground and connect it to a positive voltage source. The voltage divider off the B+ provides a positive voltage source to reference the filament supply to. In this case it looks like about 40-50 volts.
Thank you for this very concise description. Makes total sense and is so much more clear than most of what I found online.

A monoblock with 50mA output stage, and perhaps input stage 5mA, and a bleeder/heater voltage offset circuit current 3mA, gives about 58mA.
In the US (60Hz / 120Hz full wave) . . . 350 / 58 = 6 Henry critical inductance.
A quality 10Henry 100mA choke should work well.
If you build a stereo amplifier, then the higher load current lets you use my favorite choke, the Hammond 193H, 5Henry 200mA.
Super helpful! I'll be in Japan in September and intend to buy a Hashimoto choke specified as 15H-200mA. Seems like it should work but will do some math!

I prefer Gordon Rankin's design with the 2A3 at: 300V plate to filament, 50mA, 2.5k OPT primary, and 50V Self Bias (350V, not including the DCR voltage drop in the output transformer's primary. That works really good.

I actually prefer using a 3k, 3.5k, or 5k output transformer primary versus Gordon's 2.5k selection.
(A little less power, you will not notice; But with the advantages of lower distortion and higher damping factor; those you will notice).
Curious to know how you calculate the plate-to-filament voltage. Is that the 350V B+ at the plate minus the 50V self-bias? Also, the output transformers I have will output at both 2.5k and 3.5k so I'm excited to try your suggestion.

With some high voltage HEXFRED rectifiers, you can easily use the 420V secondary taps, and gain back the rectifier tap.
I need to get more HEXFREDs, I like the ones I have in one of my amplifiers.
Have to look these up as I've never heard of them. I assume it's a solid-state rectifier. Very interesting!

420Vrms is a little more friendly, 588VDC, and 1176 Peak Inverse Volts across the rectifier.
An additional 50K 25 Watt bleeder resistor can help to reduce the "unloaded" B+ voltage.
Put the bleeder After the choke.
Will definitely go with the 420V and had planned to add a bleeder resistor for safety. Does the position of the bleeder resistor change at all with a choke input? Are you suggesting putting it to ground immediately after the choke? Or after any filtering? For example, after the 100uF capacitor in the original schematic.
 
For Choke input filter . . .
Put the bleeder right after the choke, the B+ is higher there, and you get maximum bleeder current through the choke, when there is no other load for the B+ (cold amplifier tubes).
Just be sure that All B+ series resistors in the filter that go from cap to cap to cap are high enough wattage that they will not burn out if any amplifier circuitry fails in a short to ground. Instead, make the load of a short circuit cause the fuses to blow.
Use 2 fuses to protect the power transformer and rectifiers.
A typical low power amplifier of mine has Both a 0.6A (600mA) Slow blow fuse in series with a 1.5A fast blow fuse.
Power mains Hot, fuse, fuse, power switch, power transformer primary; other lead of the primary to Neutral.
Fuse protection for power on inrush and power on hot start: and fuse protection for amplifier long term overload.

I used [solid state] HEXFREDs that were 1600V PIV, and 8 Amps. The 8 Amps was over the top overkill, but I wanted the 1600 PIV rating, and that was the part I found.

An example:
Calculating some voltage with 50mA of plate current
B+ Voltage 360V
Output transformer primary DCR 200 Ohms (drop 10V) 350V
2A3 Plate 350V
Self bias resistor 1k Ohms, 50V, 2A3 filament at + 50V
350V @ plate, and + 50V @ filament = + 300V plate to filament

Your 2.5k and 3.5k will have different DCRs to B+
 
  • Like
Reactions: 1 users
A typical low power amplifier of mine has Both a 0.6A (600mA) Slow blow fuse in series with a 1.5A fast blow fuse.
Power mains Hot, fuse, fuse, power switch, power transformer primary; other lead of the primary to Neutral.
Fuse protection for power on inrush and power on hot start: and fuse protection for amplifier long term overload.
Belt and suspenders! I'm into it and fuses are cheap.

An example:
Calculating some voltage with 50mA of plate current
B+ Voltage 360V
Output transformer primary DCR 200 Ohms (drop 10V) 350V
2A3 Plate 350V
Self bias resistor 1k Ohms, 50V, 2A3 filament at + 50V
350V @ plate, and + 50V @ filament = + 300V plate to filament
Just to be clear and make sure Im not missing something... This example calculation is based on knowing the current draw of the driver and output tube, the value of the high voltage AC secondary and the DCR of the output transformer primary. Correct?

Your 2.5k and 3.5k will have different DCRs to B+
Noted and will adjust according to the example you provided, adjusting the B+ before the output transformer to accommodate the DCR of the chosen primary.
 
joneci,

I have been making calibrated measurements since 1959 (though just a very simple measurement back then).
All my installation, maintenance, repair, overhaul, manufacturing test and cal, marketing, engineering, and technical support relied on careful and accurate measurements.
Most of my years, I worked for a Test and Measurement company.
But I was both out to sea and in port for 2 years on a US Navy destroyer, and that teaches you a lot (including how to troubleshoot, and then repair something without having the correct parts, there are none in the middle of the Pacific Ocean).

A nice oscilloscope differential measurement across a 0.1 Ohm resistor in the Neutral line, and many many power-up of a Cold amplifier, caught the maximum inrush current of my amplifiers. Correct trigger setup is paramount; and so is the patience to catch a transient when the physical on/off switch is closed at exactly the Crest of the power mains sine wave.
On one amplifier, I settled on a 1.25A Fast Blow fuse, and a 0.5A Slow blow fuse. After about 3 months, the 0.5A Slow Blow fuse fatigued, and opened.
It was replaced by a 0.6A (600mA) slow blow fuse, which has worked ever since.
So much for measurements . . . there are other factors, like fatigue.

Caution:
Never connect a scope probe's alligator clip to Neutral.
Neutral and Ground wires are different voltages (voltage from a very low impedance source, which has too much current for a scope ground lead).
The only place where Neutral and Ground are 0 Volts is all the way back to the Power Panel.
I once saw 8 Volts from ground to neutral in an industrial location, the scope ground smoked and burned the insulation of the probe ground wire!
(if Hot drops by 8 Volts in the long power mains wiring, then Neutral drops by 8 Volts in the long power mains wiring).
You do not want to make the mistake just because the voltage difference is only 0.5V between ground and neutral.

You probably have a resistor between a filter cap and the next filter cap that is the B+ filter cap that 'runs' your output transformer.
Just adjust that resistor, according to the different DCR of 2.5k and 3.5k primary. Done!
 
So it took a minute, but I have something I'm pretty happy with and would love feedback.

I've attached both the original PS schematic and my own adaptation. The tweaks are as follows:
  1. It will be a stereo amp so pulled a second leg of 350V B+.
  2. I am implementing a choke input instead of a capacitor input to smooth things and drop the voltage from my PT.
  3. I've added a virtual center tap to the 6.3V secondary to get it in line with the original circuit.
  4. There is now a bleeder resistor immediately after the choke.
That said, I have a couple of questions...
  1. I've modeled the PS in PSUD and it looks pretty good regarding ripple, but it won't model the PS with the bleeder resistor and potential divider. Is the best approach to keep these components as is and adjust the 100R resistor in the filter network to get my B+ to 350V?
  2. I was going to use electrolytic caps in the PS filter. Should I consider poly, etc. in this application?
  3. Would 3W resistors in the potential divider be sufficient?
Any feedback would be much appreciated. I hope to finalize the design this weekend prior to my trip to Japan - and component shopping. :)

Thanks!
 

Attachments

  • BabyOngaku_PS_5AR4-01.png
    BabyOngaku_PS_5AR4-01.png
    13.2 KB · Views: 94
  • Screen Shot 2023-08-25 at 10.52.05 AM.png
    Screen Shot 2023-08-25 at 10.52.05 AM.png
    132.8 KB · Views: 85
joneci,

Your 2A3 tubes each draw 50mA (50V bias/1000 Ohms).
That is 100mA.
You also have the input tubes plate current; and the 50k Bleeder resistor; and the 400k total resistance of the voltage divider for the filament voltage offset.

350V/50k = 7mA (use a 10 Watt resistor there to keep it from getting very hot)
350V/400k = ~1 mA 350Vx1mA = 0.35Watt total (2 Watt resistors or more).
1.5V/333 = 4.5mA input tube x 2 = 9mA


Your B+ current is at least:
100mA
7mA
1mA
9mA
. . . = 117mA
What is the current rating of your 15 Henry Choke?
I suggest a 200ma rating or more

I have never found it necessary or useful to use anything other than electrolytic capacitors for the B+ filtering.
Other Peoples Mileage May Vary (OPMMV).
 
Last edited:
400k total resistance of the voltage divider for the filament voltage offset.
Curious how you got to 400k. Just the 390k, no? Probably missing something.

What is the current rating of your 15 Henry Choke?
I suggest a 200ma rating or more
The choke specs are as follows:

15H (200mA)
40mA min. / 240mA max

I think it should be fine considering the current draw.

1.5V/333 = 4.5mA input tube x 2 = 9mA
Could you explain this calculation?

Thanks again for you help and consideration. So much to learn and appreciate your time.
 
joneci,

Your schematic is too small for me to read.
I thought the divider was 300k and 100k; that equals 400k
I guess your divider is 390k? and 100k; that equals 490k

The input/driver tube is serially connected.
The current is the same through the two [stacked] triodes, not 2X the current.
There are 1.5V across the 330 Ohm resistor.
1.5V/330 Ohms = 4.5mA; so 9mA for stereo.

I hope that explains it.
 
Your schematic is too small for me to read.
I thought the divider was 300k and 100k; that equals 400k
I guess your divider is 390k? and 100k; that equals 490k
Sorry about that... a little zoomed in below... but yes, 390K and 100K.

Screen Shot 2023-08-26 at 1.27.23 PM.png


The input/driver tube is serially connected.
The current is the same through the two [stacked] triodes, not 2X the current.
There are 1.5V across the 330 Ohm resistor.
1.5V/330 Ohms = 4.5mA; so 9mA for stereo.
That makes perfect sense as I was doubling the current. That said, have I over-complicated things by wiring the input tubes in series? If I went parallel, I could just run my wires from the single 6.3V AC secondary to pins 4/5 (bridged) and 9 and continue them to the same pins on the next tube.

Either way, seems as if 200mA in the 15H choke should be fine for this application.

When it comes to raising or dropping the B+ to hit the 350V, would it be the most economical to adjust the 100R resistor in the PS filter (between the caps)? Would be nice to not add an additional dropping resistor.

I've never implemented a virtual CT in a circuit but seems 220R for both legs is pretty standard. Is there any compelling reason to change this value?

All said, I feel pretty good about the power supply design and unless there are any glaring omissions or errors, I will proceed with this direction. Will post when I have the PS together later this fall.
 
The 4.5mA serial current I am talking about is:
Ground, 330 Ohms, cathode, plate, 330 Ohms, cathode, plate, 7.5k, 350V.
That is the SRPP input gain and driver stage.
Stereo is 9mA.

The 4.5mA and 9mA has nothing to do with the way you wire the filaments.
Filament current is 150mA at 12.6V (4 to 5), and 300mA at 6.3V (4 And 5; to 9).

You have 840V center tapped; that is 420VAC to each rectifier plate.
By the way, the maximum voltage to each plate for a 5AR4 is 450VAC; perhaps cutting it close depending on your power mains voltage; and the secondary DCR voltage drop with a choke input filter.
The 5AR4 rectifier drops perhaps 30V.
The choke input filter is 0.9 of the AC volts.
So, you have approximately 420 - 30 = 390; 390 x 0.9 = 351V (not exact, the rectifier is a DC drop, and 420VAC is 594V peak; so can not simply subtract 30V from 420V.
There is the voltage drop in the choke DCR.
You are close to your target voltage of 350VDC.
The power transformer secondary DCR voltage drop is lower for a choke input filter, than for a cap input filter.

You might have to use a rectifier that has a little more voltage drop, perhaps 50 or 60V, depending on all the factors above, that deliver X VAC to the 5AR4 plates.
You will not know for sure until you build it, too many variables above.

Normally, good tubes do not have large leakage current from the filaments to the cathodes.
So 220 Ohms is fine.
If there is large leakage from filament to cathode, the circuit might not work properly, in that case the tube needs to be replaced anyway.
 
The 4.5mA serial current I am talking about is:
Ground, 330 Ohms, cathode, plate, 330 Ohms, cathode, plate, 7.5k, 350V.
That is the SRPP input gain and driver stage.
Stereo is 9mA.

The 4.5mA and 9mA has nothing to do with the way you wire the filaments.
Filament current is 150mA at 12.6V (4 to 5), and 300mA at 6.3V (4 And 5; to 9).
I read all of it way too fast the first time and jumped to the filament current draw. Sorry for the confusion.

There is the voltage drop in the choke DCR.
You are close to your target voltage of 350VDC.
The power transformer secondary DCR voltage drop is lower for a choke input filter, than for a cap input filter.
Just found the choke specs at the DCR is 130 Ohm... so pretty low. Is there a good way to calculate secondary DCR voltage drop?

By the way, the maximum voltage to each plate for a 5AR4 is 450VAC; perhaps cutting it close depending on your power mains voltage; and the secondary DCR voltage drop with a choke input filter.
My mains are a bit low... Maybe my reading of the data sheet was off. It shows both 450V and 550V in the Characteristics and Typical Operation section and seemed like it would be okay with 420V off the PT. That said I've been modeling with a Max Off Load Voltage of 462V AC.

You might have to use a rectifier that has a little more voltage drop, perhaps 50 or 60V, depending on all the factors above, that deliver X VAC to the 5AR4 plates.
You will not know for sure until you build it, too many variables above.
I do have a 5V4G and a 5U4GB lying around that would give me more voltage drop and have a higher max plate voltage of 500V and 550V AC RMS respectively.
 

Attachments

  • 5AR4.pdf
    495.3 KB · Views: 42