I have built a clone of a popular speaker amp for headphone use, using headphone output transformers. It has worked fine for the last year. I am now trying to re-clone it to a higher model that uses tube regulation. The problem I'm having is trying to determine the value of a couple of resistors for the input tube PSU, using an OD3 regulator. I am going by pictures that the builder posted, but the pix showing them are not clear, and I can't quite make out the stripes. I've diagramed the circuit, with question marks for the resistors in question.
For the resistor at the top of the diagram, the stripes are dark-dark-yellow-gold. I've listed the possible values on the diagram.
For the resistor at the bottom, I can only see the first stripe which is yellow.
Any help will be greatly appreciated.
For the resistor at the top of the diagram, the stripes are dark-dark-yellow-gold. I've listed the possible values on the diagram.
For the resistor at the bottom, I can only see the first stripe which is yellow.
Any help will be greatly appreciated.
You need to give us more information about your complete circuit.
Your posted schematic is not a constant voltage regulator.
Instead, your circuit is using the OD3 as a constant voltage offset.
If B+ varies the voltage to the 7DJ8 plate load resistors is not constant.
If the 7DJ8 plate current varies, the voltage to the 7DJ8 plate loads from the OD3 is not constant.
Start with a description of what you need the regulator to do:
What constant Voltage?
What current load (minimum and maximum current with no signal and with maximum signal)?
Your posted schematic is not a constant voltage regulator.
Instead, your circuit is using the OD3 as a constant voltage offset.
If B+ varies the voltage to the 7DJ8 plate load resistors is not constant.
If the 7DJ8 plate current varies, the voltage to the 7DJ8 plate loads from the OD3 is not constant.
Start with a description of what you need the regulator to do:
What constant Voltage?
What current load (minimum and maximum current with no signal and with maximum signal)?
Is the +325V actually 0V, and 0V at the line side of the 2W resistor actually +325V and are the 0D3 elements drawn upside down?.
Otherwise, the whole thing is drawn upside down, and you have 175V drop across 16k5 for 10.5mA.. and min current thru the 0D3 is 5mA, which leaves about that againfor the signal tube.. and the plate then sits at 0.005x either 100k/600k/700k minus the >5mA drop over the 28k plate load.
Otherwise, the whole thing is drawn upside down, and you have 175V drop across 16k5 for 10.5mA.. and min current thru the 0D3 is 5mA, which leaves about that againfor the signal tube.. and the plate then sits at 0.005x either 100k/600k/700k minus the >5mA drop over the 28k plate load.
ISCMMS,
In the schematic of Post # 1,
The OD3 plate is closer to B+:
+325V through two parallel 33k to the OD3 Plate
And the OD3 cathode is closer to ground.
the OD3 cathode is connected through 100k/600k/700k to ground.
The OD3 is connected in the correct polarity.
But the rest of the correctness of that schematic is still in question.
As I said, the voltage to the 7DJ8 56k plate load resistors is not regulated.
In the schematic of Post # 1,
The OD3 plate is closer to B+:
+325V through two parallel 33k to the OD3 Plate
And the OD3 cathode is closer to ground.
the OD3 cathode is connected through 100k/600k/700k to ground.
The OD3 is connected in the correct polarity.
But the rest of the correctness of that schematic is still in question.
As I said, the voltage to the 7DJ8 56k plate load resistors is not regulated.
Last edited:
Thanks for the replies. TBH, I really can't tell where the line below the two film cap's terminates from the pix. It could go to ground instead of thru the .5W resistor to the B+. That might make more sense? I've read that OD3's can be noisy, thus the cap bypass. I'm not all that savy with circuitry, but am a great thief..... er, copy cat.
The 13th and 14th (last two) pix from the top show the completed circuit. The power tubes (Russian SV83, EL84 equivalents) each use an OA3. I have to agree that I don't think that the VR tubes are actually being used as reg's per se.
https://www.decware.com/cgi-bin/yabb22/YaBB.pl?num=1510887200/400
This is the circuit I am now using, with the addition of a rectifier tube instead of diodes. Power transformer is a Hammond 272FX.
https://www.decware.com/newsite/ZKIT1.pdf
https://www.decware.com/cgi-bin/yabb22/YaBB.pl?num=1510887200/400
This is the circuit I am now using, with the addition of a rectifier tube instead of diodes. Power transformer is a Hammond 272FX.
https://www.decware.com/newsite/ZKIT1.pdf
If the drawn schematic is correct, then the purpose of the resistor at the top is to allow a current path to run the OD3 at the minimum load, assuming the 7DJ8 draws no current (such as when the filament is cold).
So for top resistor:
Minimum OD3 current is 5mA. So
(325V - 153V) / (33K|33K + ?ohm) = 0.005A
(325V - 153V) / 0.005A - 16500ohm = ?ohm
?ohm = 17.9Kohm
The lower resistor would be to limit inrush current from the +325V rail into the 3.3uF cap. A few kohm is probably in the ballpark. 4.7K perhaps?
This leaves at most 90V on the plate of the 7DJ8 so I would question if the schematic is drawn correctly. Are you sure the 56K resistors are no connected to the anode of the OD3 instead of the cathode? Either that or the OD3 is being run below it's minimum current, or they disregard starting the OD3 before the 7DJ8 tube warms up, in which case I'm not sure why even have the top resistor.
Just find a reliable circuit and build it exactly as intended.
So for top resistor:
Minimum OD3 current is 5mA. So
(325V - 153V) / (33K|33K + ?ohm) = 0.005A
(325V - 153V) / 0.005A - 16500ohm = ?ohm
?ohm = 17.9Kohm
The lower resistor would be to limit inrush current from the +325V rail into the 3.3uF cap. A few kohm is probably in the ballpark. 4.7K perhaps?
This leaves at most 90V on the plate of the 7DJ8 so I would question if the schematic is drawn correctly. Are you sure the 56K resistors are no connected to the anode of the OD3 instead of the cathode? Either that or the OD3 is being run below it's minimum current, or they disregard starting the OD3 before the 7DJ8 tube warms up, in which case I'm not sure why even have the top resistor.
I don't mean any offense, but if you aren't able to calculate reasonable resistor values for the circuit in the OP, I would not recommend making arbitrary changes to a circuit such as changing silicon diodes for a tube rectifier. A rectifier tube may not be happy with the peak current that a silicon diode rectifier circuit experiences and the power supply voltage is going to be way lower than with silicon diodes. To convert that circuit to tube rectification you're probably going to need to make all the filtering caps smaller, resistances higher and pick and transformer with a higher voltage secondary. The power rating of the resistors and voltage rating of the capacitors may need to be uprated.
Just find a reliable circuit and build it exactly as intended.
Last edited:
I took the liberty of redrawing it ..
All that is certain is that there is 150V across the VR tube
To my mind, the shunt resistor 'x' could assist in ensuring the VR tube strikes.
The resistor 'x' in series with the 3.3uF.. well, clutching at straws, my guess is that its there to flatten the impedance of the slightly inductive VR tube, why?.. no idea.
You need to know the voltage at the -ve end of the VR tube and a lot more about the 7DJ8 stage/s.
Having seen some of this guys stuff before, that VR tube is there to drop voltage and 'provide a vacuum which the power supply hash cannot pass through' - or something fun like that.
Anyway...
All that is certain is that there is 150V across the VR tube
To my mind, the shunt resistor 'x' could assist in ensuring the VR tube strikes.
The resistor 'x' in series with the 3.3uF.. well, clutching at straws, my guess is that its there to flatten the impedance of the slightly inductive VR tube, why?.. no idea.
You need to know the voltage at the -ve end of the VR tube and a lot more about the 7DJ8 stage/s.
Having seen some of this guys stuff before, that VR tube is there to drop voltage and 'provide a vacuum which the power supply hash cannot pass through' - or something fun like that.
Anyway...
Valid point, but at one time I had a Decware SE84 amp, and sketched the circuit. It's identical to the kit schematic but uses the rectifier tube. I chose the 272FX tranny because it matched the spec's exactly. Later I learned that Decware uses that very tranny.
My diagram is correct: the 56K resistors come off of pin 2 of the OD3.
Thanks so much for all your help; things to ponder.
My diagram is correct: the 56K resistors come off of pin 2 of the OD3.
Thanks so much for all your help; things to ponder.
ISCMMS,
Parallel voltage regulator:
If an OD3 is not a low impedance to 20kHz, then it would not work as a constant voltage regulator for a Pentode Screen, or a Beam Power Screen. But it does get used that way, and very successfully.
Series voltage drop:
An OD3 passes DC current, And . . . it will also pass power supply hash current.
Want to use an OD3 to reject power supply hash . . . then use the OD3 as a voltage regulator;
Do not use it as a series voltage drop.
An OD3 is not a vacuum tube, it is a Gas Tube Regulator (even though the gas is at a very low pressure).
Just my opinions.
Parallel voltage regulator:
If an OD3 is not a low impedance to 20kHz, then it would not work as a constant voltage regulator for a Pentode Screen, or a Beam Power Screen. But it does get used that way, and very successfully.
Series voltage drop:
An OD3 passes DC current, And . . . it will also pass power supply hash current.
Want to use an OD3 to reject power supply hash . . . then use the OD3 as a voltage regulator;
Do not use it as a series voltage drop.
An OD3 is not a vacuum tube, it is a Gas Tube Regulator (even though the gas is at a very low pressure).
Just my opinions.
The impedance is definately non-linear, whether you'd call it a low impedance is another thing.
If the R + 3.3uF could be connected wrt 0V (as the OP states), perhaps it is final decoupling filter/s for the 7DJ8?.
IIRC he uses these small values (also 3.3uF) elsewhere in his designs.
Whether it works as described or not, he does use them in this way.
Funch, if you have built the basic version, what is the voltage at either side of the 56k plate load resistor, and if you havent built it, can you find the value of the cathode resistor 'x3'?. Also, above at #8 you mention 0A3 - is the tube 0D3 or 0A3?.
Last edited:
ISCMMS,
Nice graph!
Where did you get your OD3 impedance graph?
The nominal regulation voltage rating of the OD3 is 150V.
The OD3 current rating is from 5mA to 40mA (a range of 35mA).
Over the current range of 5mA to 40mA, the change in regulation voltage is 4V.
4V / 0.035A = 114.3 Ohms.
The mystery is . . . How is the OD3 "impedance" * at DC 114.3 Ohms,
and yet from your graph, the impedance shows 75 Ohms from 3 Hz to 40Hz.
* Yes, I realize that using the term "impedance" at DC is incorrect.
And, your OD3 graph shows un-bypassed impedance of 400 Ohms at 22kHz, the highest red-book CD frequency.
Two 56k Ohm loads in parallel is 28k.
28,000 / 400 = 70:1
So, at 22kHz, the most voltage variation of the OD3 is 1/70th of the signal voltage across the two 56k resistors; and that worst case is only when the signal across both 56k resistors is the same voltage, and is the same phase (Mono, not stereo). For anything other than Mono, the voltage variation across the OD3 is less than 1/70th of the signal voltage(s).
Look at it this way, at 20Hz the plate load is 56,075 Ohms, and at 22kHz, it is 56,400 Ohms.
And, if a power supply has hash at 22kHz, fix the power supply first.
Do not use an OD3's 400 Ohm series impedance at 22kHz to remove the hash.
Another concern of signal level versus frequency:
How much does the output transformer primary impedance vary from 20Hz to 22kHz, when the secondary is loaded with a non-inductive load resistor of the secondary rated impedance?
Now, How much does the output transformer primary impedance vary from 20Hz to 22kHz, when the secondary is loaded with your loudspeaker's varying impedance across the audio band?
To worry, or not to worry, that is the question.
Try taking the outer layers off of the Onion first, before attempting to remove the Onion core.
Nice graph!
Where did you get your OD3 impedance graph?
The nominal regulation voltage rating of the OD3 is 150V.
The OD3 current rating is from 5mA to 40mA (a range of 35mA).
Over the current range of 5mA to 40mA, the change in regulation voltage is 4V.
4V / 0.035A = 114.3 Ohms.
The mystery is . . . How is the OD3 "impedance" * at DC 114.3 Ohms,
and yet from your graph, the impedance shows 75 Ohms from 3 Hz to 40Hz.
* Yes, I realize that using the term "impedance" at DC is incorrect.
And, your OD3 graph shows un-bypassed impedance of 400 Ohms at 22kHz, the highest red-book CD frequency.
Two 56k Ohm loads in parallel is 28k.
28,000 / 400 = 70:1
So, at 22kHz, the most voltage variation of the OD3 is 1/70th of the signal voltage across the two 56k resistors; and that worst case is only when the signal across both 56k resistors is the same voltage, and is the same phase (Mono, not stereo). For anything other than Mono, the voltage variation across the OD3 is less than 1/70th of the signal voltage(s).
Look at it this way, at 20Hz the plate load is 56,075 Ohms, and at 22kHz, it is 56,400 Ohms.
And, if a power supply has hash at 22kHz, fix the power supply first.
Do not use an OD3's 400 Ohm series impedance at 22kHz to remove the hash.
Another concern of signal level versus frequency:
How much does the output transformer primary impedance vary from 20Hz to 22kHz, when the secondary is loaded with a non-inductive load resistor of the secondary rated impedance?
Now, How much does the output transformer primary impedance vary from 20Hz to 22kHz, when the secondary is loaded with your loudspeaker's varying impedance across the audio band?
To worry, or not to worry, that is the question.
Try taking the outer layers off of the Onion first, before attempting to remove the Onion core.
6A3sUMMER,
Graph was pulled from the web. I'd measured the impedance of 0A2 types many years ago when there was some talk about these things, and I found very consistently the impedance peak as per the graph, so I knew it was eaisier to find something online than to go digging for my notes from 20 years ago.
I'm not here to critique or redesign the OP's ambition, merely to try and work out what the schematic is - a puzzle and a bit of fun.
I'll be interested to hear back from the OP to confirm some voltages, or cathode resistor values. I think then, there will be enough info to close out the puzzle and also for him to then build it if he wishes to.
Cheers
Graph was pulled from the web. I'd measured the impedance of 0A2 types many years ago when there was some talk about these things, and I found very consistently the impedance peak as per the graph, so I knew it was eaisier to find something online than to go digging for my notes from 20 years ago.
I'm not here to critique or redesign the OP's ambition, merely to try and work out what the schematic is - a puzzle and a bit of fun.
I'll be interested to hear back from the OP to confirm some voltages, or cathode resistor values. I think then, there will be enough info to close out the puzzle and also for him to then build it if he wishes to.
Cheers
Theres not a lot of bias on 7DJ8 tube with Rk 159 ohm and 1.9mA of plate current, 1.59k ?
Anyway, take the B supply for the second channel at that 225V node and duplicate the RC decoupling 1k + 3.3uF.
Values in blue will enable the VR tube to strike without the plate current.. values in red are more in line with the values you've suggested.
That 33k//33k doesnt fit, nor does the 0D3.
NOTE: The 100k and 39k resistors below dissipate ~0.5W and ~1.5W respectively.
Edit: Could that be 3k3//3k3 .. ~15mA total and ~11mA through a ~20k shunt resistor?.
Anyway, have fun.
Cheers
Anyway, take the B supply for the second channel at that 225V node and duplicate the RC decoupling 1k + 3.3uF.
Values in blue will enable the VR tube to strike without the plate current.. values in red are more in line with the values you've suggested.
That 33k//33k doesnt fit, nor does the 0D3.
NOTE: The 100k and 39k resistors below dissipate ~0.5W and ~1.5W respectively.
Edit: Could that be 3k3//3k3 .. ~15mA total and ~11mA through a ~20k shunt resistor?.
Anyway, have fun.
Cheers
Last edited: