One way to get to a lower voltage would use a vacuum rectifier such as those in the 5R4G family with choke input.
Interpolating a 575-0-575 supply on the curves gets to less than 500VDC after IR drop in the choke under a typical load.
More filtering would be required for this SE circuit, There would also be more DC drop in the cathode bias resisters,🙂
The current in each tube will be limited by their dissipation spec, That means their rp will be high, not good for Damping Factor.
Typical SEUL have DF of approx One under normal conditions. This amp as is will most likely have a DF of less than One.
Something like 6 db of NFB from the OPT secondary to the driver cathode would help a lot. 👍
Interpolating a 575-0-575 supply on the curves gets to less than 500VDC after IR drop in the choke under a typical load.
More filtering would be required for this SE circuit, There would also be more DC drop in the cathode bias resisters,🙂
The current in each tube will be limited by their dissipation spec, That means their rp will be high, not good for Damping Factor.
Typical SEUL have DF of approx One under normal conditions. This amp as is will most likely have a DF of less than One.
Something like 6 db of NFB from the OPT secondary to the driver cathode would help a lot. 👍
OK so based on the comments here the first step in the modification process is to add a choke input.
What parts will this require and how do we make the modification?
(aiming to make these changes carefully!)
What parts will this require and how do we make the modification?
(aiming to make these changes carefully!)
In terms of adding a choke input filter, I've just checked my spare parts box and have a 8H 156 Ohm Choke made by Danbury Electronics. Would this work, and where should it be inserted in the circuit?
What is the Current rating (mA) of the choke?
It needs to be high enough so that the choke will not saturate under load.
Post # 1 schematic:
After the B+ rectifier diodes, there is a 10 Ohm resistor from the diodes to the two 150uF caps.
Connect the choke between the 10 Ohm resistor and the top of the top 150uF cap.
Now you have rectifiers, 10 ohms, choke, and two 150uF caps to ground.
Simple?
Check the new B+ voltage at the choke output to the two 150uF caps (the B+ needs to be loaded; two output tubes, and two input tubes.
Suppose: the output tubes are set for 40mA each (80mA), and the input tubes draw 2mA each (4mA), then the load is 84mA plus the bleeder resistors current.
Unfortunately, with a 575-0-575V primary, and solid state rectifiers, you are going to get almost 518V, still pretty high.
That makes the peak inverse volts across the two series diodes (and likewise the other 2 series diodes), to be about 1600V before the output tubes warm up.
With 575V x 1.4 = 805VDC, that is almost as much as the two 450V caps in series (900V).
And, before the output tubes warm up, the B+ is unloaded, so the 575V will be higher, perhaps 600V or more.
Even a tube rectifier will only drop about 40 to 60V, not enough drop from over 500V.
You really do need to purchase a different power transformer (in my opinion).
You can get rid of more B+ voltage by using self bias resistors and bypass caps for the EL34 tubes.
Your output transformers are 15k Ohms right?
Are the UL taps at 40%?
If they are, you can connect the plate to the plate tap, and connect the B+ to the UL tap.
That makes the portion of the primary that you are using be 5,400 Oms.
Then Triode wire the EL34 tubes (remove the RFC between the screen and the UL tap) and connect a 100Ohm resistor from the screen to the plate.
There is a lot to do in the above suggestions, but otherwise you need to purchase a new power transformer.
It needs to be high enough so that the choke will not saturate under load.
Post # 1 schematic:
After the B+ rectifier diodes, there is a 10 Ohm resistor from the diodes to the two 150uF caps.
Connect the choke between the 10 Ohm resistor and the top of the top 150uF cap.
Now you have rectifiers, 10 ohms, choke, and two 150uF caps to ground.
Simple?
Check the new B+ voltage at the choke output to the two 150uF caps (the B+ needs to be loaded; two output tubes, and two input tubes.
Suppose: the output tubes are set for 40mA each (80mA), and the input tubes draw 2mA each (4mA), then the load is 84mA plus the bleeder resistors current.
Unfortunately, with a 575-0-575V primary, and solid state rectifiers, you are going to get almost 518V, still pretty high.
That makes the peak inverse volts across the two series diodes (and likewise the other 2 series diodes), to be about 1600V before the output tubes warm up.
With 575V x 1.4 = 805VDC, that is almost as much as the two 450V caps in series (900V).
And, before the output tubes warm up, the B+ is unloaded, so the 575V will be higher, perhaps 600V or more.
Even a tube rectifier will only drop about 40 to 60V, not enough drop from over 500V.
You really do need to purchase a different power transformer (in my opinion).
You can get rid of more B+ voltage by using self bias resistors and bypass caps for the EL34 tubes.
Your output transformers are 15k Ohms right?
Are the UL taps at 40%?
If they are, you can connect the plate to the plate tap, and connect the B+ to the UL tap.
That makes the portion of the primary that you are using be 5,400 Oms.
Then Triode wire the EL34 tubes (remove the RFC between the screen and the UL tap) and connect a 100Ohm resistor from the screen to the plate.
There is a lot to do in the above suggestions, but otherwise you need to purchase a new power transformer.
Last edited:
This is clearly a concept amp and Jack Elliano was no fool. I too was at first utterly baffled by the high voltages until I read the description. If the amp has been working for you without blowing tubes then I would not change the design. It is what it is. However, at those extreme conditions it's hardly surprising that you have trouble matching the output tubes.
The first thing I would do is remove the output tubes, turn the amp on and measure the DC voltage from pin 5 of one of the EL34 sockets to ground. Don't turn the amp upside down, there will be about 700vdc floating around under there. ;-) Turn the bias pot to both extremes and check the range of voltage it supplies.
Then I would add a second 20K pot and 36K resistor to ground, and connect the center of the pot to the second EL34 grid. Add another 47K resistor across the existing 47K resistor to halve the value of that resistor. You should get about the same range of negative bias voltage. Adjust the 47K value until you get the same range as before. Does that make sense?
I would not increase the bias pot values. You don't want to change the grid resistance on the EL34s if you can avoid it.
The first thing I would do is remove the output tubes, turn the amp on and measure the DC voltage from pin 5 of one of the EL34 sockets to ground. Don't turn the amp upside down, there will be about 700vdc floating around under there. ;-) Turn the bias pot to both extremes and check the range of voltage it supplies.
Then I would add a second 20K pot and 36K resistor to ground, and connect the center of the pot to the second EL34 grid. Add another 47K resistor across the existing 47K resistor to halve the value of that resistor. You should get about the same range of negative bias voltage. Adjust the 47K value until you get the same range as before. Does that make sense?
I would not increase the bias pot values. You don't want to change the grid resistance on the EL34s if you can avoid it.
I am worn out with this thread.
Build it exactly as described in Post # 1 schematic; and Either fight with the changes in EL34 plate currents, and any other problems,
Or change the circuit.
I am not convinced Post # 1 schematic is an Exact copy of Jack Eliano's schematic.
Please prove me wrong . . . Get me a link to His (Jack's) schematic.
I went to Jack's home / business at least twice, and talked to him at multiple trade shows.
I really miss those old days when I did business with him. I think the world of him.
Thanks in advance for the link!
Pull out the EL34 tubes? . . . No load on the B+???
Well, 575VAC = 805VDC (oh, a little less than that, the solid state rectifiers drop about 1V each).
And a good 575VAC secondary is more than 575VAC when it is un-loaded. You get more than 805VDC un-loaded.
Un-loaded choke input power supplies DCV is also 1.4x VAC.
The 575VAC transformer is more suited to DC coupled 2 stage amplifiers, like a 300B modified 'version' of the loftin white 2A3 amplifier;
Or to Jack Eliano's 300B DRD amplifier.
Build it exactly as described in Post # 1 schematic; and Either fight with the changes in EL34 plate currents, and any other problems,
Or change the circuit.
I am not convinced Post # 1 schematic is an Exact copy of Jack Eliano's schematic.
Please prove me wrong . . . Get me a link to His (Jack's) schematic.
I went to Jack's home / business at least twice, and talked to him at multiple trade shows.
I really miss those old days when I did business with him. I think the world of him.
Thanks in advance for the link!
Pull out the EL34 tubes? . . . No load on the B+???
Well, 575VAC = 805VDC (oh, a little less than that, the solid state rectifiers drop about 1V each).
And a good 575VAC secondary is more than 575VAC when it is un-loaded. You get more than 805VDC un-loaded.
Un-loaded choke input power supplies DCV is also 1.4x VAC.
The 575VAC transformer is more suited to DC coupled 2 stage amplifiers, like a 300B modified 'version' of the loftin white 2A3 amplifier;
Or to Jack Eliano's 300B DRD amplifier.
Last edited:
You're indeed wearing yourself out. ;-) What makes you think that isn't Jack's own schematic? I'm sure the OP can verify where it came from. And why should he change the amp if it's been working, except to allow for individual tube bias? Unfortunately the old Electra-Print web site is gone so the original 300B concept article is no longer available.
Hi guys it's Jack's schematic - he sent it in his own handwriting with the amp. Also I'd qualify that he is no fool rather than was no fool, still very much alive and writes to me regularly!
As I said this thread was about the bias mismatch but along the way we have raised the issue of the alarmingly high voltages.
I have relayed to Jack and his comments are as follows.
re the bias:
"No real need to bias them perfectly the same, just within 35-40ma idle.
The current goes up with signal, they are not class A, they are A3!!
Class A (12AX7) stays constant current (.9ma) while doing its job, the 12AX7 is in class A.
We used LED’s for bias once, looked on the scope on its output and flipped the scan rate up and saw an impressed oscillation up in the megahertz!? LED’s oscillate when making the light, and it adds it to the signal. !! who knew!?
Well that was that. we did use just standard diode, 1N4002’ in series, to bias and it did work but no big deal. Use just a good resistor and go to bed happy!!"
re the high voltage:
"In a vacuum, the voltage between elements is in the 1000’s of volts.The curves they show in the common tube manuals are for many usual uses and in a common use type explanation to sell product.
The most I am questioned about is pushing the tube hard and lower life.
In fact, the lower the current in A3 operation increases the life span!!
Curves are made for many uses even for digital on/off use too!
Just use these books for pin out connections and common use info.
A3 basically is a particle accelerator, much like an old TV picture tube, the screen acting as an accelerator to speed up the electrons and hit the plate harder creating more power!!! Doing this, the tube has another set of rules not seen in any tube manual!!
You hear it, and those voltages are in there and they work, been doing this A3 for 12 years now."
As I said this thread was about the bias mismatch but along the way we have raised the issue of the alarmingly high voltages.
I have relayed to Jack and his comments are as follows.
re the bias:
"No real need to bias them perfectly the same, just within 35-40ma idle.
The current goes up with signal, they are not class A, they are A3!!
Class A (12AX7) stays constant current (.9ma) while doing its job, the 12AX7 is in class A.
We used LED’s for bias once, looked on the scope on its output and flipped the scan rate up and saw an impressed oscillation up in the megahertz!? LED’s oscillate when making the light, and it adds it to the signal. !! who knew!?
Well that was that. we did use just standard diode, 1N4002’ in series, to bias and it did work but no big deal. Use just a good resistor and go to bed happy!!"
re the high voltage:
"In a vacuum, the voltage between elements is in the 1000’s of volts.The curves they show in the common tube manuals are for many usual uses and in a common use type explanation to sell product.
The most I am questioned about is pushing the tube hard and lower life.
In fact, the lower the current in A3 operation increases the life span!!
Curves are made for many uses even for digital on/off use too!
Just use these books for pin out connections and common use info.
A3 basically is a particle accelerator, much like an old TV picture tube, the screen acting as an accelerator to speed up the electrons and hit the plate harder creating more power!!! Doing this, the tube has another set of rules not seen in any tube manual!!
You hear it, and those voltages are in there and they work, been doing this A3 for 12 years now."
Hello...
Asnain kolarkar here..
First you change the tube to pure pentode.. El34 only...... Not 6ca7 is a beam pentode the beam former will not accelerate the electrons like pure pentode..... And about the bias +-3-4 ma will not make more difference . .
That is what makes the power and the imaging!! Which means, each tube will be on its own when delivering power!! . .And
Each tube could have a different bias voltage giving it its idle current, I suggested fairly matched tubes to be used so you can have one bias pot...
And.. Don't modifie any resistor or anything just maintain your line voltages ... I'm building a3 amp under Jack sir.. Guidance and I have all the rights from him to build a3 amp and sell here in India and outside... So if you have any problem again let me know.
This amp .was sold to Australian person..... Specially for headphones too
Kolarkar audio & research
India 🇮🇳
Kolarkar Auido,
Perhaps I missed something.
The only schematic I see in this post is Post # 1, a single ended EL34 in Ultra Linear, and there is no other negative feedback in the amplifier.
UL is the only negative feedback.
The schematic does not show the connection of the Suppressor grid, g3, Pin 1. And, the 6CA7 beam formers are tied to Pin 1.
So, I am guessing Pin 1 is tied to the cathode Pin 8; the most common application of an EL34 and a 6Ca7.
You said the Beam former of the 6CA7 does not accelerate the electrons like a pure pentode.
Well . . .
Unless you connect the beam formers to a positive voltage, and unless you connect the pentode suppressor grid to a positive voltage . . .
There is very little acceleration of electrons of either tube type; and the few electrons that are collected by both g3 and beam formers . . . are secondary electrons from the plate.
As to the difference of the performance and sound of an EL34 versus a 6CA7, there may be some difference; primarily because there is no global negative feedback that sometimes covers up differences.
But it is also true that different EL34s will not all perform and sound the same; and it is true that different 6CA7s will not perform and sound the same.
Perhaps I missed something.
The only schematic I see in this post is Post # 1, a single ended EL34 in Ultra Linear, and there is no other negative feedback in the amplifier.
UL is the only negative feedback.
The schematic does not show the connection of the Suppressor grid, g3, Pin 1. And, the 6CA7 beam formers are tied to Pin 1.
So, I am guessing Pin 1 is tied to the cathode Pin 8; the most common application of an EL34 and a 6Ca7.
You said the Beam former of the 6CA7 does not accelerate the electrons like a pure pentode.
Well . . .
Unless you connect the beam formers to a positive voltage, and unless you connect the pentode suppressor grid to a positive voltage . . .
There is very little acceleration of electrons of either tube type; and the few electrons that are collected by both g3 and beam formers . . . are secondary electrons from the plate.
As to the difference of the performance and sound of an EL34 versus a 6CA7, there may be some difference; primarily because there is no global negative feedback that sometimes covers up differences.
But it is also true that different EL34s will not all perform and sound the same; and it is true that different 6CA7s will not perform and sound the same.
And greatly increased 2H. Doesn't anybody use first principles to design anymore?
A 2nd gross error. a 50K AC loadline vs 100K DC loadline on the driver, Another gross source of 2H,
easily plotted on the tube plate curves.
This circuit is a great example of why I don't spend much time here on DIY anymore. Bunk.
Trying to educate those with strange beliefs already soldered into their brains is a time waster.👎
jhstewart9,
Thanks for your contributions to diyAudio's Tubes / Valves threads.
Thanks for your articles in Glass Audio.
And thanks for giving me another person to compete against when I worked in Tektronix Engineering, Marketing, and Technical support.
The old, slightly aged, 1985 HP catalog was one of my best tools to compete with HP.
When those products went out of long term support, there was a goldmine of sales opportunity for me.
My most memorable time of talking to HP engineers, marketeers, and sales persons, was at the 1989 MTTS symposium at Long Beach, Calif.
Tektronix and HP spent 2 evenings doing show-and-tell to each other; after the booths were closed to the public.
I think I talked to Siegfried Linkwitz before I even knew who, and what he was, Mr. 8568 and/or 8566, and Mr. loudspeaker crossover engineer. RIP.
At that 1998 MTTS symposium, Tektronix, 25 years in spectrum analyzer business, . . . Introduced the 2782 spectrum analyzer that was much more advanced than the 8566, something that most people thought could not be done.
Those were great T&M days for both companies.
Much has changed in the T&M industries, and the companies that use T&M equipment.
The Golden Days of Audio.
The Golden Days of T&M.
Thanks for your contributions to diyAudio's Tubes / Valves threads.
Thanks for your articles in Glass Audio.
And thanks for giving me another person to compete against when I worked in Tektronix Engineering, Marketing, and Technical support.
The old, slightly aged, 1985 HP catalog was one of my best tools to compete with HP.
When those products went out of long term support, there was a goldmine of sales opportunity for me.
My most memorable time of talking to HP engineers, marketeers, and sales persons, was at the 1989 MTTS symposium at Long Beach, Calif.
Tektronix and HP spent 2 evenings doing show-and-tell to each other; after the booths were closed to the public.
I think I talked to Siegfried Linkwitz before I even knew who, and what he was, Mr. 8568 and/or 8566, and Mr. loudspeaker crossover engineer. RIP.
At that 1998 MTTS symposium, Tektronix, 25 years in spectrum analyzer business, . . . Introduced the 2782 spectrum analyzer that was much more advanced than the 8566, something that most people thought could not be done.
Those were great T&M days for both companies.
Much has changed in the T&M industries, and the companies that use T&M equipment.
The Golden Days of Audio.
The Golden Days of T&M.
Honest question: Why do you need balanced bias between the channels? If I remember my physics correctly, to be noticeable, the imbalance needs to be greater than 3dB in volume. That's a power ratio of 2:1. Anyone who tries can bring two channels, with new or equally worn tubes, in range.