This is an amp I recently built. It sounds good but the NFB circuit resistor values are just guess work because calculating the right values is beyond my ability at this point. Can the two cathode resistors in the preamp be the same value as they are or should the lower resistor be smaller? Also, I wanted a bit more gain to drive the 6V6 so I added that bypass capacitor to the top resistor. Is that a correct way of increasing the gain? Also, what if the lower resistor is eliminated and the top resistor made 1.5K with the bypass capacitor in place. Would that just short the NFB to ground?
Nice idea!
However,
You listed a 510 Ohm resistor and 20V at the 6V6 cathode, and 297V B+ before the DCR voltage drop of the output transformer's primary.
20V / 510 Ohms = 39ma cathode current.
I think you will get less current than 39mA, so the cathode voltage will probably will be less than 20V.
Take a look at the 6V6GT tube date curves.
I looked at one data sheet, the curves indicated for a screen at 285V and plate at 285V, with the cathode self bias at 20V, you only get 25mA plate current.
(cathode current = plate current + screen current).
If the plate current is 25mA, you would need to get 14mA screen current to get 39mA cathode current (is not going to happen).
And at 14mA screen current and 285V, the screen would dissipate 4.3 Watts, OUCH! screen burn out.
look at those data sheet curves, and re-calculate the cathode resistor.
Good luck designing, building . . . and happy listening!
However,
You listed a 510 Ohm resistor and 20V at the 6V6 cathode, and 297V B+ before the DCR voltage drop of the output transformer's primary.
20V / 510 Ohms = 39ma cathode current.
I think you will get less current than 39mA, so the cathode voltage will probably will be less than 20V.
Take a look at the 6V6GT tube date curves.
I looked at one data sheet, the curves indicated for a screen at 285V and plate at 285V, with the cathode self bias at 20V, you only get 25mA plate current.
(cathode current = plate current + screen current).
If the plate current is 25mA, you would need to get 14mA screen current to get 39mA cathode current (is not going to happen).
And at 14mA screen current and 285V, the screen would dissipate 4.3 Watts, OUCH! screen burn out.
look at those data sheet curves, and re-calculate the cathode resistor.
Good luck designing, building . . . and happy listening!
@6A3sUMMER Attached is the 6V6 load line I was going by. It specifies a screen voltage of 250 volts but I didn't pay any attention to that. Instead I'm measuring 294 volts on the screen so that's gotta come down.
There are lots of variables, including how much your B+ voltage varies (versus load current, and versus the power mains voltage).
Try starting with a 390 Ohm self bias resistor, see what B+ voltage you get, and what cathode current you get.
You could adjust the self bias resistor from there.
Just be sure to dissipate no more than the data sheet specification.
Use the Beam Power maximum dissipation specification, not the Triode Wired max dissipation specification.
Plate voltage - bias voltage, and Screen voltage - bias voltage are almost the same.
Example, 385V - 15V = 370V.
Suppose15V / 390 Ohms = 38.5mA (0.0385A)
370V x 0.0385A = 14.25 Watts plate + screen dissipation, (14 Watts plate plus screen dissipation total, so we are OK).
I suspect your 385 volts B+ will be slightly lower when loaded, and also the screen and plate current will drop the voltage a little, because of the primary DCR times the screen and plate current.
Check your wiring, solder connections, parts values (like color codes on resistors, or printed values on power resistors; capacitor polarity markings, etc.) With no power mains connected, and before first power on, you can check the resistances with a DMM, example, 390 Ohms from cathode to ground.
Be sure to put a load resistor on the output transformer secondary.
We can not see your B+ schematic. Do you have bleeder resistor(s)?
Safety first.
Prevent the "Surviving Spouse Syndrome".
Short the input connector.
Remove the 10k Ohm negative feedback resistor (no global feedback).
Let us know the cathode, screen, and plate voltages with 390 Ohm self bias resistor, it will probably be near optimum.
Measure the cathode current first, to see if it is way to much.
Then, Quick testing of those measurements (tubes warmed up for 20 or 30 seconds, will not ruin the 6V6 even if the dissipation is slightly exceeded, but if the predicted current is lots more than the predicted, you may have incorrect wiring, a wrong part value, or a bad tube.
After all that is OK, AND the power mains are removed, AND the B+ is discharged, then reconnect the negative feedback resistor.
You might have an oscillation.
Barring a problem listed above, you should be able to get it up and running OK.
Try starting with a 390 Ohm self bias resistor, see what B+ voltage you get, and what cathode current you get.
You could adjust the self bias resistor from there.
Just be sure to dissipate no more than the data sheet specification.
Use the Beam Power maximum dissipation specification, not the Triode Wired max dissipation specification.
Plate voltage - bias voltage, and Screen voltage - bias voltage are almost the same.
Example, 385V - 15V = 370V.
Suppose15V / 390 Ohms = 38.5mA (0.0385A)
370V x 0.0385A = 14.25 Watts plate + screen dissipation, (14 Watts plate plus screen dissipation total, so we are OK).
I suspect your 385 volts B+ will be slightly lower when loaded, and also the screen and plate current will drop the voltage a little, because of the primary DCR times the screen and plate current.
Check your wiring, solder connections, parts values (like color codes on resistors, or printed values on power resistors; capacitor polarity markings, etc.) With no power mains connected, and before first power on, you can check the resistances with a DMM, example, 390 Ohms from cathode to ground.
Be sure to put a load resistor on the output transformer secondary.
We can not see your B+ schematic. Do you have bleeder resistor(s)?
Safety first.
Prevent the "Surviving Spouse Syndrome".
Short the input connector.
Remove the 10k Ohm negative feedback resistor (no global feedback).
Let us know the cathode, screen, and plate voltages with 390 Ohm self bias resistor, it will probably be near optimum.
Measure the cathode current first, to see if it is way to much.
Then, Quick testing of those measurements (tubes warmed up for 20 or 30 seconds, will not ruin the 6V6 even if the dissipation is slightly exceeded, but if the predicted current is lots more than the predicted, you may have incorrect wiring, a wrong part value, or a bad tube.
After all that is OK, AND the power mains are removed, AND the B+ is discharged, then reconnect the negative feedback resistor.
You might have an oscillation.
Barring a problem listed above, you should be able to get it up and running OK.
Last edited:
Typically a lower value cathode resistor increases both screen and plate dissipation even though B drops.
markw51,
Of course there is lots more cathode current when the screen is at 294Volts, instead of at 250V.
Because you have chosen Ultra Linear mode, there is not a good way to reduce the screen voltage, unless you also reduce the plate voltage.
You do that by reducing the B+ voltage to the output transformer primary.
If you decide to use a large resistor to the screen, the UL operation will be degraded.
If you decide to use a zener from the UL tap, that is asking for trouble and non-linear opertation (the screen current varies widely, the zener voltage varies, and perhaps even comes out of zener action.
There is nothing wrong with a 6V6 in UL operation and 294V on the screen, just make sure to adjust the self bias resistor so that both the plate dissipation and the screen dissipation are both within their data sheet maximum ratings.
You may have to refer to more than one data sheet.
I do not know what simulation software you are using.
But one thing I noticed (does not make any difference, but is a fine point) you said the load was 5k Reactive.
When the output transformer is loaded with a non-inductive load resistor . . .
Then at mid frequencies, the load is Resistive.
At low frequencies the load is Reactive (inductive).
At high frequencies the load is Reactive (capacitive).
And . . .
When loaded with a Loudspeaker, the Load is Anything It Wants to Be, according to the Loudspeaker and the frequency.
I also see a term g1 listed next to the 38ma current.
g1 does not have any current.
g2 is the screen, and we care about its voltage and current.
You also list power at A and A2.
You are using RC coupling from the input/driver triode to the output tube g1.
There is not any Class A2 operation; and attempt to do that will charge the cap, and shift the bias voltage and current.
As I often do, I added edits to my Post # 6, you may need to re-read it.
Have Fun getting the amplifier up and running!
Of course there is lots more cathode current when the screen is at 294Volts, instead of at 250V.
Because you have chosen Ultra Linear mode, there is not a good way to reduce the screen voltage, unless you also reduce the plate voltage.
You do that by reducing the B+ voltage to the output transformer primary.
If you decide to use a large resistor to the screen, the UL operation will be degraded.
If you decide to use a zener from the UL tap, that is asking for trouble and non-linear opertation (the screen current varies widely, the zener voltage varies, and perhaps even comes out of zener action.
There is nothing wrong with a 6V6 in UL operation and 294V on the screen, just make sure to adjust the self bias resistor so that both the plate dissipation and the screen dissipation are both within their data sheet maximum ratings.
You may have to refer to more than one data sheet.
I do not know what simulation software you are using.
But one thing I noticed (does not make any difference, but is a fine point) you said the load was 5k Reactive.
When the output transformer is loaded with a non-inductive load resistor . . .
Then at mid frequencies, the load is Resistive.
At low frequencies the load is Reactive (inductive).
At high frequencies the load is Reactive (capacitive).
And . . .
When loaded with a Loudspeaker, the Load is Anything It Wants to Be, according to the Loudspeaker and the frequency.
I also see a term g1 listed next to the 38ma current.
g1 does not have any current.
g2 is the screen, and we care about its voltage and current.
You also list power at A and A2.
You are using RC coupling from the input/driver triode to the output tube g1.
There is not any Class A2 operation; and attempt to do that will charge the cap, and shift the bias voltage and current.
As I often do, I added edits to my Post # 6, you may need to re-read it.
Have Fun getting the amplifier up and running!
Last edited:
@6A3sUMMER Ok thanks for your help as always and I will get to work on all that.
https://www.vtadiy.com/loadline-calculators/loadline-calculator/
https://www.vtadiy.com/loadline-calculators/loadline-calculator/
I think the values of the OPs circuit are ok.
The screen current of a 6V6 with equal screen and plate voltage should be around 10% of the plate current (see the 250V SE class A example in the datasheet).
So with a cathode current of 39mA the screen current would be around 4mA, resulting in a screeen dissipation of 1.1W and a plate dissipation of 9.7W which is fine.
Rikaro,
I agree with you, it is OK.
But he asked for my opinion(s), and I gave them.
Plus, those readers who do not know much, will have learned about factors that interact; load impedances;
possible problems with negative feedback (check operation without it first); etc.
Some do not have much in the way of test equipment. That can make troubleshooting difficult.
Simulation software, and plug-in-the-value calculators are great!
They simplify, allow you to do "what if", etc.
But . . . they fall short of teaching how to do real complex analysis.
What if something goes wrong, or is completely different than the expected outcome.
Just my opinion, and my experience.
I agree with you, it is OK.
But he asked for my opinion(s), and I gave them.
Plus, those readers who do not know much, will have learned about factors that interact; load impedances;
possible problems with negative feedback (check operation without it first); etc.
Some do not have much in the way of test equipment. That can make troubleshooting difficult.
Simulation software, and plug-in-the-value calculators are great!
They simplify, allow you to do "what if", etc.
But . . . they fall short of teaching how to do real complex analysis.
What if something goes wrong, or is completely different than the expected outcome.
Just my opinion, and my experience.