I am getting confused about an issue my line stage is having.
The stage, easier to describe than picture probably: Elma input selector - Elma 100k stepped attenuator with 470k resistors parallel to the output - transformer coupled 6SN7 using Sowter 1367 transformers running at 200V, LED bias at 4.8V. Heaters are DC.
Recently, as the stage heats up it appears that a few volts DC is appearing on the input connectors and I am struggling to even think of a mechanism by which this would happen given that the attenuator and the parallel resistors should by definition reference pins 1 and 4 to ground.
I have replaced the valve socket, it is happening with either of the valves I have to hand.
Before I go completely bonkers, am I missing something obvious?
The stage, easier to describe than picture probably: Elma input selector - Elma 100k stepped attenuator with 470k resistors parallel to the output - transformer coupled 6SN7 using Sowter 1367 transformers running at 200V, LED bias at 4.8V. Heaters are DC.
Recently, as the stage heats up it appears that a few volts DC is appearing on the input connectors and I am struggling to even think of a mechanism by which this would happen given that the attenuator and the parallel resistors should by definition reference pins 1 and 4 to ground.
I have replaced the valve socket, it is happening with either of the valves I have to hand.
Before I go completely bonkers, am I missing something obvious?
Input connector open, nothing connected:
A Gassy 6SN7, and . . .
A medium to high Ohms Rg resistor to ground, or a medium to high Ohms input attenuator (volume control) . . .
Will cause DC Volts across Rg or the input attenuator.
Feedback parts including a leaky capacitor that has DC volts at one end, will put DC volts at the other end.
Is there rosin and dirt between the socket pins, or wiring, or PCB?
Makes a great resistor, conducts DC.
Sorry, Fluorine was banned, but alcohol cleans.
A complete and accurate schematic is worth 1000 words, and saves 50 Posts from being needed in a thread.
Now, back to doing those properly designed and properly conducted blindfold listening tests . . . .
A Gassy 6SN7, and . . .
A medium to high Ohms Rg resistor to ground, or a medium to high Ohms input attenuator (volume control) . . .
Will cause DC Volts across Rg or the input attenuator.
Feedback parts including a leaky capacitor that has DC volts at one end, will put DC volts at the other end.
Is there rosin and dirt between the socket pins, or wiring, or PCB?
Makes a great resistor, conducts DC.
Sorry, Fluorine was banned, but alcohol cleans.
A complete and accurate schematic is worth 1000 words, and saves 50 Posts from being needed in a thread.
Now, back to doing those properly designed and properly conducted blindfold listening tests . . . .
Does the DC voltage present on the grid go away when the volume control is set at minimum? It should.
If there are "a few volts" on the grid when the control is at maximum, the current must be on the order of 30-50 uA.
The 6SN7 either has excessive grid current, or else there is a leakage path from the cathode or plate to the grid.
If there are "a few volts" on the grid when the control is at maximum, the current must be on the order of 30-50 uA.
The 6SN7 either has excessive grid current, or else there is a leakage path from the cathode or plate to the grid.
A 6SN7 with 200V Plate to Cathode; and 4.8V cathode to grid bias; will have approximately 10mA plate current.
The plate dissipation is 2 Watts (for a 5 Watt rated plate).
Even with nothing connected to the input connector, and the volume control full, the grid "sees" 77k Ohms to ground.
There better not be any significant grid voltage under that condition.
I have measured 6L6GC, KT77, EL34, KT88, KT66, 5881, and 7591 grid voltage (in Triode Wired mode, and in Ultra Linear Mode), and the grid resistor to ground was 100k Ohms.
Those typically had no more than +10mV on the grid. Those are all Power Output tubes, not a medium power tube.
10mV and 100k Ohms, Only has 100nA grid current (0.1uA grid current). Sometimes I selected tubes because if they had 20mV or more, I replaced them.
Are all sockets, tie points, volume control, wires clean and without dirt, grease, etc.? If they are clean, and you have grid volts, then . . .
Get a good 6SN7.
Done
Oops, perhaps you have created a very high frequency oscillator.
In that case, install a 1k Ohm grid stopper resistor, right at the socket grid pin.
Now done.
The plate dissipation is 2 Watts (for a 5 Watt rated plate).
Even with nothing connected to the input connector, and the volume control full, the grid "sees" 77k Ohms to ground.
There better not be any significant grid voltage under that condition.
I have measured 6L6GC, KT77, EL34, KT88, KT66, 5881, and 7591 grid voltage (in Triode Wired mode, and in Ultra Linear Mode), and the grid resistor to ground was 100k Ohms.
Those typically had no more than +10mV on the grid. Those are all Power Output tubes, not a medium power tube.
10mV and 100k Ohms, Only has 100nA grid current (0.1uA grid current). Sometimes I selected tubes because if they had 20mV or more, I replaced them.
Are all sockets, tie points, volume control, wires clean and without dirt, grease, etc.? If they are clean, and you have grid volts, then . . .
Get a good 6SN7.
Done
Oops, perhaps you have created a very high frequency oscillator.
In that case, install a 1k Ohm grid stopper resistor, right at the socket grid pin.
Now done.
Have done that on both sockets as a matter of course, the tag boards for hifi collective as per https://www.hificollective.co.uk/tag_boards/glasshouse-octalux4-valve-holder-board.html make this very easy.
How many 6SN7 tubes have you tried?
How many Volts are you getting on the grid with the volume control turned all the way up, and nothing connected at the input connector?
Even a grid stopper can not stop high frequency oscillations that are caused by wiring placement, wire lengths, etc.
12 Posts and we do not seem to be closer to solving your problem.
How many Volts are you getting on the grid with the volume control turned all the way up, and nothing connected at the input connector?
Even a grid stopper can not stop high frequency oscillations that are caused by wiring placement, wire lengths, etc.
12 Posts and we do not seem to be closer to solving your problem.
The voltage on the input on the right channel (triode 1 as I have wired it) was varying with volume setting up to about the planned bias voltage. Everything else would need to depend on having time without small children to measure with the lid off.
The right channel was what I measured to start with as the symptom was the signal fading out from my DAC on that side unless the volume setting was at full. The left channel (triode 2) did not react in the same way.
I have took the input selector and attenuator out into another box for now to try and simplify problem solving and give scope for playing with layouts/wiring.
The right channel was what I measured to start with as the symptom was the signal fading out from my DAC on that side unless the volume setting was at full. The left channel (triode 2) did not react in the same way.
I have took the input selector and attenuator out into another box for now to try and simplify problem solving and give scope for playing with layouts/wiring.
euro21,
If you short out the grid leak resistor Rg, there better be Zero, 0.000 Volts, on the grid (Rg is from the grid to ground).
For a properly working 6SN7, if the grid is at zero volts,and the plate load resistor has B+ at the top and the plate at the bottom; and the cathode is tied to two 2.4 Volt LEDs in series, there better be 4.8 Volts on the cathode. With 2.4 Volts across each LED, they will be lighted (if they are broken, the voltage across each LED will be something other than 2.4V).
Anything other than that, is in another universe.
If you short out the grid leak resistor Rg, there better be Zero, 0.000 Volts, on the grid (Rg is from the grid to ground).
For a properly working 6SN7, if the grid is at zero volts,and the plate load resistor has B+ at the top and the plate at the bottom; and the cathode is tied to two 2.4 Volt LEDs in series, there better be 4.8 Volts on the cathode. With 2.4 Volts across each LED, they will be lighted (if they are broken, the voltage across each LED will be something other than 2.4V).
Anything other than that, is in another universe.
@6A3sUMMER
I think that this /half/ of tube is defected, may occur relatively low resistance between grid and cathode (if the tube is warmed up).
If the light of LEDs is altering (shorting grid to ground), even extinguished, this is the reason.
Update on this: I first split the volume/input selector off into a separate box as effectively a passive pre into a valve gain stage/buffer, which gave a bit more space to arrange so that the LEDs and ground reference resistors were on separate tagboards. This also swapped over which halves of the valve did which channel to keep the wiring vaguely sensible. Result: slightly more time before DC kicks in but it does, only this time on the left channel rather than the right.
At this point I checked the heater voltage just to be sure. The power supply as originally built had an LT1085 set to deliver pretty much 6.3V DC. In theory...
Following some very odd readings, I then dug out a 6V traffic and set up a CRC smoothing circuit to actually give 6.3V and it appears to be problem solved. Am starting to be distinctly suspicious of how some 3 pin regulators behave in practice...
At this point I checked the heater voltage just to be sure. The power supply as originally built had an LT1085 set to deliver pretty much 6.3V DC. In theory...
Following some very odd readings, I then dug out a 6V traffic and set up a CRC smoothing circuit to actually give 6.3V and it appears to be problem solved. Am starting to be distinctly suspicious of how some 3 pin regulators behave in practice...