This may be a dumb question, I'm not sure. I'm on my second tube amp build and starting to thinking a bit harder about the details. Here goes:
In regards to noise, does the heater phase matter?
For example, lets say I have a two stage SE preamp.
First stage is a grounded grid which runs into a cathode follower- oldest topology in the book probably.
Now let's say I have two heater windings both as part of the main power transformer.
On the grounded grid stage I have the heater windings in one phase, but the cathode stage heater windings are flipped, so 180 degrees out of phase with the first stage. Would I get noise cancellation benefit in 60Hz, but also possibly in 120/180/240/300?
Seems that there can be some way to benefit from hum cancellation even in a SE preamp, especially if there are two stages. I tried searching but I haven't found too much talk of this.
In regards to noise, does the heater phase matter?
For example, lets say I have a two stage SE preamp.
First stage is a grounded grid which runs into a cathode follower- oldest topology in the book probably.
Now let's say I have two heater windings both as part of the main power transformer.
On the grounded grid stage I have the heater windings in one phase, but the cathode stage heater windings are flipped, so 180 degrees out of phase with the first stage. Would I get noise cancellation benefit in 60Hz, but also possibly in 120/180/240/300?
Seems that there can be some way to benefit from hum cancellation even in a SE preamp, especially if there are two stages. I tried searching but I haven't found too much talk of this.
Hum from heater gets amplified by 1st stage tube, but hum from 2nd stage CF does not.
Will it cancel then .... ?
Will it cancel then .... ?
Heater to cathode at peaks if you’re on the limit or elevated may need consideration but you’d have the same problem regardless of phase.
Noise has several origins. Looking at HV current induced voltages (switching noise), two LV windings result in double noise voltage if no shield has been applied. The distance from heater to cathode is small and results in a couple pF coupling. RF circuits may for this reason decouple heaters to ground with small capacitors.
Heater hum is caused by poor quality valve manufacturing where the heater 'leaks' into the cathode. Filament heaters need a balanced supply as one side will counteract the other so manufacturers either have a heater supply with a centre ground or use a virtual ground.
If you have to swop phases then replace the valve for a better quality valve and use the faulty one further down the chain where gain is not important.
If you have to swop phases then replace the valve for a better quality valve and use the faulty one further down the chain where gain is not important.
I'm not talking about a faulty tube. Even if perfectly wired, good tubes, and no noise from other sources like B+, if you put the output on a spectrum analyzer you will see bumps at 60 and some harmonics. Let's not get sidetracked with all the other reasons for noise etc. Will cancellation happen in series stages if heaters wired in opposite phases at each stage. If not, why not?
It depends on how the hum is ingressing each stage.
Both stages may suffer from resistive leakage across the cathode - heater interface. That leakage is very valve dependant, but is suppressed by a lower value of cathode load, and by maintaining a voltage difference (elevation) between heater and cathode that is not say less than 15-20V at heater waveform peak, or above stress limits.
And there also may be capacitive coupling to the second stage grid, which may be wiring layout dependant - a typical reason for having a humdinger pot available.
Its also plausible that the first stage heater may be asymmetric within and as it leaves the cathode tube, with some net capacitive coupling from one side of the heater to the cathode that isn't balanced by the other side of the heater, and the net effect accentuated by first stage gain.
Heater voltage is typically not a nice sinewave, and flat-topping from the mains source, and/or via the B+ rectification, may also be a significant hum contributor due to the harmonics not balancing out.
Due to input stage gain, there may or may not be benefit from swapping heater phases, depending on if one stage is the dominant hum contributor or not.
Both stages may suffer from resistive leakage across the cathode - heater interface. That leakage is very valve dependant, but is suppressed by a lower value of cathode load, and by maintaining a voltage difference (elevation) between heater and cathode that is not say less than 15-20V at heater waveform peak, or above stress limits.
And there also may be capacitive coupling to the second stage grid, which may be wiring layout dependant - a typical reason for having a humdinger pot available.
Its also plausible that the first stage heater may be asymmetric within and as it leaves the cathode tube, with some net capacitive coupling from one side of the heater to the cathode that isn't balanced by the other side of the heater, and the net effect accentuated by first stage gain.
Heater voltage is typically not a nice sinewave, and flat-topping from the mains source, and/or via the B+ rectification, may also be a significant hum contributor due to the harmonics not balancing out.
Due to input stage gain, there may or may not be benefit from swapping heater phases, depending on if one stage is the dominant hum contributor or not.
Last edited:
Anchan,
You said the input stage is a Grounded Grid stage.
That means you are sending the input signal to the Cathode of the input tube.
Right?
If not, it is not a Grounded Grid input stage.
How about posting a schematic, please?
You said the input stage is a Grounded Grid stage.
That means you are sending the input signal to the Cathode of the input tube.
Right?
If not, it is not a Grounded Grid input stage.
How about posting a schematic, please?
He is probably referring to a stage where the grid is "DC grounded" through kilo ohms of carbon, not "AC grounded" through a capacitor 😀
Short answer is possibly, partially, but the amount is likely negligible. Sorento stated the reason.
Since the problem is very undefined, many scenarios are possible but this is the simplest answer.
Calling a Common Cathode amplifier stage a Grounded Grid amplifier . . .
. . . is like saying that a World War II Thunderbolt Fighter Airplane, has an Inline 6 Cylinder Engine.
Not even close!
. . . is like saying that a World War II Thunderbolt Fighter Airplane, has an Inline 6 Cylinder Engine.
Not even close!
Ok sorry. I got my terminology mixed up. I didn't want to crowd out the 6v6 preamp thread, so posted here. Basically I built the 6v6 preamp gain stage and it is feeding the 6v6 cathode follower preamp stage in the Salas thread. But I was really just trying to ask if there are two SE stages in succession, one with the heaters connected one way, and the other stage heaters are reversed in polarity, and if that would result in some cancellation. I tried it quickly but haven't had time to analyze it, so I was really just asking if others have done it, if it's a thing, and if offered a theoretical advantage. Could be any SE tube stages in succession. Sorry for any confusion
Consider a first stage that has 20dB gain (10X voltage).
Consider a second stage that has 20dB gain (10X voltage).
Now, if the first stage (isolated by itself) 10uV hum caused by its filament to cathode leakage;
And if the second stage (isolated by itself) has 10uV hum caused by its filament to cathode leakage;
Then the 10uV hum of the first stage gets amplified by the second stage, 10uV x10, so it is 100uV.
If the phase is just right to cancel, you have 100uV -10uV = 90uV hum.
You get a reduction of hum by 10%, -1dB.
Just plug in the hum level of your own filament to cathode leakage for each individual stage (by itself).
Then plug in the gain of the second stage, and calculate the hum reduction improvement just like I did.
Does that answer the question?
Consider a second stage that has 20dB gain (10X voltage).
Now, if the first stage (isolated by itself) 10uV hum caused by its filament to cathode leakage;
And if the second stage (isolated by itself) has 10uV hum caused by its filament to cathode leakage;
Then the 10uV hum of the first stage gets amplified by the second stage, 10uV x10, so it is 100uV.
If the phase is just right to cancel, you have 100uV -10uV = 90uV hum.
You get a reduction of hum by 10%, -1dB.
Just plug in the hum level of your own filament to cathode leakage for each individual stage (by itself).
Then plug in the gain of the second stage, and calculate the hum reduction improvement just like I did.
Does that answer the question?
Last edited:
That makes sense. I hadn't considered the gain of each stage as being a factor. Will try to do this in reality and measure with an old quantanalyzer and scope I have and report back.
In my case, from memory the first stage is 15db gain, and I measured between -75 and -85 db noise peaks in the 60/120/180 Hz range. It goes into the cathode follower stage which of course is just shy of unity gain, and I had -90db peaks around those same points. Will try in real life.
Thanks!
In my case, from memory the first stage is 15db gain, and I measured between -75 and -85 db noise peaks in the 60/120/180 Hz range. It goes into the cathode follower stage which of course is just shy of unity gain, and I had -90db peaks around those same points. Will try in real life.
Thanks!
Last edited:
If you have a volume control between the 1st and 2nd stage (like in many guitar amps) there can be a point on the control where the 2 hums cancel. Same with PS hum.
In electronics, Nulls are Very Sharp (just happen at one Needle Point).
Special Cases are abundant, often not practical.
Keep hum at low levels by using other time tested techniques.
Works more often, works under more conditions, for several years, etc.
Causes of hum:
Poor B+ filtering; tubes that have leaky filament to cathode insulation, especially with un-bypassed cathodes; poor orientation of signal transformers versus power transformers and chokes, and too close spacings to signal transformers; poor wiring dress and placement; and Ground Loops.
B+ Ground Loops; and input connector/input stage/signal source ground loops.
Any and all of these may get you every time.
Special Cases are abundant, often not practical.
Keep hum at low levels by using other time tested techniques.
Works more often, works under more conditions, for several years, etc.
Causes of hum:
Poor B+ filtering; tubes that have leaky filament to cathode insulation, especially with un-bypassed cathodes; poor orientation of signal transformers versus power transformers and chokes, and too close spacings to signal transformers; poor wiring dress and placement; and Ground Loops.
B+ Ground Loops; and input connector/input stage/signal source ground loops.
Any and all of these may get you every time.
Last edited:
I'm aware of the causes of hum, and have minimized. The question was very specific related to phase.
I once had a constant current source for the cathodes of a dual triode phase splitter.
The constant current source consisted of a choke and a resistor in series to properly bias the phase splitter triodes.
The choke had enough inductance to work well at low frequencies, and a low enough distributed capacitance to work well at high frequencies.
All worked very well, or so I thought.
But the "CCS" choke was picking up magnetic fields from either the power transformer or from the B+ choke (lots of hum).
There was no hum at the amplifier output. But the power out was too low.
The common mode hum of the choke and phase splitter, was sending common mode hum to the push pull output tubes.
Since it was common mode, it cancelled because of the push pull circuit.
But . . . It robbed the amplifier of power, the power of the common mode hum was subtracting power from the total output power of the push pull stage.
Live and learn.
PS (Post Script)
A solid state CCS should eliminate the above problem, but not necessarily, here is why:
Suppose the filament to cathode leakage is faulty (a low resistance). That will cause common mode hum, just as the choke "CCS" did.
"There are more things that can go wrong, than that can go right"
The constant current source consisted of a choke and a resistor in series to properly bias the phase splitter triodes.
The choke had enough inductance to work well at low frequencies, and a low enough distributed capacitance to work well at high frequencies.
All worked very well, or so I thought.
But the "CCS" choke was picking up magnetic fields from either the power transformer or from the B+ choke (lots of hum).
There was no hum at the amplifier output. But the power out was too low.
The common mode hum of the choke and phase splitter, was sending common mode hum to the push pull output tubes.
Since it was common mode, it cancelled because of the push pull circuit.
But . . . It robbed the amplifier of power, the power of the common mode hum was subtracting power from the total output power of the push pull stage.
Live and learn.
PS (Post Script)
A solid state CCS should eliminate the above problem, but not necessarily, here is why:
Suppose the filament to cathode leakage is faulty (a low resistance). That will cause common mode hum, just as the choke "CCS" did.
"There are more things that can go wrong, than that can go right"
Last edited: