Heaters & Filaments: AC vs. DC

[nazaroo]

In an interesting post on another forum, I came across this disturbing phenomenon. I haven't confirmed the claim, but it seems to deserve examining.

IN my view this may only be relevant to cathodes in 'modern' beam power tubes, and not in classic thoriated tungsten or carbon heater/filaments.

Just thought I'd pass this along...

We thought we'd box clever and cheat a little on a new series of amps - in order to reduce the build time, PT weight and size we thought we'd use 12VDC to heat the tubes... no need to laboriously twist heater wires... just hook the damned things up in series and forget about it... WRONG

:

Pre-amp and PI don't seem to be affected but power tubes don't like it... why? Something called 'plating' occurs... where the cathode becomes charged and thermionic emission becomes inhibited... takes a while to happen but eventually the amp starts sounding a bit thin and weedy... I'd been warned by the guy who helped us develop the DC heater supply that this might happen and sure enough it did...

So now we merrily twist our heater wires (we're use a very thin silver plated PTFE jacketed wire) and have just made up some jigs to knock up heater wire assemblies before offering them into the amps and hey presto.... very quiet operation and no plating... we're still acheiving the goal of smaller and lighter PT's by virtue of using 12.6VAC (in series of course) and there is no audible difference...

So according to this experimenter,
there is a definite cathode-smothering effect on some common beam tetrode/pentodes, traceable to D.C. heater-current.

The explanation appears to be 'magnetization/saturation' and/or a static field causing attenuation of cathode emission and plate current.

If this is true, it should be traceable to:

(1) materials used to manufacture the heater filament.

(2) physical size, length, and/or orientation/polarity of the cathode-heaters and/or cathode arrangement.

Any thoughts?

This seems important, as it implies that for yet another reason, we can't generalize regarding "best practices",
without specifying the actual tubes to apply the concepts to.

 

[nazaroo]

This is especially disconcerting, as experienced builders like Tubelab.com have noted the following:

Observations so far:

The 10 amp filament puts out a serious AC magnetic field that gets into everything. Hum is evident in the speakers and on the analyzer BEFORE you turn on the power supply. As soon as you plug in the filament transformer the hum appears. Moving the transformers or the tube around has no effect. The hum balance pot has no effect on this hum. DC filaments WILL be required.

This suggests strongly that D.C. heater supplies are actually preferred,
from the perspective of noise-free amplification.

 

[popilin]

I have not thought about the problem, well I think slowly.
But 10A AC surely is not the problem, from Maxwell's equations, we can see that the source of magnetic field due to AC is voltage, not current.
The noise is from the AC electric field.

Returning to the initial problem, how to charge a cathode if it is at a potential close to zero if fixed bias is used?
What is the potential difference of cathode to filament?
We should probably look that way, the DC electric field due to cathode-filament voltage can retain/push some electrons produced by thermionic effect.
With an AC electric field the effect is less perceptible.

Moral: Make the voltage cathode-filament negative, floating filament at a potencial higher or closer than cathode potential, and enjoy DC.


Best regards
Johann

 

[DigitalJunkie]

Moral: Make the voltage cathode-filament negative, floating filament at a potencial higher or closer than cathode potential, and enjoy DC.

That's what I was thinking. Elevating the heater circuit above ground(EDIT: above cathode voltage(s).) should remedy this,no?

 

[popilin]

That's what I was thinking. Elevating the heater circuit above ground(EDIT: above cathode voltage(s).) should remedy this,no?

Classical Electrodynamics says yes.

 

[trobbins]

Sometimes it is a little too easy to shoot from the hip when offering explanations about what really is just a snippet of information.

I don't think that one marketing blog lends much information to allow a reasonable appreciation of a technical issue. Not that that should stiffle discussion, but better for the discussion to start with eeking out what the blogger was actually discussing for starters, and have others described such an issue in the last 60 years.

I also suggest digging out those text books again about EM fields and induction, and also looking around for a few tutorials on the subject - as most will indicate it is actually the heater current that 'does the damage' so to speak.

And who knows what extreme configuration tubelab was referring to - a heater for one tube? or many tubes? I have a heater supply requiring more than 25A from a PT that also services HT, and there is no hum issue at all - but as I haven't indicated the configuration, it could be hard for you to guess the type of amplifier I'm referring to.

Ciao, Tim

 

[popilin]

I also suggest digging out those text books again about EM fields and induction, and also looking around for a few tutorials on the subject - as most will indicate it is actually the heater current that 'does the damage' so to speak.

You do not need a textbook for something as simple, furthermore Maxwell's equations are in the forum, here

http://www.diyaudio.com/forums/tubes-valves/181072-designing-transformers-j-c-maxwell.html

If you can follow the math, you'll see that equation (21) clearly indicates that the AC magnetic field source is the AC voltage.
While the equation (28) tells us that the DC magnetic field source is the DC current.
As shown in equation (14), the AC voltage generates an AC electric field, which affects the grid, introducing noise.

But this reasoning is flawed, was thought to derive the equations of transformers.
Equation (4) shows a time-varying current produces a time-varying electromagnetic field, and then electromagnetic radiation, which induces noise through the grid.

I apologize, Tubelab.com reasoning is correct, only that it is poorly expressed.
It was not my intention to offend anyone, but it happens I'm going crazy doing several transformers, I'm kind of obsessed with transformers, the tree will not let me see the forest, also as I said, I think slowly.

Thanks Tim for making me see my mistake.
I think the following reasoning was correct.

Best regards
Johann

 

[nazaroo]

Originally Posted by DigitalJunkie
That's what I was thinking. Elevating the heater circuit above ground(EDIT: above cathode voltage(s).) should remedy this,no?

Classical Electrodynamics says yes.

Popilin and Junkie:

I think you've got the right idea, but one thing backwards:


If you want to drive away electrons from the heater,
and leave them available for tube current to plate,
then you want your heater more NEGATIVE, i.e.,
repelling the electrons and leaving them floating
in the vacuum ready for attraction to screen and plate.

Thus if the effect were purely electrical (electric field),
I would have to say you "float" your heaters (and entire D.C. supply) BELOW Ground, say about 50 volts!.

If they are separated from the Cathode electrically,
and you have a -ve Bias supply already, why not use it to
submerge your heaters -40 volts below ground?

But if the effect is in part magnetic (and lets see how this could come about,
other solutions must be added.

We all know that D.C. current magnetizes many metals and alloys,
i.e., iron, steel, nickel, etc. and that both heaters and cathodes
are made of various materials.
It may be that in some tubes, either the heaters, or the cathodes (or both) can be magnetized.

If so, the next question is:
How or why would this magnetization affect the electron release and/or
the electron cloud?

Finally, could the original poster be mistaken or unclear in
his description of the phenomena?
Could he be referring to the well-known (or supposedly better understood)
polarization of the heater in a Directly heated cathode,
whereby bias (and current) is affected across the filament surface?

 

[DF96]

from Maxwell's equations, we can see that the source of magnetic field due to AC is voltage, not current.

Any current generates a magnetic field, both AC and DC. So an AC current generates an AC field. In addition, an AC voltage generates an AC magnetic field.

 

[zigzagflux]

the source of magnetic field due to AC is voltage, not current

With a shielded power cable with grounded copper shield, I can energize the inner conductor with voltages over 69 kV and would not be able to tell the conductor is energized using either electric or magnetic field sensitive equipment. So much for that magnetic field. Remove the shield and I can detect the presence of voltage. But pass 1A through that shielded cable, and I can clamp it and measure that current by means of the magnetic field which blows right past the copper shield.

I think the fact that Maxwell's equations recognize an AC magnetic field propagates an AC electric field and vice versa adds to the confusion on this subject, but IME you can simplify in practical application by knowing that current produces the magnetic field in significance. AC voltage produces an electric field in significance.

That is why you can shield with copper from energized wires, but twist to cancel magnetic fields in heater leads. If the heater leads are not carrying the copious current, they don't induce into the magnetics of the amp.

 

[DF96]

It depends on frequency. At 60Hz the magnetic field created by an AC voltage is very small. At 60MHz it is a million times bigger, because the magnetic field comes from the time differential of the voltage.

 

[DavesNotHere]

I'll share some data with heaters with ya.

signal tubes run more evenly throughout their life span is heater voltage is approximately 80-90% of typical heater voltage rating.

when you either use ac or dc, the voltage should be clean. and noise on the heater voltage (usually caused by lower quality power components) will be impressed onto the cathode via heater-cathode detection ( a diode action effect and electron modulation via magnetic induction (the dual of the inter electrode capacitance between the heater and cathode that rectifies the noise). ) to examine this phenomenon, take an oscilloscope to ground and measure this noise on the cathode with just the heater voltage applied.

 

[DavesNotHere]

there are tactics to suppress the noise on ac heaters. the one I developed takes the hum balance circuit and instead of directly connecting the center tap of the hum circuit to ground, it is AC coupled to the ground by a capacitor ( I typically use a .1uf poly or pio cap for this)

but none of these things are new. I imagine there are things out there that used this that were made way before I was born.

 

[DF96]

If you do this you still need to establish a DC reference for the heater circuit.

 

[DavesNotHere]

also you can use this effect to create a automatic power noise reduction system by tying the center tap of the hum balance to a capacitor based voltage divider network ( two caps in series instead of resistors)

 

[franzm]

About what kind of tubes you are talking, direct (heater=catode) or indirect heated?
Magnetic field of the heater depends on the geometry of the heater, especially the folding scheme of the tungsten wire coil.

 

[DavesNotHere]

If you do this you still need to establish a DC reference for the heater circuit.

not really cause the diode effect of the heater to cathode will dynamically cause a dc reference to form in the hum balance tap.


and if the heater needs to be elevated ( some cascode and dc couple circuits) the center tap would be on a rc voltage divider network (a cap and resistor divider network with their center taps tied together and connected to hum balance tap.

 

[Bigun]

I read somewhere that with dc supplies it's recommended to reverse the polarity from time to time. If I remember correctly, it was something to do with dissimilar metals and the like and deterioration at contacts. Anyhow, can't remember where I saw it (somewhere on the internet) but the recommendation was simply to reverse the polarity every once in awhile. I think the poster had some fancy scheme to auto-change direction at power up.

 

[DF96]

not really cause the diode effect of the heater to cathode will dynamically cause a dc reference to form in the hum balance tap.

So the valve manufacturers who specify in their datasheets that there should be a maximum resistance of, say, 20k between heater and cathode (except for a CF when it can be higher) don't understand their own devices?

I read somewhere that with dc supplies it's recommended to reverse the polarity from time to time.

Yes, I have seen that somewhere too; something to do with metal migration or electrochemistry. I have also seen it declared as nonsense. Does anyone know the truth on this? It sounds plausible. It could also be an audiophile myth.

 

[Yvesm]

I read somewhere that with dc supplies it's recommended to reverse the polarity from time to time.

YES !
I usually do that fifty times per second