Vacuum tubes - DC heaters?

[T3sl4co1l]

Yup.

There is some risk of IMD in DHTs, where the nonlinear cathode-grid transfer function causes some mixing, but this can be avoided by 1. choosing a frequency far above 20kHz (say a few 100 kHz), so that not only signal but harmonics as well are well attenuated at comparable frequencies where IMD might occur, and 2. using low voltages -- since typical DHTs have quite low gain, e.g. 50V grid input required for full swing, and the e.g. 2.5V RMS of the heater is a tiny fraction of that range, and therefore looks like a nearly-linear section of that curve.  That is, applying the calculus principle of continuity of derivatives, where the heater voltage swing is small in comparison to the overall range of the transfer function, and what curvature it has (nonlinearity, distortion), thus the distortion and mixing of that signal is small.  Conversely, types with high filament voltage may be discouraged for such service; this mainly means transmitter types anyway, for which the power level is much higher, or for RF purity reasons, so you might want to avoid the issue entirely and stick with mains/DC for them.

Also easy way to get isolation; if you need to float something with a relatively fragile heater, well, pretty easy to add more windings to a small ferrite transformer and be done with it.

Tim

[Xena E]

Audio amplifiers using DHTs back in the day, most often relied solely on a filament winding on the line transformer that was centre tapped to the signal ground through grid biasing arrangements.

In an experimental small (sub 1W) class A Push Pull amplifier that I was shown by a student, directly heated miniature FSU rod pentodes were used throughout.

The whole demonstration was powered by a 12V SLA battery through an HF SMPS giving the 1.2V HF filament supply, and the 90VDC B supply.

The push pull output tube filaments were powered directly from the HF transformer of the PSU the regulation for both (HF) A and (DC) B+ was taken from monitoring the 1.2Volt HF A supply.

The bias for the output stage is generated via a CCS whose load  is the filament of the voltage amplifier (20mA ISTR ?), all very elegant: and sounds superb as well, with no vestigial HF on the output.

[tridac]

Have seen that before, but better stiil, use a 500r trimpot, whatever, then you can adjust to compensate to any imbalance...

[SteveThackery]

My thanks so everyone who has contributed. I've found it to be a very useful and enlightening discussion.

[TimFox]

A corollary question to DC or AC heater power choice:
Vacuum-tube heaters or filaments have strong temperature dependence:  the cold resistance is much less than the resistance at operating temperature.
When driven directly from a transformer secondary or "stiff" DC supply, there is a very high inrush current, that stresses the heater at power-on.
In operation, it is good practice to maintain the heater voltage to within +/-10% or better.
There are many ways of reducing this effect, for example:
1.  A resistor in series with the heater wastes power during operation, but reduces the initial current.  (AC or DC)
2.  A NTC thermistor in series with the heater or transformer primary will reduce the initial current, but wastes less power during operation after it heats up.  (AC or DC)
3.  If a normal voltage regulator circuit is used to stabilize the heater voltage at 6.3 VDC (or whatever), its current limit can reduce the initial current while maintaining the voltage at normal current.  (DC)
#3 is perhaps the most accurate method with simple components, assuming DC power.
When tubes were dominant, small-signal tubes were inexpensive and extending their lifespan was often ignored.
I remember 12AX7s, when in production, were < $1.00 at retail.  Have you priced an NOS Telefunken 12AX7 recently?  (Maybe $500 ea.)

By the way, an important limit is the voltage between heater and cathode (for indirectly heated cathodes), which is why the heater cannot be allowed to float, and important for cathode follower and phase-inverter operation, where the cathode is at a substantial voltage above ground. 
This varies from tube to tube:  some are good to +/- 200 V (peak), but many have much lower insulation voltage ratings. 
The rating (specified on the tube data sheet) is often asymmetric, or with a limit on the DC value (for heater positive with respect to cathode) less than the peak value.

[Bud]

I remember something along the lines that DC heating may cause uneven emission and uneven utilization of the cathode material and lead to reduced operational life of the tube, because of the uneven potential of the cathode along the heater, since the heater has one end grounded, at zero potential and gradient potential increasing toward the other end which is under a higher potential. To mitigate this, heater voltage polarity was flipped every time the device is turned on, or set it randomly.

Old farts here may confirm or provide a better explanation

Edit: to an extent this effect can be visually observed on VFD displays if the heater is not center biased. The display will have a noticable brightness gradient along the tube.

[David Hess]

No, I'm not sure why the Ef 7.5 line is for as it is significantly higher than 6.3V +10% to start with. Any H-K voltage above 100 - 200V (tube dependent) is outside the datasheet maximum specs anyway, so extending the graph to 300V is somewhat academic too.

That is what I mean.  Reliability should have dropped and been lower than the 6.9 volt line.  I suspect the graph was just mislabeled.

The relevant text from Getting The Most Out Of Vacuum Tubes - Tomer is:

There are many heater failures that appear to be simply
that and nothing more, but which in reality are the end
result of heater insulation breakdown. This is particularly
true where a relatively high heater-to-cathode voltage ex-
ists. Until a few years ago, all indirectly-heated cathode-
type tubes had a heater-to-cathode rating of approxi-
mately 90 volts. This was based on the known limitations
of the insulating material used on heaters. There has been
no change in this material over the years, yet heater-to-
cathode voltage ratings have steadily increased. This
doesn't make much sense—from an engineering point of
view—even though it does permit the design of many cir-
cuits not possible under the older, more conservative rat-
ings. As a result, we have far more breakdowns of this
type in the field today than we had some years ago.

As stated DC heating of preamplifier tubes is possibly the only real advantage for sound quality.

Tektronix used regulated DC heaters at least once in their 1A7 differential vertical amplifier which used pairs of 8393 medium-mu triode Nuvistors in parallel for 10 microvolt per division sensitivity.  The DC heaters kept AC ripple out of the most sensitive part of the signal path, and regulation controlled voltage drift.  The heater current was trimmed to balance the differential input stage.

[David Hess]

I remember something along the lines that DC heating may cause uneven emission and uneven utilization of the cathode material and lead to reduced operational life of the tube, because of the uneven potential of the cathode along the heater, since the heater has one end grounded, at zero potential and gradient potential increasing toward the other end which is under a higher potential. To mitigate this, heater voltage polarity was flipped every time the device is turned on, or set it randomly.

That applies to directly heated cathodes, where the heater itself is the cathode.  Indirectly heated cathodes which became the dominant technology have no such issue.

[TimFox]

Yes.  Another term for indirectly-heated cathode is “equipotential” cathode.

[radar_macgyver]

Some RF tubes like Klystrons have datasheet specifications for amplitude and phase pulling vs. heater voltage. In radar applications, this manifests as 60 Hz and harmonics in the received spectrum. Typical mitigations are running the filaments at the same frequency as the radar's pulse rate. This still produces spurious emissions, but they lie at the Nyquist frequency of the received spectrum and can be ignored. High power tubes have rather high filament drive requirements (7.5V, 33A) so it's rare to find DC driven filaments since the filament supply must pass through the high voltage pulse transformer used to power the cathode.

[Xena E]

Where low power tubes were once used in domestic portable radios a range was developed that allowed economical (for the time!), powering from dry cells.

The filament current that these devices needed was so relatively low that designers had to be mindful of the fact that the plate current added to the current through the filament and not only were there potential gradients this inequality of filament current could also cause limited life span, this was a particular problem in sets that had 5 tubes say, with nominal 1.5V heaters that ran in series across a 7.5 Volt supply.

If the schematics of the day are inspected those sets so designed often had bypass resistors wired across sections of filaments to counteract the effect.

[Gyro]

No, I'm not sure why the Ef 7.5 line is for as it is significantly higher than 6.3V +10% to start with. Any H-K voltage above 100 - 200V (tube dependent) is outside the datasheet maximum specs anyway, so extending the graph to 300V is somewhat academic too.

That is what I mean.  Reliability should have dropped and been lower than the 6.9 volt line.  I suspect the graph was just mislabeled.

The relevant text from Getting The Most Out Of Vacuum Tubes - Tomer is:

There are many heater failures that appear to be simply
that and nothing more, but which in reality are the end
result of heater insulation breakdown. This is particularly
true where a relatively high heater-to-cathode voltage ex-
ists. Until a few years ago, all indirectly-heated cathode-
type tubes had a heater-to-cathode rating of approxi-
mately 90 volts. This was based on the known limitations
of the insulating material used on heaters. There has been
no change in this material over the years, yet heater-to-
cathode voltage ratings have steadily increased. This
doesn't make much sense—from an engineering point of
view—even though it does permit the design of many cir-
cuits not possible under the older, more conservative rat-
ings. As a result, we have far more breakdowns of this
type in the field today than we had some years ago.

No, that graph makes no sense to me either, the text doesn't really help either - Yes, it follows that increasing heater-cathode ratings without change of materials would increase breakdown events, but why does it show the lowest failure rate (irrespective of incomprehensible heater voltages) at ~90V h-k rather than 0V. I think it's just drawn wrong!

[Gyro]

Indirectly heated tubes have a reasonable heater thermal mass that tends to reduce thermal shock, the low cold resistance of several parallel heaters tends to load down the transformer secondary at switch-on too at least for the 6.3V ones, rather than the series chain connection 300mA ones. The exception were the 'quick heat' tubes that would flash alarmingly at the termination end.

Most alarming are the directly heated Thorated Tungsten filament tubes, where the filament undergoes a physical transition at ~900k due to the Miller-Larson effect* discovered in 1944. This causes the crystal structure to re-orient as it passes through the transition temperature, leading to necking, cracking and ultimate failure. Although the filaments are rated for around 30k hours, each cycle reduces life by around 0.2%! The only way of mitigating this is to keep the temperature above 900k [Edit: and limiting inrush current to reduce the stress on the increasing number of defects from when you do cycle them]. Not a happy situation for owners of big, expensive, and potentially rare DHTs.


* Morgan Jones, Valve Amplifiers, third edition, pp 261.

[floobydust]

Diodes (selenium) and large value electrolytic (filter) capacitors were expensive back in the day.
A critical application needing DC filament power is in condenser microphones.

For DC filaments, designs pulled it from B+ through a dropping resistor, making a lot of heat. It also contributed to cathode current, so the heater and cathode were pretty much at the same potential, and cathode resistor could be small. e.g. Altec, Telefunken or Neumann U47 (VF14 60V 50mA filament, undervolted) and many other mic's used the same scheme for low noise.

[David Hess]

No, that graph makes no sense to me either, the text doesn't really help either - Yes, it follows that increasing heater-cathode ratings without change of materials would increase breakdown events, but why does it show the lowest failure rate (irrespective of incomprehensible heater voltages) at ~90V h-k rather than 0V. I think it's just drawn wrong!

I trust that part of the graph.  Biasing the heater positive relative to the cathode prevents emissions from the heater to the cathode, which would otherwise damage the insulation.  The graph is based on empirical results.

[CaptDon]

I have no explaination for the following, but repeated experiments showed the same results. 12VAC RMS VS. 12VDC Observation, the lamps operated on D.C. had a much shorter life. The lamps were lit for an 8 hour shift each weekday (actually 8.5 hours daily). The A.C. operated lamps showed marginal darkening of the glass and a broken filament which seemed to retain its original thickness at end of life. The D.C. operated lamps showed intense darkening of the glass and the reduction of filament thickness was visible to the naked eye. This experiment was undertaken do to customer complaints of short bulb life. It was amazing how fast a 12VDC regulated supply could kill a 14.4 volt bulb!! But all of this is confusing because we all know of cars operating at a somewhat variable 12 to 13.8 volts D.C. in cars 20+ years old and the dash lamps are still working fine!!! Maybe fixed vs. variable voltage is some sort of clue? Controlled testing directly opposing reality?? Bulbs in question #53, very common in older cars and motorcycle instrument clusters, and in this case arcade game back lighting.

[T3sl4co1l]

Indirectly heated tubes have a reasonable heater thermal mass that tends to reduce thermal shock, the low cold resistance of several parallel heaters tends to load down the transformer secondary at switch-on too at least for the 6.3V ones, rather than the series chain connection 300mA ones. The exception were the 'quick heat' tubes that would flash alarmingly at the termination end.

Particularly Mullard types were/are famous for this; to be clear, it's a normal / intended effect, for parallel operation.  Such tubes should not be used in series chains, IIRC.

Most alarming are the directly heated Thorated Tungsten filament tubes, where the filament undergoes a physical transition at ~900k due to the Miller-Larson effect* discovered in 1944. This causes the crystal structure to re-orient as it passes through the transition temperature, leading to necking, cracking and ultimate failure. Although the filaments are rated for around 30k hours, each cycle reduces life by around 0.2%! The only way of mitigating this is to keep the temperature above 900k [Edit: and limiting inrush current to reduce the stress on the increasing number of defects from when you do cycle them]. Not a happy situation for owners of big, expensive, and potentially rare DHTs.


* Morgan Jones, Valve Amplifiers, third edition, pp 261.

Hmm, I'm not aware of any phase change in the W-Th system.  Although I don't find a phase diagram offhand, but, it's present in small amounts so is probably either an insoluble impurity or solid solution, and either Th (or ThO2) itself.  Pure Th has a phase change (FCC to BCC at 1360°C). Pure W has no phase change, it's the same crystal up to MP.  Jones seems to misrepresent this here -- it appears Miller and Larson published in the 50s, regarding creep, not phase change: https://en.wikipedia.org/wiki/Larson%E2%80%93Miller_relation  I don't see anything for either creep or phase transformation for W-Th mixtures, but https://inis.iaea.org/collection/NCLCollectionStore/_Public/32/053/32053126.pdf?r=1&r=1 is interesting for creep, in the same temperature range, of a couple other alloys.  Creep and recrystallization proceed normally at higher temperatures it seems, which is certainly evident in tungsten lamps (especially the blocky recrystallization and sublimation in a failed halogen bulb), but creep should be modestly relevant for thoriated cathodes.  (Literal "filament" cathodes, wires and ribbons, are strung up between spring supports, so stress and creep are relevant; very large or specialized transmitters may have self-supporting cathodes, with creep nearly irrelevant?)

I would be inclined to think thermal cycling causes differential stresses, and perhaps activation of grain boundaries/migration (perhaps under electromigration forces too? maybe not with AC though).  Maybe that spreads into a stress riser that eventually snaps the filament, maybe it's governed by creep (might the creep timer be reset every time it's cooled down?), no idea.

This probably all isn't too germane for Jones' purposes, more just to say something on the topic, than to give the academic rigor that vacuum tube designers and manufacturers would need, so I suppose it doesn't matter all that much.

Tim

[coppercone2]

tungsten does all sort of weird stuff

1) if you snap it, it can splinter like wood
2) if you heat it too much, the tip of a tungsten welding electrode turns into basically,, if you took a tree trunk, cut it into a cone, and left it in the sun and rain for 15 years. It starts turning into a 'blooming onion'.

I am convinced its actually a type of lumber


metals not supposed to do this
https://www.reddit.com/r/Welding/comments/an5atq/my_friends_tungsten_split_after_he_broke_his_cup/


Its higher temperature effects, but I am convinced there tungsten is a tricky material




I think tungsten actually comes from trees. I don't think they mine it, its probobly extracted from the center of a specific pine cone.

[floobydust]

"The Influence of Notching Upon the Life of Miniature Lamp Filaments"
Notching is the phenomena in which a saw-toothed surface appears over portions of the filament. The notching grows due to the electro-migration of tungsten ions and becomes more noticeable after long operation. Notching depends on the lighting condition, either DC or AC. At AC voltage, the notching occurs near the area where the filament is supported by the anchor or is connected with the lead-in wire. These areas have a temperature gradient. At DC voltage, since tungsten ions move in only one direction, the notching occurs over the entire filament. Therefore, the lamp life at DC voltage becomes shorter than at AC voltage because of more severe notching. In order to reduce the notching, a rhenium tungsten filament, which has a higher temperature recrystallization point, can be used." source

edit: tungsten seems to be crystalline in structure, as well as the fact it oxidizes too. SEM pictures are pretty interesting.
Blown filament (vacuum loss they say) https://www.sciencephoto.com/media/601254/view/blown-tungsten-filament-sem

[Andy Chee]

Therefore, the lamp life at DC voltage becomes shorter than at AC voltage because of more severe notching.

What is the lowest AC frequency at which severe notching becomes pronounced?

For example, what if one were to reverse DC polarity of the filaments at every device switch-on power up?  Does that count as AC?

[Xena E]

Indirectly heated tubes have a reasonable heater thermal mass that tends to reduce thermal shock, the low cold resistance of several parallel heaters tends to load down the transformer secondary at switch-on too at least for the 6.3V ones, rather than the series chain connection 300mA ones. The exception were the 'quick heat' tubes that would flash alarmingly at the termination end.

Particularly Mullard types were/are famous for this; to be clear, it's a normal / intended effect, for parallel operation.  Such tubes should not be used in series chains, IIRC.

Yes, heater flash was one of their traits, particularly the 12AT7, 12AU7, and 12AX7 series. The phenomenon of heater flash is now one of audiophooleries urban legends that signifies better quality.

One spectacular 'flasher' is the 6AL5/EB91. These were designed for use in 300mA controlled warm up chains, however,  had so little thermal mass would light up a room at switch on.

They were not particularly known for open heater failures but subject to a common fault of h/k insulation breakdown.

In bright emitter electron tubes, embritlement and wasting of the filamentary type cathode was always a factor in life determination. When the minature low power battery tubes were introduced it was difficult to produce the filament with any consistency in tungsten, the answer that some manufacturers adopted was the use of nickel.

The problem with nickel filaments being that it was more sensitive to supply voltage variations and it's narrower temperature/resistance range meant that overvolting damage occurred more readily.