Vacuum tubes - DC heaters?

[SteveThackery]

I've been checking out a 100 watt tube (valve) amplifier over on the official KICAD forum, and a question occurred to me.

Traditionally the cathode heaters are fed from an AC source - almost invariably a winding on the mains transformer. Voltages are typically 6.3VAC RMS, or sometimes multiples thereof on transformerless circuits where the heaters are connected in series across the mains.

Anyway, do you think there is any mileage in running the heaters off a DC supply? Nowadays smoothed and regulated DC is straightforward to deliver compared with the early days of valve gear. I wonder if AC-fed heaters can inject hum into the signal path - something that might be avoided with DC heaters.

[Gyro]

No, or very little, mileage. The little bit comes from very low level signal stages, where you may get some hum reduction. Even then, low level tubes like the EF86 have non-inductive biflar wound heaters to reduce coupling. Even standard tubes have a folded arrangement that cancels hum to a large extent.

It is vaguely possible for the hot (but insulated) heater to behave as a cathode, but this possibility is eliminated by elevating the whole heater winding to a few 10s of volts above the highest cathode potential.

Methods of producing a DC heater supply, eg. a bridge rectifier or SMPS are far more likely to introduce noise in the audio pass-band.

[TimFox]

The main advantage of DC heaters (assuming clean DC) is easier wiring of the leads to the tube socket on low-level stages (e.g. RIAA preamp).
Note that RIAA stages have relatively high gain at 50/60 Hz.
AC requires careful wiring there, which is covered in tube books.

[schmitt trigger]

Of all the “classic” amplifier schematics that I recall seeing, only the phono preamp or mic preamp stages used DC powered heaters. Not all of time though, there were other tricks to minimize hum.

What I also recall seeing is half wave rectification applied to all the heaters while the amp was in standby, for power saving and a faster turn on.

[Gyro]

Yes, heater wiring always needs to be very carefully implemented, usually using tightly twisted solid core wire. Heater currents don't really have a place on PCBs.

Of all the “classic” amplifier schematics that I recall seeing, only the phono preamp or mic preamp stages used DC powered heaters. Not all of time though, there were other tricks to minimize hum.

When you get down to those sort of levels, microphonics can be as much problem as any heater induced hum. Mullard recommended resilient mounting of such stages, again typically with the EF86 pentode.

Resilient mounting is / was also used for things like the Nuvistor board on the HP3400A.

What I also recall seeing is half wave rectification applied to all the heaters while the amp was in standby, for power saving and a faster turn on.

Ah yes, the old 'instant on' modes on guitar amps and TVs of the era. Sometimes the reduced cathode temperature was enough to prevent tube life reduction due to cathode poisoning (no anode supply / cathode current). That was a real problem in TVs when the standby modes were first introduced.

[Benta]

Don't forget electromigration on the filaments when using DC.

[TimFox]

I have heard that DC shortens heater life in indirectly-heated cathodes in small-signal “receiving tubes”, but have been unable to locate a definitive discussion in my extensive library of quaint and curious volumes of forgotten lore.

[TimNJ]

Yes, heater wiring always needs to be very carefully implemented, usually using tightly twisted solid core wire. Heater currents don't really have a place on PCBs.

Of all the “classic” amplifier schematics that I recall seeing, only the phono preamp or mic preamp stages used DC powered heaters. Not all of time though, there were other tricks to minimize hum.

When you get down to those sort of levels, microphonics can be as much problem as any heater induced hum. Mullard recommended resilient mounting of such stages, again typically with the EF86 pentode.

Resilient mounting is / was also used for things like the Nuvistor board on the HP3400A.

What I also recall seeing is half wave rectification applied to all the heaters while the amp was in standby, for power saving and a faster turn on.

Ah yes, the old 'instant on' modes on guitar amps and TVs of the era. Sometimes the reduced cathode temperature was enough to prevent tube life reduction due to cathode poisoning (no anode supply / cathode current). That was a real problem in TVs when the standby modes were first introduced.

Heater current for a typical power tube like a 6L6 is only like ~1A. It's not *that* much. A 2mm wide trace should not give much temperature rise for a few tubes.

I designed a PCB based on the Fender 5F6A design about 7 years ago. I routed the AC heaters as a parallel pair of traces. Anecdotal, but it is a quiet amp.

[tridac]

At least one early HP tube based  spectrum analyser (853 ?) used a solid state regulated dc supply to drive the 2-4GHz bwo heater, the 1st mixer local oscillator. I think that might have been done for stability reasons, or perhaps to eliminate modulation of the lo via the heater.  Only know that because the bwo starts to have dead spots in the output, as the tube ages, and the one I got for scrap price with that issue, was fixed by turning up the heater volts by around 0.1v, restoring the bwo to full output across the band. A bit like the old tv crt restoration trick

Other than that, can't think of any other reason, but perhaps very low level preamplfiers, tube dc op amps, again for stability or noise. Can't see it causing any problem though...

[TimFox]

Local oscillators in HF receivers often had stabilized heater current for thermal stability.  The AC current was passed through a “ballast” tube, such as the 4H4C used in the National HRO-60, to compensate the current against voltage difference.  These were PTC resistances in a tube envelope.
With modern semiconductors, an appropriate regulator IC is easier with DC operation.

[retiredfeline]

I've only ever seen it for the input tubes in some fussy audio preamp designs. I don't see any problem powering heaters with a DC supply. Batteries were use for the A supply before mains.

[floobydust]

Capacitive coupling from heater to cathode does add hum if the cathode resistor is unbypassed, which is the case in some preamp circuits and guitar amp front ends etc.
If there is a lone cathode resistor unbypassed then you get hum from AC powered filaments and running them on DC has benefit. Also the wiring radiates a lot of hum, need a twisted pair for AC filaments this has been known for 100 years.
For a power amplifier in push-pull configuration, hum cancels in the output transformer so there is no benefit to using DC filaments there.

I'll use Schottky rectifier diodes and an LDO to generate DC for filaments.

[themadhippy]

slightly related,is there any  advantage to the following heater arrangement? nothing  exotic ,just ecc82 ,ecc83 and el34's

[TimFox]

A useful circuit if the transformer secondary doesn’t have a center-tap.

[themadhippy]

But theres no need for a center tap if its only powering the heaters,which is what this is doing

[TimFox]

You need to establish a (common mode) DC voltage on the heater, and not float it.
If you ground one side of the heater, you have a common-mode AC voltage on the heater which can couple capacitively to the grid or cathode.

[MathWizard]

I had to make a supply for a vacuum florescent display this winter. And I made some sort of a full bridge oscillator. The rails were between -20, GND, and the DC level of the low side of the load was down around -18V, and the resultant AC was only a few volts. Something like that. But it was fairly high current, up over 100mA if I wanted.

If I had to make a VT heater supply, I'd try something like that again, but without any DC bias, or very little.

[themadhippy]

You need to establish a (common mode) DC voltage on the heater, and not float it.

That probably explains why my experiments with valves never turned out as i expected

[T3sl4co1l]

Rectifying AC doesn't make much sense: the advantage is small while the cost is relatively large (and, not that rectifiers and capacitors are exactly expensive themselves, but like I said, the advantage is just that small).

The advantage is larger for low-level stages, guitar pickup, phono, etc., where the ripple makes a bigger difference.  The cost of a simple rectifier-filter is more acceptable, too (i.e., rectifier, cap-input filter, RC filter to heater -- R necessary to adjust voltage down (can get some extra filtering out of it too), wasting some power).

DC bias also helps, to reverse-bias the heater-cathode diode.  The heater still has some emission, and this helps reduce current flow in general -- thus reducing ripple input as well.

You could go fancier and use an SMPS, but 6.3VDC is somewhat uncommon, and if you're going to go the trouble, why not make the 300V or whatever with it too?  But now you're doing an entire power supply design for very little amplifier, and EMI is a real problem.

Tim

[Andy Chee]

You could go fancier and use an SMPS, but 6.3VDC is somewhat uncommon

5VDC is very common these days.  I wouldn't imagine it's too difficult to build or modify one for 6.3VDC operation?

[David Hess]

DC bias also helps, to reverse-bias the heater-cathode diode.  The heater still has some emission, and this helps reduce current flow in general -- thus reducing ripple input as well.

I am not sure what the Ef 7.5 line is, but heater to cathode voltage also affects reliability.

[T3sl4co1l]

E_F should be filament voltage, though I don't get why it would go up >6.3V (makes sense) but then down again at 7.5V? Otherwise, that's about what I understand though, staying on the positive and highish side of H-K ratings.  I wonder what exactly the failure mechanism is being plotted... is that just regarding H-K failure (short, or H-H open)?

You could go fancier and use an SMPS, but 6.3VDC is somewhat uncommon

5VDC is very common these days.  I wouldn't imagine it's too difficult to build or modify one for 6.3VDC operation?

Depends.  Flyback almost certainly; forward maybe; LLC probably not.

Tim

[Gyro]

You could go fancier and use an SMPS, but 6.3VDC is somewhat uncommon

5VDC is very common these days.  I wouldn't imagine it's too difficult to build or modify one for 6.3VDC operation?

Compact 5V SMPSs these days use (coincidentally) 6.3V rated capacitors, so capacitor replacement would probably need to be included a modification.

[Gyro]

DC bias also helps, to reverse-bias the heater-cathode diode.  The heater still has some emission, and this helps reduce current flow in general -- thus reducing ripple input as well.

I am not sure what the Ef 7.5 line is, but heater to cathode voltage also affects reliability.

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.

The heater - cathode insulation limits are rarely an important consideration though. The only time they can be at risk in thing like high voltage SRPP stages or maybe cathode followers, where it might be advisable to raise the heater DC bias to some mid point (with due consideration of other tubes on the same heater supply).

I've never found the need to implement DC heater supplies on any of my designs, even low level. Raising the DC bias of the heater winding center tap (real or artificial) to 20-30V above the highest cathode voltage has pretty much guaranteed silent operation. Microphonics a stray magnetic fields from transformers are far more likely to be a problem (and sometimes confused with heater coupling).

On RF stuff, local decoupling the heater supply to the relevant chassis ground with low inductance capacitors [and series ferrite beads] can be important to stability, but this is due to capacitive H-K coupling at high frequency rather than 50Hz. [EDIT: This can also be relevant to high frequency tubes, like the E180F and 6S45P, being pressed into audio amplifier service, but these often also need grid, cathode, and even anode stopper resistors to tame them too].

[Xena E]

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

There is another more satisfactory method that I've seen that is a revival of the methods used in soundtrack lamp exciters in vintage cinema equipment, that is to use HF heating, it can even be used with directly heated tubes, the disadvantages of DC heating are thereby negated.