Drop DC heater voltage by 1 volt, what resistor

[Chuck H]

Yupper! unfortunately a lot of the original 115V only iron is still around and as of early last year that's what I was getting whenever I ordered a PT. It would be good to know one way or the other when ordering but it's not always clear.

[R.G.]

...perhaps this is a good opportunity to ask R.G. what he thinks of the practice of putting an otherwise unused 5V rectifier heater tap in series with the primary in order to effectively add a few extra turns to a PT primary winding and thereby effectively alter a 115V transformer to one optimized for 120VA.

It's a great idea - almost.

The "almost" has to do with two things: the intangibles of how safe it is, and the current rating of the 5V winding. Primary currents on tube amps are in the 3-5A range for most of them. I haven't checked the filament currents of typical tube rectifiers. What's the heater current for a 5AR4, et al?

The safety issue is harder to figure. The 5V winding is not a tap, it's a fully floating winding, intended to float on the rectifiers up at B+. So it has to be isolated for several hundred volts to be safe to use. But primary - secondary isolation used to be 1500V, then 2000, then 2500, and it's now about 4000 if you follow the safety specs. By attaching the 5V winding to the primary, you're making the isolation between it and the windings around it have to be primary/secondary safety barriers retroactively. This may or may not be a good bet, but it's sure that the transformer designer and safety testers did not design and test it that way.

It may be fine, but for the cost of one small transformer in a box ahead of the whole amplifier ( or wired in ahead of the power trannie integrally) you can sidestep it entirely. A 10Vct 3A power transformer is pretty small and cheap and gives you the option to correct 5V and 10V. A 12Vac CT trannie lets you correct 6 and 12V and is probably cheaper because there are more of them.

And so I'm cautious. I'm probably too cautious about primary power safety, but I was beaten over the head with it for some years.

[Chuck H]

And so I'm cautious. I'm probably too cautious about primary power safety, but I was beaten over the head with it for some years.

When I was younger I took a job jerkin' hardware. This store chain required a training seminar lasting four days!!! One lesson the instructor opened with "Ok, what's the worst thing you can do to a customer?" Hands went up and answers such as "Fail to give accurate advice" and "Treat them like you don't have time for them" were offered. And the instructor said "No... The worst thing you can do to a customer is kill them." "Now begins the safety portion of this seminar."

[Gingertube]

The way a cathode works is that emission increases exponentially with temperature towards a limit, such that at absolute zero (zero Kelvin), no electrons are emitted; at room temperature, a few escape. At glowing temperatures, there are enough electrons being "boiled off" to make use of. Due to the exponential curve, you gain very little emission by increasing temperature past where it really gets going.

That means you get no performance benefit from running heaters at higher than specified voltage, only lifetime reduction.

I still can't get a sensible consistent answer about running tubes at lower than recommended heater voltages. Mant say tube life is reduced, many other say the opposite.
I would suggest that you get longer life, lower transconductance and perveance (that is lower current capacity). Many highend HiFi Preamps run heaters at lower than recommended voltage. This means that teh tubes do not meet full spec when new but maintain that "not quite new spec" way longer than the expected lifetime when operated at full heater voltage

Broskie had a few things to say about heater voltage here:
The Tube CAD Journal, Heater Concerns
If you keep clicking "NEXT" to about the 9th page there is a graph of tube transconductance vs lifetime at 5.04V, 6.30V and 7.56V heater supply. The 6.3 and 7.56 volt curves show the transconductance falling off by 25% fairy quickly over the first 1000 hrs of operation with the 7.56 curve falling off at a higher rate after that. The 5.04 volt curve starts at that 25% low point and basically stays at that point way out to over 5000 hours. Passed the 1000 hour point the 7.56V and 6.3V curves have actually fallen below the 5V curve.
The old tube guys recommended changing a tube when transconductance dropped below 70% of new value but that was because circuits were designed to operate with tubes down to 70% of nominal value - so 70% of new transconductance was considerd to be the end of life point.
If you design the circuits to work with a lower transconductance to start with then it seems you can get VERY extended tube life by running at lower voltage.

Some guys say that you get reduced tube life at low heater voltage and talk about things like cathode poisoning. I think this just applies to big transmitter tubes but possibly to audio output tubes too.

I would therefore not hesitate to run preamp tubes at lower voltage but would perhaps try to keep output tubes near nominal voltage.

Cheers,
Ian

[R.G.]

That is the essence of the GE report I'm still opening boxes to find. Down around 5V, the emission and gain are lower, but stable at that lower amount for longer periods of time. So you get longer life, somewhat reduced gain, but more stable. And for power tubes, run them at rated voltage because they need all the electrons.

And thanks for the link. I have a hardcopy of the reference Broskie used, Shea's "Amplifier Handbook". I'll add that to the search list. It may well have more info to add to the issue.

One kitibz: increase in emissions is a power law, not an exponential, although this is truly an inconsequential difference, as the shapes of the curves are almost indistinguishable. At least if I dimly remember that bit of math right.

[Tom Phillips]

...If you design the circuits to work with a lower transconductance to start with then it seems you can get VERY extended tube life by running at lower voltage...

Very good point. An amp design should be done using a whole system approach. Since each design change will often affect multiple performance parameters in addition to its main purpose one can't necessarily just make mods/changes without addressing the associated secondary effects. (Unless you are rewarded with a magic synergistic result that you love) In other words, set your chosen heater voltage before you voice your design. It appears that there is opportunity to extend tube life with a slightly lower voltage. Certainly, higher heater voltage seems to have at least the one distinct disadvantage of faster changing transconductance as the tube ages. I think this would be most important when the user's goal is to get the most use out of NOS tubes from the golden age. However, for current new production tubes, there remains the problem of other things that go wrong with them such as poor mechanical construction etc. That may swamp out any advantages of messing with the heater voltage if it is already within reasonable limits. Then there's the fact that some new production tubes have barely passing transconductance (by original standards) when they are brand new.

There seems to be no end to the tradeoffs. The successful amp builder is the one that can produce am amp that the player likes. When cork sniffers start asking about the type of parts that were used I like to just tell them to close their eyes and listen to the amp. When they persist with their questions such as "what type of resistors did you use" I just keep repeating "Play and Listen." If the user can't get passed that then they (and you) will be doomed to change parts forever as they search for the magic combination that will make them sound like the guitar hero they are trying to emulate. All that effort often uses up valuable time that would be better spent practicing their guitar playing.

[NateS]

A side-step is to go to a duty-cycled, balanced, variable resistance. The way I'd approach this is to put a fixed resistor in series with both ends of the heater string that gets the heaters below the voltage you want them, and then to use a MOSFET pair on both sides to "short out" the extra resistor in a symmetrical PWM way. The difference in voltage between MOSFETs on and MOSFETs off will be a volt or so, and can be done at a frequency which doesn't interfere with the audio.

Of course, it may be simpler to use a DC power supply. Certainly a simpler design, if more intrusive on the "no-modernness" and "retroness" of the amp.

I was looking at this problem too. My biggest problem with going to DC heaters just for the purpose of regulating them is throwing the ginormous capacitor required for 6.3VDC @ 4A from a 6.3VAC@60Hz winding. Even Switching PS frequency doesn't really help you here, the only way to reduce it is to reduce current by taking power from a higher voltage winding. It doesn't seem to matter what I do, at some point you're going to have at least one BHC in the circuit. They're not horrible on price, but they're big, it's another place inrush needs to be controlled. And keep in mind, the traditional alternative is just two wires (ie 6.3v AC heaters with no added circuitry).

I was trying to combine 1) inrush control 2) regulation, 3) DC elevation, 4) heaters all the way off, all in the same block. I think your duty cycle controlled resistance is not a bad way to regulate. I do wonder though, since heaters are essentially resistive (at least they are at mains frequencies), if you're DC modulating a series resistance, isn't the "mosfet shorted" instantaneous current going to be the same as unlimited? Does that matter, or is it average power?

[R.G.]

I was looking at this problem too. My biggest problem with going to DC heaters just for the purpose of regulating them is throwing the ginormous capacitor required for 6.3VDC @ 4A from a 6.3VAC@60Hz winding. Even Switching PS frequency doesn't really help you here, the only way to reduce it is to reduce current by taking power from a higher voltage winding. It doesn't seem to matter what I do, at some point you're going to have at least one BHC in the circuit. They're not horrible on price, but they're big, it's another place inrush needs to be controlled. And keep in mind, the traditional alternative is just two wires (ie 6.3v AC heaters with no added circuitry).

I was not being clear. I wanted to PWM the *AC* with a dual-MOSFET on each side, symmetrically.

A heater winding running from 7Vac is over-volted. A heater winding running from 5Vac is under-volted. If you switch between 7V and 5V on a PWM basis, you can get to anywhere in between. The idea was to put a resistor between the too-high 7Vac and the heaters on both sides symmetrically such that the voltage on heaters with the resistors in circuit was about 5Vac. Then use some hypothetical perfect AC switch to "short" the resistors on each side simultaneously. If the shorting switches are in sync, you do a PWM between 5Vac and 7Vac symmetrically, and have very fine control of the voltage indeed. The switches have to be (a) much faster than 60Hz, which is easy and (b) much lower resistance than maybe 5 ohms; there are many MOSFETs in the 10-50mOhm region that might work, especially if you only run them at 600Hz or 6kHz.

There is an implied c requirement that the PWM frequency, radiation, etc. not disrupt the audio. Given that it's only a few volts being switched, this may be somewhat easily managed. Or maybe not. Mother Nature is, after all, a mother.

I was trying to combine 1) inrush control 2) regulation, 3) DC elevation, 4) heaters all the way off, all in the same block. I think your duty cycle controlled resistance is not a bad way to regulate. I do wonder though, since heaters are essentially resistive (at least they are at mains frequencies), if you're DC modulating a series resistance, isn't the "mosfet shorted" instantaneous current going to be the same as unlimited? Does that matter, or is it average power?

I hadn't got that far yet. I was solving a smaller problem. In the setup I was hypothesizing, it was on DC instantaneously, but AC on a macro scale. That just makes the MOSFET switching more complicated, doesn't affect much on the heaters. And the "MOSFET shorted" condition would always have the heaters themselves in series, so no, no unlimited currents.

By the way, I finally read a note saying that tungsten filaments have about a 7-8 to one increase in resistance from room temp to heater temp, so inrush is about 7-8X operational current. Worth deploying inrush limiters.

The hyper-deluxe solution would be to use a current-limited optical servo to literally watch the heater's glow-color and servo them to a specific color temp, which implies a specific temp!

[NateS]

I was not being clear. I wanted to PWM the *AC* with a dual-MOSFET on each side, symmetrically.

I think I understood you, and you got me thinking about doing it that way. By "unlimited" I meant "not limited", not "infinite".

Cold resistance is pretty easy to measure. Grab a multimeter and tube. Doing it this way, I estimated about 15A peak for 3-12AX7's, and a pair of EL-34's, which is quite a bit less than what I was expecting to deal with, but worthy of limiting none the less, in the name of pampering our glass bottles before bludgeoning the snot out of them. (As the albino in the movie, "The Princess Bride" says, "The Prince and the Count always insists on everyone being healthy before they're broken).

The only way I came up with to regulate the heater winding without the DeLorean sized flux capacitor, is sort of Rube Goldbergesque. (PFC Buck converter running on the AC input, with a trimpot scaled version of the AC signal for reference voltage, and only enough filter to handle the switching frequency, with no bulk storage required. It sort of resembles a cross between a class D and class H amp.) I like your solution better.

Since I'm running elevated heaters, I can skip the balancing act. If you're switching that fast, and the load is purely resistive for practical purposes, why use a resistor at all?

[Mick Bailey]

I have heard extra noise from excess filament voltage. This is in critical applications like tube mic's. With adjustable heater voltage in the mic supply you could actually hear the noise increase with higher filament voltage.
Not sure if you would hear it in a guitar amp but it's worth noting.

I have a 1950-something article where Osram recommended reducing heater voltage to reduce noise.

[Chuck H]

Makes sense. If you have excessive emission there's going to be more un allocated stray electrons reaching the plate. So how do we know how much heater voltage is ideal for any given circuit? Maybe the next big thing to reduce hiss and noise in uber-gainers!!! Idealizing the filament for each stage.

[R.G.]

Re: tinkering each filament voltage for best performanc

Why stop there? I think I may open a "Certified Electrons" shop.

Imagine this: take ordinary electrons and subject them to a rigorous filtering and cleaning operation where they are sent through hot "annealing" media, then plunged abruptly through cooled media, and then subjected to an electron-ballistic filtering path in a hard vacuum that rejected all electrons that do not follow a carefully chosen path to a vacuum-electron collector and then inserted into electron transport medium.

It's worth noting that only electrons with the correct quantum spin orientation will make it through this process, and any that do not will be forcefully ejected by the filtering.

The transport medium will be made from selected and treated electron traps. Although commercial medium is used, it is first resolved to its lowest state, then normalized and finally filled with certified electrons for delivery. There is a shelf life, and any certified electrons must be used within three weeks (well, OK, 22.3 days) to get the best benefit.

Just as in a crystalization process, the selected and phased nature of the certified electrons form nucleation sites. Once released into the equipment to be treated, they certified electrons begin aligning the electrons already there. If done carefully, the in-situ electrons take spin orientation from the certified electrons and produce almost as good an effect as if all of the electrons in the equipment were certified electrons. Over time, the alignment decays, but it can be renewed at any time with more certified electrons. Regular subscribers will receive a substantial discount on an electron maintenance agreement.

Notice that to be of the best use, certified electrons must be used only in AC powered equipment. DC powered equipment runs electrons through in one direction, and the benefit is soon lost as the certified electron charge is swept out and into the batteries. AC powered equipment is different. The electricity runs into the equipment for 1/120 second, the reverses and runs the other way, leaving the charge of certified electrons largely in place, not wasted by siphoning away.

[Here's the factory: A battery charger feeding two wires; one is looped into a glass of ice water, then through a glass of hot water, and from there through a vacuum tube from cathode to plate; from the plate into a discharged Nicad battery. The battery is charged, shrink wraped, and ready for shipment. *All* electrons have the same quantum spin states, so any not having the right spin are indeed rejected - if we ever find any. ]

[Chuck H]

Why stop there? Why not operate selected tubes (in a controlled environment) from the certified electron battery and then market them as having been "impregnated with certified electrons"!!!

[R.G.]

I just gotta believe that this would be illegal in some states.

[Enzo]

I think the "Tice clock" anticipated this years ago.

What's All This Hoax Stuff, Anyhow? | Lighting content from Electronic Design