[AudioCats]

Do I have to filter the supply juice for the tube filament? will straight AC/ unfiltered DC work or will it generate hummm?

 

[dsavitsk]

Quote:

Originally Posted by AudioCats /img/forum/go_quote.gif Do I have to filter the supply juice for the tube filament? will straight AC/ unfiltered DC work or will it generate hummm?

AC almost always is fine, assuming you lay it out correctly. The only real exception is on a very high gain tube in something like a phono stage.

 

[AudioCats]

Thanks dsa, the "real" reason I asked was that I wonder if I can just use a car switching supply to lower the 30VDC down to 6V for ECC86's filament.... using a linear regulator to lower the voltage will require too big a heat sink since the current is pretty high.

Looks like as long as I can shield the switching noise from the rest of the circuitry it might be fine? (plan to use 7824 to drop the 30V down to 24 for the tube, the 7824 will also give some extra ripple rejection)

 

[dsavitsk]

Why not a big resistor? Something like a 40R/12W wirewound should drop enough of the voltage that a linear reg is fine. But, the switcher is probably fine too. Maybe put a ferrite on it to get rid of the HF noise.

Oh, a neat trick: use a lm7806, but connect the ground pin to ground through a diode that drops ~0.3V. You'll then get the 6.3V on the heater, which (you are using this as a cathode follower, right?) increases the gm slightly (over using 6V) which might be useful.

Also, the 6gm8 jis ust as happy with 30V on the plate as 24 -- maybe happier. Again, it will increase gm and lower rp.

 

[AudioCats]

Thanks, but a resistor will generate the same amount of heat, that makes me equally nervous....

no, I don't dare to use the 6gm8 as followers, thinking about using one as coloration/gain stage..... Don't think the 6gm8 can pump enough juice to justify using. Otherwise I'd use 4x 6gm8 to get enough current output to 600 ohm phones, and somehow 4x6V=24V, no need to drop voltage for filament

The main problem is the 6gm8 only has 2.4mA/V, which means it will take a lot of input voltage swing to generate a little AC current output? is there a web site where I can figure out how to do the calculation for the cathod followers? I keep wanting to use a CCD (replacing the cathod resistor) to choke the AC so all AC gets forced thought the output cap to the load, have no idea if that will work.......

 

[ericj]

Truth be told, tungsten filaments greatly prefer to be heated with AC. Heating them with DC weakens them.

Yes, there are tubes that are rated for DC heat, and there are tubes with oxide-coated equipotential heaters, but in general, the longest lifespan will be achieved with AC heat.

fwiw, how much it affects the lifespan varies from tube to tube, and is a bigger deal with hotter tubes. Most of the tubes we use have relatively low heater currents.

But if you were to take several incandescent lightbulbs, for example, and run half of them from 110vac and the other half from 110vdc, the DC-powered light bulbs would die much, much faster.

This was, in fact, the reason why so many of those "grid of lightbulbs" animated signs in the 80's always had a dozen dead bulbs at any given time. A good friend of mine was once tasked with building such a sign, and went to great lengths to solve the issue. He ended up using pairs of triacs to feed the bulbs with a modified sinewave, and avoided offending the FCC by using a random number generator to switch them on and off in a completely random order. (when he tried switching them in unison, it only took the FCC a few hours to find him and start pounding on his door. Really. A few hundred triacs in unison = noisy.)

The sad part is, they only built one sign like that, and the next engineer to fiddle with the design went back to DC-powering the bulbs. But my friend tells me he learned a lot, and has entire books on filament physics now.

 

[AudioCats]

so AC is actually better...... interesting!
maybe I should use a small trans to step the 24V ac down to around 6v, and use two 8.1v/3W zener to cut off the top of the waveform to make sure it's within 8.6V p/p .........that way the filaments might last forever

 

[tyre]

Quote:

Originally Posted by ericj But if you were to take several incandescent lightbulbs, for example, and run half of them from 110vac and the other half from 110vdc, the DC-powered light bulbs would die much, much faster.

I'm not sure that's true for incandescent light bulbs. I don't know enough about tube filaments to comment.

excerpt taken from here...
Quote:

As tungsten atoms evaporate from the filament, a very small percentage of them are ionized by the small amounts of short-wave ultraviolet light being radiated by the filament, the electric field around the filament, or by free electrons that escape from the filament by thermionic emission. These tungsten ions are positively charged, and tend to leave the positive end of the filament and are attracted to the negative end of the filament. The result is that light bulbs operated on DC have this specific mechanism that would cause uneven filament evaporation. This mechanism is generally not significant, although it has been reported that light bulbs sometimes have a slight, measurable decrease in lifetime from DC operation as opposed to AC operation. In a few cases, AC operation may shorten the life of the bulb, but this is rare. In rare cases, AC may cause the filament to vibrate enough to significantly shorten its life. In a few other rare cases involving very thin filaments, the filament temperature varies significantly throughout each AC cycle, and the peak filament temperature is significantly higher than the average filament temperature. Ordinarily, one should expect a light bulb's life expectancy to be roughly equal for DC and AC.

I personally use DC for tube heaters just for the peace of mind. There's already enough potential sources of hum and noise from other parts of a circuit, that an additional rectifier, smoothing cap, and voltage dropping resistor is worth the extra dollar or two. Also, I don't use "premium" tubes, so even if there was a slight drop in tube life, it wouldn't matter to me.