Bulb in cathode?

[dgta]

I was recently told about a scheme whereby the output tubes (in a p/p) are tied together to ground via a small lamp (light bulb). The output tubes have separate grid bias as well.

According to this guy, this was used on some older amps. I haven't seen it. Is this true, and if so, what is accomplished by having the non-linear resistor in the cathodes?

[Wavebourn]

I once used a bulb in cathode of self-split PP, it was a toy project, but it sounds with an electric guitar incredibly!

I saw some commercial amp that run output tubes in class A, and filaments of preamp tubes were used as cathode as resistor for output tubes. They were shunted by capacitor. As the result, bias was more stable (almost CCS) by current, and filaments of preamp tubes were powered by DC.

[Rod Coleman]

Originally Posted by Wavebourn

I saw some commercial amp that run output tubes in class A, and filaments of preamp tubes were used as cathode as resistor for output tubes. They were shunted by capacitor. As the result, bias was more stable (almost CCS) by current, and filaments of preamp tubes were powered by DC.

The 1950s DanElectro/Silvertone guitar amps do that, too.

check the schematic:

Silvertone Download Page - silvertone-1433-manual-amplifier-schematic.pdf

Output PP 6L6s run their combined cathode current - about 22V/120mA - into [12V mode] 12AX7 heaters.

Works OK, but I don't believe the hard life of the 6L6s (from running 75mA, 22,5W) is justified by much sonic improvement.

[dgta]

OK, I understand the filament scheme. But my source (an old timer) claims to have seen amps where they used a small light bulb, like a #47 panel bulb, as the common cathode load, sometimes in parallel with a fixed resistor.

I'm thinking a bulb has a high positive temp coefficient, so an increase in tube current would make the bulb resistance higher and bias the tube back to lower current. Could it be just a protection scheme? I don't see any way the bulb can actually enhance normal operation.

[leadbelly]

Wavebourn already answered this, but you missed it. Stop focusing on resistance. The point to the light bulb is that it has significant inductance, as in CCS, as in Class A PP. Look at some of the Nelson Pass designs using light bulbs as CCS's.

[Wavebourn]

Originally Posted by Rod Coleman

Output PP 6L6s run their combined cathode current - about 22V/120mA - into [12V mode] 12AX7 heaters.

Works OK, but I don't believe the hard life of the 6L6s (from running 75mA, 22,5W) is justified by much sonic improvement.

6L6s are transmitting generator tubes, actually. They are nice when used in RF oscillators, biasing themselves by rectified RF they generate. But for audio it was a commercial step by RCA, to repackage them and push on consumer market. If you analyze their plate curves, in order to work in linear region they have to dissipate 150W that would melt them. For Hi-fi sound they must be backed, and anyway it would not be an optimal regime, just a compromise between sound quality and lifespan.

[dgta]

Perhaps I did miss something. I'm not intimately familiar with Pass's designs, but I did read his Zen stuff. In the article I read, he says nothing about inductance. He states that he's using light bulbs as cheap high power resistors that don't need heat sinks. He actually states that he's approximating a CCS by a high voltage source and large high power resistor. Do you have a link to a paper where he discusses using the inductance of light bulbs as a circuit feature?

[SY]

I haven't measured one, but we used to use light bulbs as dummy loads for HF transmitters- I doubt that would have worked if their inductance was significant.

A lightbulb is a cheap power resistor, and it provides some stabilization since its resistance increases with average current. This property has been used for Wien bridge oscillators.

[leadbelly]

On second thought that sounds correct. I googled it and even the most complex filaments are only 2 feet of coiled wire. Sorry for the confusion.

[Wavebourn]

Originally Posted by leadbelly

On second thought that sounds correct. I googled it and even the most complex filaments are only 2 feet of coiled wire. Sorry for the confusion.

What confused you, that filaments indeed have higher dynamic resistance than static one, but it is not because of inductance, but because of temperature dependence of resistance of the wire.