Why only half-wave rectifier on bias?

[Axtman]

Why is the bias on tube amps only a half-wave rectifier? I typically only see a single diode, dropping resistor and capacitor.

Why not a full wave rectifier, like a bridge rectifier, with a high value capacitor to smooth out the ripple?

[galaxiex]

Cost, and full wave not really needed.
Half wave works just fine and amp makers are penny pinchers, or the bean counters are in charge....

[Enzo]

Agree completely. There is almost no current drawn from the bias supply, so there is no need for extra filtration and smoothing.

[loudthud]

KOC uses full wave on some of his amps where bias comes from the high Voltage secondary. It helps lower the impedance of the bias supply which is important when you have a high B+ and want individual pots for each of four output tubes.

When bias comes from a secondary tap, there is no option for full wave unless you want another tap or a floating winding like the Traynor YBA-1.

[Enzo]

Even amps with a separate winding often go half wave. for example Peavey. 5150, MAce VT, etc.

[nickb]

When bias comes from a secondary tap, there is no option for full wave unless you want another tap or a floating winding like the Traynor YBA-1.

^^^ This is the primary reason.

A secondary thing to consider as that a half wave rectifier will produce ripple with a fundamental frequency equal to the power supply i.e 50 or 60Hz. Guitar speakers are generally very inefficient as such frequencies making the ripple less audible. If your were to double the frequency for the same ripple amplitude the speaker will certainly produce more audio output.

A simple improvement for bias hum in a PP tube power amp is to add a bias balance pot.

[g1]

A secondary thing to consider as that a half wave rectifier will produce ripple with a fundamental frequency equal to the power supply i.e 50 or 60Hz. Guitar speakers are generally very inefficient as such frequencies making the ripple less audible. If your were to double the frequency for the the same ripple amplitude the speaker will certainly produce more audio output.

Plus the ripple should be cancelled out in push-pull anyway if it is reasonably balanced, or is that what you were hinting at with the balance pot comment?

[nickb]

Plus the ripple should be cancelled out in push-pull anyway if it is reasonably balanced, or is that what you were hinting at with the balance pot comment?

In a perfect stage the common mode gain would be zero. At idle the common mode signal is the bias ripple and so the output would be zero. A balance pot allows you to adjust that common mode gain closer to zero. Reality is of course more complicated, but that's it in a nutshell.

[Gtr0]

In a perfect stage the common mode gain would be zero. At idle the common mode signal is the bias ripple and so the output would be zero. A balance pot allows you to adjust that common mode gain closer to zero. Reality is of course more complicated, but that's it in a nutshell.

This may seem like common sense... but...
At which point do you look for the balance to occur? I assume scoping the output of the PI, where the output coupling caps meet the original bias resistors?

[Enzo]

Just use your ears, balance for lelast hum.

[Gtr0]

Just use your ears, balance for lelast hum.

cool... the simple route!

[trobbins]

The traditional technique is to balance for minimum hum at idle - in the olden days for PP output stage, when speakers weren't too sensitive, that was done by inserting a small signal transformer in the B+ feed, and listening to ear-phones connected to the secondary of the small transformer.

Nowadays with super-sensitive soundcard spectrum analysers, you can balance for a large signal condition - which will then take in to account PP stage tube dynamic differences - but you'd have to be keen