1/4 Power Switch For 100W Twin

[Apexelectric]

I have a Fender Twin from 2001 with a 1/4 power switch and was curious to see how they engineered this feature in this amp. I found the schematic that was posted on an older thread on this forum and it looks like they accomplished this by using the center tap to produce half of the B+ on the power transformer and the DPST switch to remove one of the legs of the winding while simultaneously rebasing the output tubes with a higher value resistor.

There are a ton of diodes in the system, especially around the filter choke, and a few other things I’m not terribly familiar with that make this set up a little confusing. Does anyone out there have a good working knowledge and some practical experience with this circuit. It seems like a worthwhile option for a higher powered amp unless I’m missing something that I’m not considering.

Any info would be appreciated

[PRR]

> There are a ton of diodes in the system

Yes, and IMHO most of it is not worthy of plagiarizing.

The raw B+ is 470V. Cheap e-caps go to 450V. So we need stacked caps.

Caps stacked do not split the voltage equally. Maybe we should force equality?

Since they were already committed to a CT for half-voltage, the first filter caps have a "natural split". The CT will be half of the HV.

First odd detail: D204 D205 form "the rest of a full wave bridge" when S203 is open for quarter-power. But the CT with D201 D203 already make a full wave rectifier. I'm not sure what D204 D205 bring to the party. It may "do something", because C203 was put in, and oddly it is 500V-rated.

Whatever. We have the hi-current 470V plate supply and need a lower and cleaner 470V for screens. So we use a choke and another cap-stack. How to equalize? We have (just like at C201 C202) a pair of 220K resistors. But when S203 is opened or close the voltages change suddenly. Voltage spikes? Who knows?? D206-D211 form a bidirectional 95V Zener. (Four '4006 were cheaper than two Z47v5.) This clamps the midpoint to C204 C205 to be within 95V of the PT CT.

Is all this necessary? The resistors tend to force near-equal split on the stacked caps. Moreover two caps from the same crate are likely to have fairly equal leakage. And when they don't, the one with the less leakage will take the higher voltage, which will increase leakage, and limit the voltage rise. 350V caps across a 470V rail, we could tolerate 340V+130V, almost 3:1 difference. Also the leakage of a 47uFd is likely to be under 1mA (a random datasheet says 0.37mA). For 1mA+0.0mA unbalance (unlikely) in 110K (two 220K) we have 110V max unbalance or 345V+125V. Fully legal even at worst-case. As for the switching transient, e-caps are pretty robust against short overloads.

So my reading is that the Designer was paid by-the-part. Or had to satisfy a less experienced Boss who kept asking "What if...?" Anyway in a large electronic gizmo, adding a diode (even 8, but apparently not 10) is small change, not worth designing-out. Or teaching the Boss.

Since the B+ and G2 voltage changes, we also need to change G1 bias.

But not heaters or opamp supply.

It does seem odd that Y+ and X+, presumably driver and preamp supply points, are caused to fall to half. IMHO the preamp up to the first gain control knob "should" have consistent supply for consistent finger-feel. I may need to cut my acoustic output, but my fingers have an optimum force-zone and having to gentle-down will change my fingering, which really is not the point of quarter-power mode.

Keep looking. I think I have seen simpler plans more worthy of plagiarizing.

The literature and lore of Power Scaling and similar technology does stuff like this, over a wider range, but keeps the monkey-motion down to one card and a few wires.

[Apexelectric]

PRR,

Thanks for the reply. That’s kind of what I read into it except spoken less eloquently. I too thought it was kind of odd to drop the preamp voltages as well but I guess it was cheaper than not having to do so. Im sure they thought it sounded good enough.