Tweet
I've been trying various noise reduction strategies on the JMI Vox AC30 I've been working on, and I realized I'd overlooked one of the obvious ones: filament balance. Duh.
In stock form, one side of the filament supply is ground-referenced at V1. Advice on sites that discuss the AC30 is to use the standard approach of two 100 Ohm resistors to balance the filament supply at the V1 socket. I found that it made no difference in the noise level whether the filaments were balanced at the V1 socket or at the pilot light, where there's a little more room to work. On this example, at least, the result was better if the filament ground reference was on the preamp tube sub-chassis vs. the power amp tube chassis. Why? I'm not sure, other the fact that it's sometimes hard to predict which ground point will work best.
In this case, referencing the filament supply directly to chassis ground vs. elevating the filament reference to the output tube cathode voltage made no difference.
I also tried clipping in a hum balance pot, which produced better results than any fixed resistor arrangement, but mounting is an issue, depending on whether or not you want to drill a new hole in the chassis. Lowest noise was at about a 60/40 point, but it also varied somewhat depending on which volume control was turned up. Minimum noise was found at slightly different pot adjustment points for the Normal and Brilliant channel inputs.
Your mileage may vary, but those are my results :-)
In stock form, one side of the filament supply is ground-referenced at V1. Advice on sites that discuss the AC30 is to use the standard approach of two 100 Ohm resistors to balance the filament supply at the V1 socket. I found that it made no difference in the noise level whether the filaments were balanced at the V1 socket or at the pilot light, where there's a little more room to work. On this example, at least, the result was better if the filament ground reference was on the preamp tube sub-chassis vs. the power amp tube chassis. Why? I'm not sure, other the fact that it's sometimes hard to predict which ground point will work best.
In this case, referencing the filament supply directly to chassis ground vs. elevating the filament reference to the output tube cathode voltage made no difference.
I also tried clipping in a hum balance pot, which produced better results than any fixed resistor arrangement, but mounting is an issue, depending on whether or not you want to drill a new hole in the chassis. Lowest noise was at about a 60/40 point, but it also varied somewhat depending on which volume control was turned up. Minimum noise was found at slightly different pot adjustment points for the Normal and Brilliant channel inputs.
Your mileage may vary, but those are my results :-)
Comment