DC-coupled 12AU7 cathodyne PI

[EETStudent]

After spending a lot of time under the hood of my '67 Ampeg Reverberocket 2, I noticed the cathodyne PI is fixed bias(I believe). It is powered by a section of the 6U10 tube that is equivalent to a 12AU7. Ive done a lot of reading, and I understand that 12AU7's can be great platforms for a cathodyne inverter. I am however a little unsure how to tweak the circuit for optimum peformance.

Ampeg gs-12r schematic image by Travst on Photobucket .... This link should take you to a copy of the GS-12R schematic.

I understand Ampeg designed these amps to be very clean, but I would like to acheive a better overdriven tone. I have read where others have designed cathodyne PI's around 12AU7, say from a stock 12AX7, but they were AC-coupled. It was recommended that they lower the plate and cathode resistors(relative to 12AX7), and increase the bias resistor. What adjustments can I make to the given circuit to yield best performance, bandwidth and increased voltage swing. Thanks

[loudthud]

The link you provided is to a fixed bias cathodyne version. On that schematic the phase inverter is AC coupled to the gain stage. On this version: http://www.el34world.com/charts/Sche...peg_GS12R_.pdf the cathodyne is direct coupled to the gain stage.

The AC coupled circuit probably has the most headroom. In the direct coupled version the plate voltage of the gain stage must be set lower than normal to prevent clipping of the PI. Bandwidth is a function of bias current and capacitance. Capacitance is fixed by the tubes used and the layout so higher current means higher bandwidth. The AC coupled version gets the gold star because of the lower value resistors set the currents higher.

[EETStudent]

The link you provided is to a fixed bias cathodyne version. On that schematic the phase inverter is AC coupled to the gain stage. On this version: http://www.el34world.com/charts/Sche...peg_GS12R_.pdf the cathodyne is direct coupled to the gain stage.

The AC coupled circuit probably has the most headroom. In the direct coupled version the plate voltage of the gain stage must be set lower than normal to prevent clipping of the PI. Bandwidth is a function of bias current and capacitance. Capacitance is fixed by the tubes used and the layout so higher current means higher bandwidth. The AC coupled version gets the gold star because of the lower value resistors set the currents higher.

Loudthud,

I do see that schematic I posted(which is accurate to my amp) is AC-coupled. I was under the impression that because that cathodyne is fixed bias, it is different from your typical cathodyne(ala 5E3). The difference Im seeing is that the grid DC voltage source comes from the B+ via the 5.6M R, whereas in a 5E3 it comes from the 1M feedback resistor. Assuming that is correct, Im guessing the bias in my schematic is determined by the 5.6M/1.5M resistor network, as opposed to the 1.5k cathode resistor in a 5E3. Are my ramblings at all correct?

If so, I would still like to somehow calculate the best operating and component values for my amp.

[loudthud]

Loudthud,

I do see that schematic I posted(which is accurate to my amp) is AC-coupled. I was under the impression that because that cathodyne is fixed bias, it is different from your typical cathodyne(ala 5E3). The difference Im seeing is that the grid DC voltage source comes from the B+ via the 5.6M R, whereas in a 5E3 it comes from the 1M feedback resistor. Assuming that is correct, Im guessing the bias in my schematic is determined by the 5.6M/1.5M resistor network, as opposed to the 1.5k cathode resistor in a 5E3. Are my ramblings at all correct?

If so, I would still like to somehow calculate the best operating and component values for my amp.

The 5E3 uses cathode bias which is OK until you want to stick a lower gain tube in there, then the bias is hit or miss. Fixed bias would let you stick anything in there and the operating point would only move slightly. The 12AX7 is about the worst choice for headroom because of the high cathode to plate voltage resuired to keep the grid negatively biased.

In general, start with the plate curves for the tube, the B+ voltage and the plate and cathode resistors. Plot a load line using the sum of both resistors. Determine the plate to cathode voltage where the load line crosses the zero grid bias line. Subtract that voltage from B+. Divide that voltage by 4. This is the target cathode voltage. Divide that voltage by the cathode resistor to get the tagget cathode current.

This part is a little messy. You would like to know what the grid bias needs to be to get target cathode current. Take B+ and subtract the taget cathode voltage. Draw a new load line from that voltage using just the plate resistor. Find the bias voltage that will give the target cathode current. The target grid voltage is the target cathode voltage minus the grid voltage it takes to get there. Design a voltage divider to obtain that grid voltage.

You may ignore the grid bias voltage and just use the target cathode voltage if you want. It won't make much difference.

Alternate method: Warmup your scope, install a pot from B+ to ground and adjust the pot until you get nice big signals at the power tube grids or it sounds good. A circuit simulator could also be used for added pain.

Edit: Check my profile under Visitor Messages where I posted links to threads about the Paul C Mod which covers cathodyne biasing.