Here's what I have so far.
Each channel will have its own +-18V supply. A third supply will deliver +-12V for meters and +9V for controls. I will also have a soft start board and ground lifts for signal ground on each side.
Each channel will have its own +-18V supply. A third supply will deliver +-12V for meters and +9V for controls. I will also have a soft start board and ground lifts for signal ground on each side.
I wanted to scribble down a rough schematic of the grounding scheme.
The smoothing cap reference will be close to the grounding point which will be near the enclosure inputs (top of diagram). Supply ground will connect to chassis ground via (switchable?) ground lift circuit. Signal ground will connect directly to supply ground near input. I am not entirely sure about any of this, because there seems to be many differing opinions on the best grounding practices. I have also read that star grounding a preamp is not an effective scheme for reasons which remain unclear to me.
The smoothing cap reference will be close to the grounding point which will be near the enclosure inputs (top of diagram). Supply ground will connect to chassis ground via (switchable?) ground lift circuit. Signal ground will connect directly to supply ground near input. I am not entirely sure about any of this, because there seems to be many differing opinions on the best grounding practices. I have also read that star grounding a preamp is not an effective scheme for reasons which remain unclear to me.
I forgot to draw the wire for GND from the tape deck. That will connect at the common supply ground as well.
Got some parts in today and put together the mute and EQ select circuit. Happy to report it works as intended.
The top LED is simulating the output signal, which "mutes" when the wiper is between positions on the non-shorting switch. Remaining two LEDs are simulating the EQ selection, stripped down to two choices for demonstration. There is a brief delay during switchover to allow the mute circuit to disengage the output signal, after which the previous selection switches off. The new selection engages after a brief delay. Finally, the mute circuit reengages the output signal.
It seems completely reliable even when I switched positions as fast as I could, so I believe this will be an effective method of preventing thumps when switching curves.
The circuit is the same as the one I posted previously in #41 with the exception of the addition of D6 to prevent a voltage from C43 appearing on the inverter's input when the switch is open.
PS
I used 3501-05-511 relays to test this out instead of the slightly more expensive 3502-05-511. The 3501 and 3502 have the same coil specs, so they will function the same in the build.
The top LED is simulating the output signal, which "mutes" when the wiper is between positions on the non-shorting switch. Remaining two LEDs are simulating the EQ selection, stripped down to two choices for demonstration. There is a brief delay during switchover to allow the mute circuit to disengage the output signal, after which the previous selection switches off. The new selection engages after a brief delay. Finally, the mute circuit reengages the output signal.
It seems completely reliable even when I switched positions as fast as I could, so I believe this will be an effective method of preventing thumps when switching curves.
The circuit is the same as the one I posted previously in #41 with the exception of the addition of D6 to prevent a voltage from C43 appearing on the inverter's input when the switch is open.
PS
I used 3501-05-511 relays to test this out instead of the slightly more expensive 3502-05-511. The 3501 and 3502 have the same coil specs, so they will function the same in the build.
Spent a few days working on a PCB layout.
Designed to fit in a Pesante 300. I consolidated the power supply to save space. Back side of PCB is used for ground traces rather than a pour to reduce coupling. I did my best to keep the feedback traces as short as possible, but with limited success due to the footprint size of film capacitors. I also have to layout the board for the +-12V and +9V supply that will be shared between both channels (controls and meters).
Designed to fit in a Pesante 300. I consolidated the power supply to save space. Back side of PCB is used for ground traces rather than a pour to reduce coupling. I did my best to keep the feedback traces as short as possible, but with limited success due to the footprint size of film capacitors. I also have to layout the board for the +-12V and +9V supply that will be shared between both channels (controls and meters).
Thinking about the user interface to this. I'd like to consolidate the power switch and EQ select to a single rotary switch. The first position will be Off, second will be NAB, third will be IEC 7.5, fourth will be IEC 15. So a 4 position, 2 pole switch should work for this. One pole opens the live wire at position 1, and closes at positions 2, 3, 4. The other pole opens the feedback loop at position 1, and closes the loop with the respective curve at positions 2, 3, 4. The issue is that, in order for the mute circuit to mitigate thumps during EQ select, the switch must be non-shorting, thus causing the device to power off between positions 2, 3, and 4. I think I figured out a way to resolve this issue using a dual coil latching relay.
The connector pin 3 will connect to one pole on the switch, pin 2 connects to the off position (reset coil), and pin 3 connects to positions 2, 3 and 4 (set coil). With US mains power, the voltage drop across the 2880 ohm coils should average ~10mA, just slightly more than required to energize. When position 2 is selected, the set coil energizes, closing the live wire and powering on the device. It remains powered until position 1 (off) is selected and the reset coil is energized, opening the live wire and powering off the device.
At "standby", the current through R1 and the voltage drop through the diode bridge should consume ~20mA which is acceptable to me.
Haven't tried this on a bread board yet, just brainstorming ideas here.
The connector pin 3 will connect to one pole on the switch, pin 2 connects to the off position (reset coil), and pin 3 connects to positions 2, 3 and 4 (set coil). With US mains power, the voltage drop across the 2880 ohm coils should average ~10mA, just slightly more than required to energize. When position 2 is selected, the set coil energizes, closing the live wire and powering on the device. It remains powered until position 1 (off) is selected and the reset coil is energized, opening the live wire and powering off the device.
At "standby", the current through R1 and the voltage drop through the diode bridge should consume ~20mA which is acceptable to me.
Haven't tried this on a bread board yet, just brainstorming ideas here.
Wish I could edit my post... pin 1 connects to positions 2, 3, 4. And voltage drop across the coil should be about 24V, current ~10mA.