jdotw / opinst-bass-1 Goto Github PK

The header for NoiseToSubFiltMixCV goes to the opamp for NoiseToOsc2FiltMixCV, and vice versa.

Not even sure why this is there.. should be just a direct connection?

Had to snip them from the first protoboard

You never know when a pin will become useful!

e.g when rotary delta overflows

So that the MCU can cause a power-on reset on demand of all attach peripherals

We need to be able to service an interrupt on a bus without worrying about the state of a push bus.

The oscillator DACs should be on a separate bus to allow for high resolution in updates.

Anything interruptible (rotpics) should be on a separate bus.

We shouldn't use the LEDs on the rotpics because then it's a combined push and pull bus that cant be interrupted.

The periodic update (push) components should be on a shared bus + mux if needed

Looks like we are pulling up to 5v on our side of the opto isolator and sending that to the MCU. It would be better to pull it up to 3.3v.

There's not enough clearance for a header

At the moment all LEDs are expected to drain through the PIC. Switch this around so we can use PWM/CCP easily.

Surface the I2C and SPI busses as headers, along with the various GPIO lines, and allow for any MCU to be connected to this

To avoid the snake of hard-wired cables in v1

There are 4 unused local labels beneath the U5A/B/C/D in sector D/E/F 8.

Using weak pull-up in MCU?

The pinout is NOT the same as the Osc1 header. Scale & -5v, and Note & PWM CV are transposed!

Need to find a different interrupt port for the SD_DETECT

otherwise any LED that flashes will be out of sync

Doesn't connect to the voice board

M2SP has an onboard antenna

The header for Osc1ToOsc1FiltMixCV actually goes to the opamp for Osc1ToOsc2FiltMixCV, and vice versa.

Using a single opamp to do both the summing of the CV and the scaling causes the offset voltage to become the virtual ground which keeps the filter CV at 3.4v

Turns out on a 7905 that's actually Vin

That way we're always using the same address

If each graph followed the same pin pattern that would vastly simplify code

positive and negative terminals are inverted. Had to snip them to get the board to power up.

To avoid mirror/rotation in software.