The external diode is not needed in this case. There is a diode inside the ULN2803 between the output and its COM pin which is intended to serve this purpose. Just connect the COM to your +24V or whatever is the positive supply of the relays.

I shamelessly borrowed the basic design and logic from a existing open source opto isolated relay board that fit my needs (ref: https://www.electronics-lab.com/project/8-opto-isolated-relay-board/ ), so I can't claim full credit for working through all the engineering behind it, I know just enough to grab the parts I think I need to make it work and then shuffle it around a bit to make it fit my application.
I try my damndest to figure out the how of the functionality while I'm at it.

I have looked over the PCB layout and see one ground is used for everything which defeats the purpose of using opto couplers.
The ULN2803 is intended for 5 volt input. The ULN2804 is made for the 12 volts you are using.

Steve, that diode is not across relay coil

Attached is the schematic how it should be connected.

BTW I don't like the routing, and you have a common ground which voids opto isolation.
Even your board does not look like schematic that you used.
What is the 12pin connector for? Are you going to place wire links from that one to 8pin one to establish connection for series resistors for opto couplers?

Please look again. If you connect the COM pin of the ULN2803 to the +12V or +24V that supplies the relay coil, then the internal diode is indeed across the coil. That was the whole purpose of including it inside of the ULN2803 device.

The PCB traces to the relay contacts could be wider, especially the one from the 24V input which is common to all. I am guessing you will have only light logic level loads but if you control large solenoids that should be addressed. A dead short might destroy the common trace, a fuse would be good if your 24 volt supply can supply high current.

There are multi-relay opto isolated boards on eBay that would likely fit your needs.

If you have parts and are trying to save this design tell us more about what you are connecting to the inputs and outputs.

The PCB traces to the relay contacts could be wider, especially the one from the 24V input which is common to all. I am guessing you will have only light logic level loads but if you control large solenoids that should be addressed. A dead short might destroy the common trace, a fuse would be good if your 24 volt supply can supply high current.

There are multi-relay opto isolated boards on eBay that would likely fit your needs.

If you have parts and are trying to save this design tell us more about what you are connecting to the inputs and outputs.

Outputs are going to a bit of a gas piano controlled by these:
https://www.smcpneumatics.com/VQZ2151-5LO1-Q.html

Inputs are a bunch of 24v PNP sensors to replace the old NAMUR sensors that were originally there and lost their translator with the old black box controller.. Currently everything is getting mechanical limit switches that have WAY too big a variance for my comfort.

The primary aim of building an expansion board is to avoid/eliminate the rats nest that I had to clean up previously, further I'm not adverse to ripping up the existing board traces and reworking them. I left a significant amount of room on the traces anyway and could likely just expand the size based on recommendations.

Lastly, thanks for pointing out that I grounded everything on the same circuit. Considering I've got the single 24v supply point, where else would I ground it? Chassis?

Steve, that diode is not across relay coil

Attached is the schematic how it should be connected.

BTW I don't like the routing, and you have a common ground which voids opto isolation.
Even your board does not look like schematic that you used.
What is the 12pin connector for? Are you going to place wire links from that one to 8pin one to establish connection for series resistors for opto couplers?

As I've said, electricity isn't my first language, so mistakes are due to ignorance. I really do appreciate you folks taking a look and having a think on what I'm doing.

The inputs are to the PNP proximity sensors. The manual for the C11G BO board I'm using requires a 27kOhm resistor with those 24v sensors before it gets to the board, so that's what the 12 pin connector is for. Rather than adding another separate board to manage those I'm trying to consolidate it to one.

The board I'm pulling from is a source, but I didn't directly copy and paste it since I wanted to do more than just have a relay board and eat up real estate with another board for the sensors and another for the driver pin out and another for the tool changer, etc, etc, etc.

If I am decoding the solenoid part number correctly the coil is 24 volts 0.45watt with an indicator light and surge suppression. That means it draws only 19 mA so no problem with trace size. Still, it wouldn't hurt if you can increase the trace width. Who knows what the future may bring.

I looked at the C11G manual to see how the 27K resistor from 24 volts works into a 5 volt input. If you set the jumpers for pull down the 4.7K pulldown and the 27K resistor form a voltage divider that presents 3.5 volts to the buffer IC inputs. That's above the nominal 3 volt threshold.

If your not going to put opto isolation on both the inputs and outputs you could just connect the C11G outputs to the ULN2803 and run its outputs to the 24 volt solenoids, no 12 volts needed.

You need 2 supplies to have opto isolation between your inputs and ULN chip because you NEED different Commons (aka GNDs) for each side (input side and output side).
If you do not need galvanic isolation, then you can omit optos and drive ULN inputs directly. Relays are, anyway, galvanic isolation between your inputs and pneumatic valves.

Preliminary "reverse engineered" board from your picture, omitted voltage stabilisator.

If ZASto's board is using a ULN2803 it is not connected correctly.
The relay contacts are connected differently too.

Also I'm wondering where the 12 inputs connect as the C11G has only 5 inputs.